I didn't know what to expect out of Paris, and I kinda thought I'd not fuck with London - but in reality, both my partner and I really fucked with London, and kinda really didn't enjoy Paris (Other than hanging out with a friend!¹).

Paris was kinda rough initially, insanely busy and unwelcoming. It wasn't until we went a bit out of the city that we actually could enjoy how pretty everything was, and I do mean that. Paris was Beautiful - but it was just to packed and had way too many cars driving around. We met up with a local friend a few times in the city, and it was great spending time with her and yapping. Kinda stopped my partner and I from going crazy - and she helped us enjoy the city a lot more 💖.

We didn't do too many touristy things, but we did see a few things - you really can't appreciate the scale of all these monuments & buildings until you're up close and in person with them! Our friend recommended we visit when the locals are on holiday, since we ended up visiting when theres tourists and locals - which made things super packed...

After our short Paris trip we went through to London, we caught the Eurostar in and the whole process felt kinda tacked on and hacky lol. I had to scan my passport at 3 different terminals in the same room to board the train, all doing different things haha. Brexit moment I guess.

London was great because we got to spend most of the time meeting and hanging out with friends in pubs² - with a few days just spent chilling in parks! Sadly I am allergic to the UK... I've never really had allergies before but I basically was on a rotation of different meds to stop me feeling like death in and about the city, not sure if it was damp or pollen. But something was fucking me up bad !!

I've gone through my camera and I've gotten some great pictures so I'm keen to post them all a bit later in the week when I get time for the full blown holiday post!

We spent the week in a BnB in Heiloo - an actual BnB. The hosts also lived in the house still, but the floorplan had been changed so we had a private kitchen/bathroom/bedrooms. They cooked us some yummy treats during our week there, and they were very friendly!

The day we landed we met with everyone at Amsterdam Central - I got barely any sleep on the flight (RIP), but when we met with our friends they tried to keep us busy around the city. We got coffee and bagels, and just wandered around - it was very pretty, but we definetly started to crash early. It was about an hour from central to Heiloo, both my partner and I spilt off to go straight to the BnB while everyone else went to go get groceries for dinner & breakfast. Two of our friends ended up cooking bolognaise for all eight of us, and it was really yummy. But we crashed soon after...

The rest of the days there were all very low-key. Nice coffee at Dak Showroom, Stedelijk Museum, Wandering in Utrecht - lots of wandering, Meeting with old friends at The Hague, it was all very chill. Nothing crazy, nothing super touristy, just spending time with friends. It's honestly been the best holiday I've had so far.

Heiloo was absolutely beautiful, the whole country was. Amsterdam definitely felt super touristy, but it was easy enough to escape. We 100% want to go back, and ideally meet with everyone again.

I'm planning to do one more blog post with a short update on our time in Paris and London - followed by a bigger post with all the photos I took and some more details!

Most of the week was spent just doing not much and prepping for the Holiday, We flew out of Sydney on the Sunday at around midday and transferred in Jakarta for another flight directly to Amsterdam. All up it was about 23 hours or something on a plane and it sucked! We landed in AMS at about 9am, so I was really hoping I would pass out on the last leg of the flight and be well rested for the day but that didn't really end up happening lol - I got maybe an hour max of sleep ...

My partner and I did watch a bunch of movies on the plane though:

• Predator Badlands
• Godzilla: King of the Monsters
• The Prestige
• Zootopia 2
• Detective Pikachu
• The Lego Batman movie

Detective Pikachu was actually terrible, and we didn't end up finishing it - it felt like an old Smosh YouTube bit with too much budget. Otherwise, I enjoyed all the movies heaps! Predator was really good, Godzilla was mid but campy (Holy fuck they had to do so much rewriting in future movies lul), The Prestige was super intresting, Zooptopia 2 was really fun, and the Lego Batman Movie was funny. They at least helped pass the time! Will probably rewatch when I get back home on a nicer screen and speakers.

Once we landed in AMS my partner and I met up straight away with friends - ideally they would keep us awake until we got back to our accommodation that night so we wouldn't have shitty jet lag. It worked!

We did a lot of wandering around - If we sat on a particularly comfortable bench I probably would have fallen asleep, but I didn't!

I plan to write all about the trip in next week's post!

Other than that though it's been a really low-key week, just working and getting things done. Still tinkering with Nix & NixOS in my free time. Really enjoying it! I've also been more online lately and talking with oomfies on Signal & Bsky - which has been really fun.

For the long weekend, my partner and I really didn't get up too much. Mostly chilled at home, played games and read manga. I binged and caught up on an entire series yesterday¹ - was really good! Planning to wrap up the Easter Monday with a TV binge, not sure of what yet, but we've got snacks!

Next week is a short one, only Tuesday - Friday. I've got a salon appointment on Wednesday & my partner and I fly to Amsterdam on Sunday! Next week's blog post might be a little delayed since we land Monday morning in Amsterdam!

On Wednesday I had a chill day in office and did a last minute dinner with some homies at a Korean place. Didn't do BBQ, just ordered a bunch of yummy things and beers and chatted it up for an hour. It was fun just doing something out of the blue with friends + getting out of the house. I similarly had an impromptu lunch with my brother, uncle & grandpa on the Friday. I literally cannot remember the last time I had a good chat with my uncle or grandpa so it was really nice having a few drinks and eating gyoza & karaage!

I also did another shopping trip! I bought 3 new skirts from Uniqlo during my lunch break on Thursday & spent that afternoon working from a coworking space in the city. It was good to shake up the scenery and backdrop when working. I 100% need to do it more, sitting at the same desk every day at home drives you a little crazy! I also spent that night finally digging into Nix! I ended up writing a blog post just for that since I actually have finally gotten hooked on it - I set up my new laptop almost entirely with Nix! Which was actually really convenient since over the weekend the screen started having random lines show up in it ;-;;; !!! Luckily it's covered under warranty, but I'll be out of a laptop for a week, and they might wipe it - but Nix fixes this².

It was a pretty jam packed week, but I also planned catching up with my parents for lunch over the weekend. After last week I was like "I need to get outside" and I organised a lunch with them. It was really nice - but it made the whole week really busy since I didn't expect to do everything else as well haha.

That's everything for that week though, this coming week should hopefully be more chill :3c !

Well, today, I was unmedicated & while trying to work on something else, my brain decided: "Hey I reckon we could try and do nix, but like, just dotfiles".

I don't know why I never thought of this, I always see big grand setups - or huge NixOS repos, but it makes sense when thinking about things like home-manager. So I just sorta tried to do it.

I do have to admit - the docs are still fucked. But with the help of an LLM, Some oomfs configs¹, and a general idea of how things should work I ended up with a pretty simple & sane setup that I actually like and understand.

I really do not fuck with the idea of having Claude write my entire config for me, so I really mostly used it to point me in the right direction. Ever since I did up my own nvim config myself, I've very much been in the boat of "you should write your own, and understand your own dotfiles". When something inevitably breaks, you're reading someone else's code if you didn't write it. That adds even more overhead and mental energy when you're trying to debug.

I started super small, just getting git & my gpg key setup. Since I do that a lot and it's always kind of annoying. I started with a new repo and a basic flake, I want two machines - Work & Personal:

{
  description = "pfych dotfiles";

  inputs = {
    nixpkgs.url = "https://channels.nixos.org/nixpkgs-unstable/nixexprs.tar.xz";

    home-manager = {
      type = "github";
      owner = "nix-community";
      repo = "home-manager";
      inputs.nixpkgs.follows = "nixpkgs";
    };
  };

  outputs =
    { nixpkgs, home-manager, ... }:
    {
      homeConfigurations = {
        "pfych@personal" = home-manager.lib.homeManagerConfiguration {
          pkgs = nixpkgs.legacyPackages.aarch64-darwin;
          modules = [
            ./home/shared
            ./home/personal.nix
          ];
        };

        "pfych@work" = home-manager.lib.homeManagerConfiguration {
          pkgs = nixpkgs.legacyPackages.aarch64-darwin;
          modules = [
            ./home/shared
            ./home/work.nix
          ];
        };
      };
    };
}

It looks like a lot, but it's really not - Nix has some funky syntax, but it's all functions all the way down. They look like this:

{ value = { input, ... }: { foo = "bar"; }; }

In a fucked up comparison, think of it like this js:

const value = ((input, ...rest) => ({foo: "bar"}))()

So this base file defines a few variables: description, inputs, and outputs.

inputs is where all our packages come from, I've just got standard nixpkgs and home manager. These get passed into outputs! In my outputs I then configure everything via home-manager with two separate setups for both personal and work use.

What's cool about this is I can switch between them really easily:

home-manager switch --flake ".#pfych@personal"
home-manager switch --flake ".#pfych@work"

and everything flips over like nothing happened!

The actual home configs themselves are quite simple too, I point nix to the right packages and then import modules - these are just separate files super similar to this one! Let's look at personal.nix:

{ pkgs, ... }:
{
  home.username = "pfych";
  home.homeDirectory = "/Users/pfych";

  services.gpg-agent.pinentry.package = pkgs.pinentry_mac;
}

It's a function, that takes in packages (that we don't actually use 💀), and then it sets some variables that are system dependant. Like usernames, home directories, etc. Ideally I'd just have one darwin.nix and linux.nix but my Work setup is slightly different so it won't be that simple ;-;;

The more interesting file is ./shared! This is where the main config is! Shared is actually a folder containing a few files: git.nix, gpg.nix and finally default.nix - when you import a folder as a module nix looks for default.nix to import.

Here's what these files look like:

{ pkgs, ... }:
{
  home.stateVersion = "24.11";
  programs.home-manager.enable = true;

  imports = [
    ./git.nix
    ./gpg.nix
  ];

  home.packages = with pkgs; [
    gnupg
  ];
}

The default.nix first enables home manager for everyone and sets a stateVersion. I then import my git & gpg config, and then install gnupg with home.packages. This part is neat since it actually installs gnupg in my $PATH regardless of OS - so I don't need to use homebrew or pacman, etc.

My git.nix and gpg.nix then just have some basic config I'd like to share everywhere:

{ ... }:
{
  programs.git = {
    enable = true;

    settings = {
      user = {
        name = "pfych";
        email = "contact@pfy.ch";
      };

      init.defaultBranch = "main";
      commit.gpgSign = true;
      tag.gpgSign = true;
      push.autoSetupRemote = true;
    };

    signing = {
      key = "347543D390119D28232AA5D64CC16AAFFB23D192";
      format = "openpgp";
    };
  };
}

and

{ ... }:
{
  services.gpg-agent = {
    enable = true;
    enableSshSupport = true;
    sshKeys = [
      "D2785A4719E88BFEBB1F713B1E83E7CF9AB92225"
    ];
  };
}

This is really cool! So far it basically is just mirroring my existing yadm dotfile repo but using nix gives me a nicer lsp, typechecking and other goodies when working with all different applications configs!

The last little bit with nix that always confused me was "Where the fuck does all this config go?" - "Everyone shares cool repos, but when I set up xyz everything was in /etc??". Well, it turns out I'm kinda dumb - the answer is it doesn't really matter!

I've just got the repo set up in my home folder as ~/nix, but I could just as easily move it to ~/.config/nix or ~/Developer/nix if I wanted to - It'd just change the commands to build a little:

# Assuming I'm at ~ and my config is in ~/Developer/nix
home-manager switch --flake "./Developer/nix#pfych@personal"

I'm glad I pushed a little and actually tried to get something working this time. I'm going to slowly start moving my existing dotfiles repo over to my nix repo whenever I have some downtime. Just like my dotfiles repo, I'm content once it's set up and I don't find myself fiddling with them, & I'm hoping the same thing extends to nix!

So far I've managed to move over my nvim config & my Firefox config! Firefox was the coolest since I have all my chrome, plugins and settings sync'd now with nix instead of doing weird symlink stuff with yadm!

Whenever this happens I do try to be positive though, reflect on what I've been up to, and things I've done. It really helps when my partner has a really good memory¹, since she can always remind me of what we've been up to - & I'm really appreciative of it. Last weekend was a bit slow, one of those "chill & quiet" weekends - so it feels like you did nothing and wasted time, but we did actually go out and have a nice lunch and get a run in. It's small things that I want to try to focus on more.

Another thing I find that helps me remember things is taking more photos! It's soooo good to look back in my camera roll and see pictures from the past week to help me reflect on what I've been up to. We actually started last week with a quick IKEA run - IKEA is actually super chill if you go at 6pm on a Monday vs the weekend lol. We grabbed a mirror (for selfies), some fake plants and a new desk for the study. The fake plants are mostly for hard to reach background greenery, things that u don't rlly pay much attention to. We prefer real plants where we can but some spots, like on top of bookshelves, aren't really realistic to easily water, and they're back there enough that when combo'd with real plants it'll look nicer. We also grabbed a desk for our study, which until now was a glorified storage room, but we've cleaned it out and have moved the 3D printers in there for now. Excited to use the space a bit more in the future!

On Tuesday we went over to a friend's place for dinner, as well as to pick up my new Macbook Air. We just chilled out and played with their cats & they got me some gifts for my birthday too which was really nice. One friend got me the book "Strange Houses" which I was saying I really wanted to read, so I'm keen to start reading it soon!

The rest of the week was quite slow and average - nothing too cool to focus on. I think that's why over the weekend I was kind of crashing, feeling like I did nothing, when really I did do shit - it was just at the start of the week. Maybe I need to do these reflective posts more often.

It was my 25th Birthday on Saturday which I enjoyed with my partner by going out for a nice brunch, window shopping and relaxing. She also booked in a surprise birthday dinner with some friends on the Friday beforehand, we went to Yebisu and ate lots of yummy food (and drank a bit) - it was good fun. Everyone chipped in and got me lots of trinkets and goodies from TGSWIIWAGAA¹ as well as a CD of "Lonely people with power" by Deafheaven, which was really nice :3c !! I also got sent some postcards and cards from Family which are up on the fridge, even the ones from my siblings calling me old.,,

I managed to get the laptop I mentioned last week! I've got an M4 MacBook Air with 16GB of RAM for AUD$800, Absolute steal! I pick it up this week on Tuesday so it's exciting! Much better than the Neo I had my eyes on.

Otherwise, the week was kind of slow. I played a bit more IIDX & Gitadora Drums on the weekend while out with friends and I really enjoyed it. I need to get back into the habbit of going into the city and playing more, I'm a bit washed so it'd be good to get back into it. I did also manage to build out a "real time" estrogen level monitor for my homepage. It's very simple right now and just works on a basic generalized Bateman function, but I think it's funny. I'll probably do a writeup on it when it's got a bit more polish.

I've been wanting to get myself a personal laptop for a while now. I've been using an old ThinkPad T440, but it's mostly a tinkering device that I use to fuck around with things like nix and other Linux whatevers. I actually got thinking about it again with the MacBook Neo announcement. The fact the first thing I thought about was the RAM probably says it's not for me, but it did get me talking out loud again to my partner about getting a laptop. This out-loud-thinking reminded her she probably still had her old university MacBook lying around, and after a quick search we found it and got me set up.

Right now I'm typing away on this 2017 MacBook Pro with a 2.3 GHz Dual-Core Intel Core i5 and 8GB of RAM! It's stuck on Ventura, so I actually had to check out MacPorts for the first time since Homebrew didn't work, and it's actually quite nice. The heat this laptop puts out does suck though and the battery is terrible - but surprisingly I seem to be mostly CPU limited, not RAM limited... maybe I do get the pink MacBook Neo after all....

I did reach out to a friend who works at a big-box store to see if they had any old display units going into the "For Staff" pile, she's said there might be some M2 MacBook Air's dropping in. If they're under $1000¹ I'll definitely jump on it! But genuinely I kinda am surprised how usable this older MacBook Pro is, and how the RAM hasn't really made me feel like I'm missing anything! Makes me think!

Other than the Laptop, we did try to get out of the house on the weekend and go window shopping. I've got no money that isn't already allocated, so I didn't plan on buying anything... that was until Alice took me into Kaika - since she was surprised I'd never seen it before! $44 later, I'm now the proud owner of a Miku Paint Girl figurine².

I've also started playing some Shmups more, specifically Deathsmiles and DoDonPachi DaiOuJou. I've got a crazy doohickey setup with my MCon controller to play vertical shmups in bed, and it's been fun to play more casually. I do wish the ROM check was a bit faster (or even skippable) in emulators. Deathsmiles is a lot harder than I remember, but it's still super fun.

It's my Birthday this week - so hoping next week's update will be a bit more action packed. I tried to plan something with friends but everyone just eyed my partner and were like "uhhhhh yeahhhhh" so I think she's planned something for me! I'm excited :3c

On Thursday, I took a short day at work by stacking my hours up into the other days - After lunch I clocked off and went into the city to go shopping with my partner and one of our friends. The goal was simple, at least TWO outfits. Right now I have like, just a few outfits, and I lack a lot of basics - so it's hard to put something together. I wanted to change this so we booked in this shop. We first hit Uniqlo, & I picked up a blouse and some new pants. Then we went to Myer, and had zero luck finding anything that didn't give off bad white woman vibes¹. Afterward we hit H&M, and just picked up some reaaaallly basic tees with a nicer cut than my current ones, mostly to use as undershirts!

We ended up going back into Myer afterward since I made a comment about not having many comfy bras. My friend wanted me to try on and get my size for Calvin Klein, since she was telling me it was super comfy. I tried on a few bras and tops and found my size², but we had to go to a different store to actually buy the things I wanted. I got a pyjamas top & decided to order the rest online later, since I was already running my budget thin. After that we hit up target (lol) since we wanted to see if we could scrape the rest of my budget into some more basic things. Not much luck here since everything fit & felt like shit, but we managed to get me some shorts!

We ended linking up with our friends boyfriend after this, and we went for dinner together. It had started to pelt down rain while we were shopping so we had to trek to the restaurant trying to avoid getting super wet. We ended up getting Malaysian at a place called Sydney Kopitiam Cafe, since my friend had been going there with her family since she was little and was craving it. It was really fucking good. Apparently it's usually impossible to get a table, but the combo of pouring rain and random weekday meant that it was just us there³! My partner and I are suckers for small places with good family vibes so this has 100% been added to our eating out rotation.

The rest of the week was relatively chill - on Saturday we joined some friends at a Mardi Gras themed spin class before going and getting brunch together. Had the most fucked up yummy French toast to share and it went so hard, the coffee was also insanely good too!

It's kind of crazy how we're already into March. It's my birthday next week⁴ & my partner and I go to Europe together in April... This year's going so quick!

Doors opened at 7pm, with an opening act played by Lawrence English, and then Merzbow. The show was at Oxford Art Factory, a small venue in Sydney, but one with good vibes - especially for a show like this. The venue was surprisingly busy, but we managed to get in right as doors opened and secure a seat up the back on the brick wall. It wasn't until I went to get a beer that I realized the entire venue was packed shoulder to shoulder - Something I'd surprisingly not seen here before.

Lawrence started the show by coming out on stage, chatting with the crowd a little bit, and reflecting on previous shows, before starting his set. He emphasised how the show is about the people around you, and the shared experience, which I really resonated with. I was actually unfamiliar with his work, but I'm glad I got to experience it live - the ambience was great, the low rumbles and resonance of the room made his whole set an experience that I was sharing with everyone else. Closing my eyes, looking down at my feet, staring into the dimly lit crowd - it all set a mood to think. I really enjoyed it, & so did my partner!

There was a brief intermission after Lawrence's set before Merzbow came out. We grabbed another beer, and chatted about how we really enjoyed the soundscape and "journey" that set had taken us on - visualizing things in our heads, feeling the music, just - taking in the vibe. While Lawrence was playing, there was very minimal lighting, just the dull red ambient lights of the venue. Perfect for just, zoning out. It all meshed as an experience - and I think that's what made it so good.

This set was a stark contrast to Merzbow, I knew the music would be more abrasive & confronting. While Lawrence spent most of his set with bassy rumbles and ghostly overtones - Merzbow started with harsh, gritty, white noise. A whole different experience, but one that I was still excited to feel.

However, there was something different, that honestly - kinda took me out of it. Instead of low light, or even some strobes - there was a cheap projector projecting video over Merzbow onto the wall behind him. At first, I really paid no attention to it, but it was weirdly hollow - it was all looping clips, ken burns, weird looking CG renders. Like, at a glance they're unremarkable, but after seeing the same maybe, 10 clips rotate... it started to look off. They weren't hypnotic, abrasive, loud, they weren't ... well, really anything. That's when I started thinking - fuck, are these visuals all AI generated?

They had no soul, they all had the exact same duration, and they all ... felt flat. I'm always down for fucked, weird, visuals during a show - like damn, even just white noise would have had more "soul". But instead what we had were these weirdly unrelated, kinda clashing 3D renders playing behind him. It made it really difficult to zone out, and enjoy the vibe. Skill issue? I don't know. But coming from the great opening act from Lawrence, Merzbow somehow didn't clear the bar he set, and I was left disappointed. The music was good, it was what I was there for. But the set just didn't come together as a cohesive experience. Which I feel like is critical for this genre.

Afterward, my partner and I walked around the city and talked about the show. We talked a lot about pushing yourself to engage with "difficult art" or, art that you wouldn't consume normally. She also had the same feelings as me - Lawrence was great, but the disconnect of Merzbow's gritty, real and harsh music, with soulless AI visuals felt wrong, it wasn't just a "difficult show" - it was just "bad".

It's funny saying a noise set lacked cohesion, but that's exactly how it felt. I'm 100% certain I'd feel differently if he just performed the same set to a dark room.

Otherwise, really not that busy! I did end up creating an 88x31 badge for my site visible on the homepage or on the badges page. I also added some badges for sites I think are cool or sites run by people I've yapped with.

On Yapping, I ended up yapping heaps with aria.coffee on signal. I ended up liking some of her random posts on bsky, she liked some of mine, we both realised we're Australian, and it just kinda went from there¹!

It made me realise that I'm actually really missing having a cohort of friends I can yap randomly with about tech who are local to me. Like, someone pointed out this week how bsky is suffering from the "linkedinification" of discourse. It all needs to be about "What the next big moves are", "How can we scale", "What are other players in the space talking about" etc.

Originally this point was about bsky but reflecting on my conversations with coworkers, and other ppl I've worked with etc, it's all like this. I'm not saying I don't enjoy chatting with them. But I've realised that I actually yearn for standard friendly banter with people that's tech adjacent².

I do already have a space with friends online where I can talk like this, but it's really refreshing to have people in the same timezone as me (or close to me) where we can just... yap. IDK, maybe I'm just yearning for trans community or something, people who are like me and have shared interests. For ages this was really just my partner, so I'm kind of excited to try to get out a bit more and interact with more ppl.

I've got some fun stuff planned for this week though so I'm excited for that!

Last week I finally got the Azur lane figurine I ordered 4 weeks ago! The box was actually HUGE and the figurine doesn't even fit on our shelf ;-; ... She's SOOOOOO pretty though, and I'm glad I finally got her. I'd fucked up the post code on my delivery address so it got sent literally everywhere except my suburb. Shoutout to the AusPost phone support honestly, once you get past the shit bots the people I chatted too were super understanding and explained how the new label is getting printed & what the tracking would look like etc. All really great friendly ppl!

Other than that and Sunday though my week was a bit uneventful! Sunday was crazy though - I had a friend basically disappear over the last few months, like she was still around, but never free. Turns out she was scouted to be an idol. Like the japanese idol kind. She invited us all to one of her shows and she was crazy good (and so was her other group member!) - it's kinda crazy tho that there's a scene for this in Sydney. Like, it makes sense, but its cool that niche communities like this can form and be big enough that it can actually afford to pay her and the other girls who perform. Was really cool!

Oh! Also, my NAS that died last week is all fixed! New drives arrived at the same time as the figurine & I swapped them in. The 8hr resilver was stressful, but it all went well in the end. Trying to orchestrate a RaidZ2 migration in my head without dropping crazy $$$ though, but I don't think it'll be possible. Glad I had good backups, but I dread having to restore them...

Otherwise, the main big thing that happened this week was a disk dying in my NAS RaidZ1 array! Completely out of nowhere, the controller board died & it refuses to do really anything. I ordered 2x replacement drives straight away, 1 to drop in and the other to keep as a cold spare from Server Part Deals and ended up paying almost 2-3x what I paid late 2024! CRAZY! I've currently got the NAS powered off while I wait for the replacement drive (that should hopefully come today), I don't want to risk complete data loss. I did run a few final backups of things before shutting it all down, and it did highlight some things that I didn't have automatically backing up correctly which was nice.

I migrated my ATProto PDS to my VPS so I could continue to use bsky with the NAS down and spent ages fiddling with what I thought was file corruption, when it was instead actually just the remote encryption I set up ... glad to know that all works!

Socially I had some friends over for a dinner and to hang out, we watched some YouTube mostly since my NAS was dead, but we did spend a HEAP of time browsing neocities and nekoweb for other peoples personal webpages. SOOOO much sauce on some of those sites, and we came up with a cool list of changes I can make to my site to make it less static/flat/booring while still staying readable. Stay posted for those updates!

That's really it for this week - really hoping my figurine I ordered 3 weeks ago shows up today, its meant too. Would be awesome to have it and the new drives rock up together!

I started my week getting two new estrogen implants put in - my levels and bloods look good, and I'm feeling great! Even if the implants are a bit pricey ($240 for two) vs other methods like injections or gels - The convenience factor makes up for it entirely. Honestly my main inconvenience medication wise now is actually making it out to the Dr who puts them in my cheeks lol. It is something I would like to maybe DIY someday.

Wednesday I actually went into the city and got a few arcade credits in for the first time in a while! I've been mostly playing BMS at home, but last week I talked about my friend who's moving to Perth, and we wanted to get one more set of IIDX in before he left. After playing some games at the arcade we went to Ume Burger @ Darling Quarter and had burgers and fries for dinner. Nothing crazy, but imo - Ume Burger does the best burger in Sydney.

On Friday - my MCON finally arrived after backing it on Kickstarter like a year ago. It was kind of a shit-show fulfillment wise, but it's nice to finally have it in my hands. Sadly - now that it's in my hands, it feels like shit to use. I mostly game stream from bed with my GameSir G8 or Steam Deck and was hoping that this could be a compact little controller I could keep on me to play games. But instead, when using both sticks and triggers its super crampy and small! Really is a shame, apparently it's going to be fixed in V2, but I really don't want to give OhSnap my money again due to how fucked messy the whole shipping internationally thing was. It's fine for D-Pad and face-button gaming though, so I'll still be using it. But I'm going to try to keep mine in its box in nice shape since I got the Kickstarter exclusive translucent colour. Hopefully someone who does like the controller will pay 2x what I paid ($99 USD) sometime in the future haha.

That same day my brother came over to avoid the heat & try out my sim racing rig. It's nothing fancy, just a Moza R5 combo with a clutch, but it's still a blast to drive. I've really got to get a base to screw it all into because it slides around a lot when playing, especially with the load cell brakes upgrade! The next day, my partner and I drove up to my parents place to catch up, have a few drinks and have pizza. I've been craving this pizza from the pizza place nearby for like 2 months now - nothing here in Sydney hits the spot the same. I went full fat-fuck mode (It's a mindset) and ate nearly a kilo of pizza that night. Went hard, and I felt like shit the next day but was 100% worth it.

I did get some exercise in this week though. Not as much as I would have liked, but I did a spin class this morning. Absolute hell after a massive pizza dinner - but was still fun!

Otherwise, that's about it! I've been playing heaps more Warframe as usual too - I got my first prime this week which is fun! Excited to keep grinding and enjoying the game :3c

This week has really been quite chill - lots of not much. I managed to fit two spin classes and a run in! Honestly it's how every week should be - but I'm glad I got in all in.

Otherwise, work was fun this week. We did an offsite at The Pillars club in one of their boardrooms - was crazy swanky & was fun, having staff keep my coffee topped up was dangerous but really nice. We did lots of just yapping about the company and planning for the new year. Excited for the push this year & keen to work on heaps of cool stuff. After all that work we had some drinks in the clubs bar and yapped about movies. I'm constantly reminded that I'm behind and haven't watched so many good movies. Someday ... I'll binge "the classics" - but not any time soon lol.

On watching things - my partner and I finished watching Ruri Rocks this week. It's a show about a highschooler getting into geology with the help of two Uni students. It's a really pretty show & it made geology actually interesting lol! Was cool to learn about minerals from such a cozy show :3c.

On Saturday I went to a farewell dinner for one of my friends who's moving to Perth for Med School - We went to Duk Inn which was a weird funky restaurant that's hard to describe... The menu had things like "Experimental beef cubes" and "Mystery ribs? (Order them)". It was all very very yummy though. As much as it sucks to have to move to Perth - I'm really excited for my friend and I hope he does well at Med School! My partner and I will have to come visit :3c

Next week is a "short week" since this Monday is part of the long weekend. Hoping that I can slot in some shopping at some point - I've really got to get some more business-casual summer outfits ... I can only get away with rotating the same 2 outfits in the office for so long!

Work wise - I've mostly spent the entire week planning, researching and building a proof of concept project that's pretty cool. It's nothing user-facing, so I can't share details, but I'm enjoying the challenge it brings. I'm hoping to have something production ready by late next week but time will tell :3c. We also had a new starter at the company I work at on Wednesday!

It's actually someone I've worked with in the past. It just so happened we ended up catching up for drinks, I mentioned that I was stressed hiring, which he mentioned being stressed job hunting haha. I didn't end up interviewing him since I didn't want to skew the teams opinions, but he got along well with the team and passed the interview pretty well! I'm excited to continue working with him.

Other than work I actually did some cool things with friends this week. One of my friends works at a store that houses an e-waste drop-off point. Often cool things come through, and she'll grab them (since most things still work). But this time she managed to snag an old stereo/radio/CD player thing. It's nothing crazy but the vibes it radiates made it worth the snag. Sadly one of the speakers was really quiet, and when opening it up, she had literally no idea what to do.

She sent me a message with pictures of it's guts and asked if I knew how to fix stereos, and I lied and said yes!¹ We organised an "e-waste pizza party" at my place and her and her partner (one of my rlly close friends) came over, and we got to it. We gave the whole thing a good clean and scrub - lots of corrosion from batteries on the inside, but no real visible damage. However - just the clean seemed to make the quiet speaker slightly louder.

I poked and prodded around for a little bit but couldn't pinpoint the exact issue. Which is when I tried a hail-mary. "The right side speaker works right? But the left doesn't? Are you cool with no left channel?". She was. So I daisy-chained the right speaker to the left and it worked! We played a few CDs, and it all sounded fine (if you don't mind missing out the left channel :3c).

After fixing the player we ordered pizza and sat around the TV while we scrolled AmiAmi looking at anime figurines. During the repair, I'd mentioned that I'm a little bummed I never caved and bought my own cool big figurine like my partner has (multiple times :3c), and that I'd like a big figure of my own in the cabinet that lives in the living room. We ended up pulling up a figurine I'd actually wanted but passed on before ... a figurine of New Jersey from Azur Lane in her Oath skin.

Being surrounded by friends - and a lot of back and forth umming and ahhing about budget etc... I ended up pulling the trigger. Since it was on sale for only ¥28,000 down from its retail ¥43,780 (What a steal!). However, if you've shopped with AmiAmi before you're likely aware of the issue here - shipping. After placing the order, a few days later I was forwarded the quote for shipping. ¥17,410 with EMS or ¥42,100 with DHL... Needless to say, I gritted my teeth and selected the EMS option. Making the figurine come out to ¥49,951 in total.

I think it's worth it though. She's one of my favourite ships from Azur & I've wanted the figurine in the past (I'm treating myself²).

Other than all that though, the week has been a bit slow - my partner and I spent the weekend indoors avoiding the storms & working on world-building/lore for a future project! I'm excited to share more details eventually... but not right now!

This week was my first week back working. I'm glad I had time off over the holidays - but it's nice to get back into my routine of getting things done at work. Nothing too crazy going on career wise, spent some time researching workflow engines and messing with AI. Hopefully can share more in the future!

Personal life wise - this was the first week when I really tried to stay on top of my medications, actually taking them & following my phone alerts. My latest ADHD medication gives me a little bit of Nausea, but it's not as bad as the last one which made me really grumpy and depressed. So we call that a win! Still on health, I caught up with my Physio - They've changed due to some merger whatevers, so I was a little worried, but they were good! Getting back onto those exercises this week since I was really lazy during my time off haha.

I managed to get three runs in this week as well, trying to stick to a routine with my partner on this... I set a record today for longest running workout (4.47km) and General running record (1,335Kj) today - it was quite cold after an insanely hot day yesterday so I feel like the cool breeze and overcast conditions helped me not gas myself out as quickly as usual. It felt good! Good splits & pace too - I've been doing 2:1 splits with a target of 6:20, and I'm definitely getting better at holding pace. As expected though - I do fall off at the end lol.

I also got my hair done this week - I saw my usual stylist in Marrickville, was good to yap and get my hair tidied up a little. Went back to the curly bangs instead of the side-part I'd had for a while (mostly due to me being lazy about maintaining the bangs lol). Was informed that due to my blonde hair that I need to worry about my hair bleaching in the sun!! Apparently it's nothing too bad as long as I use a good conditioner and care for my hair properly (which I will now). Also had a catchup with my Psychologist this week, and she complimented my hair which was nice! We discussed a lot of stuff regarding discipline & fear of "wasting time" and where that all stems from - was very productive!

Tech wise - my friends and I have started using IRC a little bit. I set up ZNC & ErgoChat on my NAS for a bit to tinker and then migrated them to my VPS for better reliability. It's quite unintuitive how it all works, but we've had fun so far & discussions on IRC seem to be a lot more interesting and productive than chats on discord or other platforms. We're still figuring it out, I'm interested to see how long we stick to it before dropping off haha.

I also watched my partner beat Dispatch a few nights ago - I really enjoyed the writing and story, and I'm happy with the ending we got! We played it over a couple nights, doing 2 episodes after dinner like we normally do with TV shows. Highly recommend. Otherwise, for games - I mostly just played Warframe. It's hooked us both lately, and we're just grinding through the base star-chart & campaign. I'm playing on my old account back from when I was in High school with my Xbox One - it's crazy because I remember playings HOURS with friends after school but logging back in all I have are a bunch of frames a friend bought us, a massive dojo, and like ZERO campaign progress LOL. No clue what we were up too lol.

It's now 2026, which is crazy to think about. The oldest post on this blog is from late 2021, making this site ~4 years old. I want to start blogging more here, consistently. Like yes, I did make ~16 posts, but ideally I'd like to have more general updates, maybe weekly. Mostly for my own sake - since I really struggled to do this retrospective lol, I had to dig through my phones camera roll for the year to remind myself of things I've done.

I was originally going to write some grand big long post, but it's a bit of a hodgepodge, so I've instead done whatever this is :3c

Places I went in 2025

• I traveled to Melbourne for my Partners birthday. I'd never been, so it was really great staying in the CBD & enjoying heaps of good food & drink. I definitely fell in love with the city & I would love to live there someday.

Small wins in 2025

• I really feel like I started spending time and chatting with more people in 2025, being a bit more outgoing and trying to connect. Less of a shut-in outside of work. I want to keep this going! It is exhausting constantly being the one to organise things IRL but its worth the hassle and exhaustion when I see people I care about having fun and hanging out :3c
• I also started focusing on my health more - I picked up running and cycling, and I've started seeing a physio. Again - my poor discipline kills me with staying up on these, but I'm really trying to push myself to do something since it's better than nothing. I'm proud of some of the milestones I've hit and I feel a lot better now that I am exercising more often.
• I got really into Riichi Mahjong in 2025, starting at the beginning of the year. I haven't played as much as I'd have liked to, but I've tried to make it a habit to play with friends often. I've found it really good for working on my focus, and trying to keep my mind sharp.

Big Changes in 2025

• I got a new job! After working in the same role at my previous employer for almost 6 years I'm happy to have moved on to something new. The previous role was great - but I needed a change. I'm excited to see where it takes me in the new year.
• I feel like I've hit a good point in my Transition where I'm really happy with who I am. Everyone at my new role calls me by my new name & I feel like I'm slowly figuring out my style, fashion and general vibe. It still needs work, but I'm really happy with my progress in 2025.
• My partner and I got a cat! Her name is Cami & we love her so much. She's a 9-month-old stray that a friend rescued, they think she may have been abandoned once they got past the "cute kitten" stage of their life. She's an absolute angel and a big baby who loves people & enjoys napping. I'm excited to care for her and give her the best life possible!
• I finally got medicated for ADHD, initially with Dexamphetamine, now with Ritalin. Dexamphetamine made me grumpy, irritable and angry - which lead to me being depressed (because I didn't like to be angry). However, Side effects aside - I found it incredibly helpful for my focus and decision paralysis! Never being able to start anything was always my biggest hurdle, and I do really hope that I can start more things in this new year

Things I want to change from 2025

• I didn't really take as many photos as I would have liked - Honestly I really started dropping photography in 2024, I used excuses like "Oh I mostly take family and personal photo's now which is why I don't share them". But while party true was really an excuse I tricked myself into believing. I want to pick it up again, because I really did enjoy it. The photos in this post are random ones I took during 2025 that I have not shared.

Goals for 2026

• Run 10km at a pace I'm proud of
• Read at least 3 full books and write about them here
• 100% a game and write about it here
• Hit credits in a game and write about it here
• Skateboard again (I miss it, but I'm also afraid)
• Wear makeup consistently for a week (Laziness kills me, especially when I WFH. I need practice)
• Get my nails done (I've never done this before, and it looks like fun)
• Cosplay at a convention with friends
• Start work on a game project (Even if its shit and I don't show anyone)
• Turn my expenses planner into an app
• Create an OC or a Glorbo (I want to be more creative)
• Learn to use a sewing machine
• Clear SL2 in BMS
• Ride front row at a spin class and not die
• Try my best to blog weekly about what I've got up too
• Stream more
• Stop telling people about project ideas that I haven't started

For $9.99 in the middle isle who wouldn't be tempted? I've been fiddling with assorted solar projects so thought even if they broke instantly I'll at least get maybe a battery cell & a dinky panel out of it!

They lasted about 2 days before they'd randomly start flickering on and off quickly and strobing our neighbors! It took me a while to finally get around to taking them down after turning them off (laziness etc.). But once they were down I opened them up and was surprised to see the solder connection for the LEDs to the control board was broken! Re-soldering this joint fixed the lights entirely.

Inside was a simple board with pads for the battery pack (1800mAh), the solar panel, & pins for the basic LED strip. I don't own a working multimeter right now so I didn't look too deep into everything. Thats a project for next time!

Ideally I'd like to put an ESP-32 inside here and have them be "smart" lights, but I lack the connectors to do it safely without soldering everything directly to the board. Might tackle this another day.

I have a virtual memory card with FreeMcBoot mount on startup and then when I launch a game either via OPL or with a Disc, the Memcard Pro2 automatically mounts the game specific memory card. However, I've got an actual PS2 memory card with all my game saves and I had no idea how to go about transferring these over. The first thought was to create a virtual copy of my existing memory card:

1. Create a new Virtual memory card on the Memcard Pro2
2. Insert both into your PS2
3. Use the PS2 memory card browser to copy the saves from the old memory card to the virtual memory card

Now I've got a .mc2 file with all my saves on the Memcard Pro2's SD card! By using a program like MyMC++ I can browse the memory card:

# I use uv for python apps/packages since it makes managing them easier.
# This will take a while to launch initially while it installs!
uvx -v "mymcplusplus[gui]"

Using the MyMC++ GUI I can export each of the actual game saves off the .mc2 file into individual .psu files using export button in the header.

Once exported I created the following script in the same directory:

#!/bin/bash

for save in *.psu; do
  FILENAME=$(basename -- "$save")
  NO_EXTENSION=${FILENAME%.*}
  NO_PREFIX=${NO_EXTENSION#"BI"}
  GAME_ID=$(echo "$NO_PREFIX" | cut -f1 -d" " | cut -c1-10)

  MEMCARD_FOLDER="memcards/$GAME_ID"
  mkdir -p "$MEMCARD_FOLDER"

  MEMCARD="$MEMCARD_FOLDER/$GAME_ID-1.mc2"
  uvx mymcplusplus -i "$MEMCARD" format
  uvx mymcplusplus "$MEMCARD" import "$save"
  curl -s "https://raw.githubusercontent.com/niemasd/GameDB-PS2/refs/heads/main/games/$GAME_ID/title.txt" >"$MEMCARD_FOLDER/name.txt"

done

echo ""
echo "Done!"

This script loops over each of the saves, pulls the GAME_ID from the file name, and then creates a dedicated .mc2 folder & file for the game. Ready to add directly to the Memcard Pro2!

Once copied onto the SD card the dir tree should look like this:

OS
PS1
PS2
├ SLPM-66621
│ ├ name.txt
│ └ SLPM-66621-1.mc2
└ SLPM-55117
  ├ name.txt
  └ SLPM-55117-1.mc2

Since most venues offer "Bonuses", costs can vary. For a breakdown of this math, check the "The Math" section below.

• The Venues
• The Games
  • Beatmania IIDX
  • Sound Voltex
  • Gitadora
• The Math

The Venues

The following venues offer deals where the "Real Dollar Value" can get quite low!

The Games

The following games have been logged, all tables can be sorted by clicking the header:

Beatmania IIDX

The cost of Beatmania IIDX at relevant venues:

Credits to play

Credit Breakdown

Sound Voltex

The cost of Sound Voltex at relevant venues:

Credits to play

Credit Breakdown

Gitadora

The cost of Gitadora at relevant venues

Credits to play

Credit Breakdown

I stared, watching the people there playing Beatmania for a couple of minutes before one of them, in a really friendly tone, asked if I'd like to play. I'd played a few rhythm games before, but never at the level I just watched them play at. But I said "ok" and they helped navigate me through the menus, set up an account, and configure my settings - All while a dreaded 60-second countdown loomed between each menu!

Once I picked a song I had an absolute blast. It was some of the most fun I'd ever had playing a Rhythm game, and I could see the potential as a game that I could potentially play ... forever.

After I played my credit I asked the twins who'd just helped me if there was any online community and was told that there was a Facebook Chat. I didn't have Facebook, so we had a quick back and forth to try and find some shared social platform, so we could at least add each other and stay in touch. We ended up all having WeChat accounts, which they found funny - "The white girl with WeChat" was how they described me to the guys in the Facebook group at the time which I also found funny.

Coming to Puri became a routine for me, I eventually made a Facebook account just for the chat and connected with heaps of really cool people. On weekends, at its peak, there would be at least ten to fifteen of us all hanging around, waiting to play (and sometimes we'd talk!).

Through the people here, I got into the wider Sydney arcade scene, which had really only just started growing as both Sega & Konami had expanded their online services to Australia.

It was a great community, and without Puri, I never would have met all the amazing people I now call some of my closest friends or even my partner. It was a place we could all easily meet up, go for dinner and hang out. And without this community, I don't know where I'd be, or who I'd be right now. It was the first group of people since moving out of home I could really call close friends.

Over the next 5+ years, the foot traffic dwindled. Newer, shinier arcades opened around the city pulling more people. They had convenient cards¹ you could load with money, so you didn't need to carry coins. And they had more variety.

I'd still go to the new venues to hang out with people, play new games, and just chill. But I'd always end up coming back to Puri. They were a quieter, cheaper, and safer-feeling venue. The big open atrium above the cabs and the low foot traffic made the place feel ... homely. It was always a constant, and a safe place I could fall back on that wasn't my shitty apartment or office at work.

Sadly, as of the 21st of June 2025, Puri staff have quietly let us know that they're selling all their Rhythm game cabs. There's no known date when they'll be sold, or when their last playable days are. But they've said they'd let us know when it is, so we can all get one last game in.²

It felt weird getting emotional over a venue selling machines. I genuinely got really upset, like losing a close friend.

Reflecting, it felt weird to hold a place in that high of regard. But it had been a constant in my life, every weekend, every tough day after work, and whenever I was down in the shits. I knew I had a place where I could rock up with change, hang out with friends and kill a couple hours. There are other venues, yes, but the rhythm game setup at Puri meant a lot to me. Even after a majority of the community went elsewhere.

It's not really anyone's fault that Puri is selling the machines, it makes sense as a business. They take up valuable floor space, and they probably don't pull in much money anymore. But that doesn't make it any less sad.

I genuinely don't know where my next "place" will be. All the other arcades are crowded, cramped and noisy³. I'll probably still go to get a few credits into certain games and hang out with friends. One of them just got Gitadora Drums, which I've been waiting for somewhere to get for years. But with the only other location that has a Beatmania IIDX cabinet charging $7 for a premium credit, when Puri charged $3, I definitely see myself playing BMS at home a lot more than before.

I'll miss Puri. A lot.

Goals

The primary goal of this project is to offer a torrent of uncompressed BMS charts, allowing for partial downloads & easy contribution by seeding existing BMS folders. A secondary goal of this project is to offer a hosted index of this torrent mapping md5 & sha256 to the path of the chart in the torrent, allowing for quick discovery, partial downloads and potential future integrations with clients.

What about the Big Sister BMS Project (BSBP)?

The BSBP is an amazing resource! However, the bandwidth and storage costs are a potential burden, and relying on donations to cover costs may be unsustainable in the long term.

A generalised public torrent, uncompressed, allows for longer term preservation of BMS charts, allowing for users to seed directly from their active BMS folder (Assuming they follow a flat directory structure).

This project is not a replacement, or competitor, to the BSBP!

Benefits

By distributing charts via BitTorrent, users can quickly and easily pull specific charts & updates in a distributed manner without relying on a 3rd party to host and bear the cost for serving these files.

New versions of the core .torrent file can be distributed which, through most torrent clients, will only download changed &/ new files. Allowing for quick library updates.

Methodology

This projects core will consist of a portable & easily runnable script which generates both the torrent and an index given a path.

This script will ideally be run on an existing folder containing all charts in popular and common tables.

node generate-torrent.js -- /path/to/folder # Language not set in stone!
ls /out
bms-2025-01-01.torrent    bms-2025-01-01.json 

This torrent can then be distributed by the user who generated it.

Ideally an initial centralised torrent will be created and distributed alongside the index, This simplifies onboarding of new players and ensures a high seed-count for users when downloading. A centralised distributer also allows easier updates to the core torrent when a new version is released with notifications provided to seeders via a chat service¹.

Torrents are tagged with their creation date to facilitate easier discussions with people wanting to update &/ check if they are using the latest version.

Foo: "Hey whats the latest torrent"
Bar: "2025-01-01 is the latest right now"
Foo: "Ah I'm behind, can you send the file or magnet link through?"
Bar: "Here you go: <attached file>"

In theory users could seed multiple versions of the torrent. However, there is a small chance of conflicting files if content of a distributed .bms file is changed, without changing the filename.

Folder Structure

Folder structure may be a contentious topic, with some users potentially wanting to merge the contents of the torrent with their existing filesystem & collection of BMS.

A flat hierarchy makes the most sense for simplicity: /root/<chart-name>/<chart>.<extension>

However some people may organise their collections in different ways. This is worth investigating to discover if there are any better ways to organise the torrent.

File formats

Some charts are distributed wth multiple versions. Some with and without videos, .wav & .mp3, etc.

I believe this torrent should distribute "the most correct" version of the chart, which may be a divisive topic. This means that the if a no video version is offered alongside a video version, this torrent will distribute the version with the video.

Regarding audio formats, it's worth investigating and discussing with players. The format .wav makes sense, however some charts are distributed "originally" with massive uncompressed .flac sounds. This may however, be a non-issue in practice.

The Index

Assume the torrent is laid out in the following way:

bms
└ chart-name
  ├ chart-normal.bms
  ├ chart-hyper.bms
  ├ chart-another.bms
  ├ hitsound.wav
  └ video.mp4

Provide a JSON file &/ HTTP endpoint to convert the sha of chart-normal.bms to a path:

GET /path/bf6bcb1370fe4956ceb5ea436bdb63bb HTTP/1.1
content-type: application/json

{
	"path": "/bms/chart-name"
}

{
	"bf6bcb1370fe4956ceb5ea436bdb63bb": "/bms/chart-name",
	"f4d1d4c7bbfa6007835aaab6bfe11435": "/bms/chart-name",
	"911dc9e1f0549ca12990892e113d255b": "/bms/chart-name"
}

A standardised index file can allow for a potential future integration with popular clients, allowing them to integrate with either an external (or internal) bitTorrent client. Allowing charts to be downloaded on the fly.

However, recently I had a friend reach out to me to see if I could extend it to also ping them if chapters were updated for a specific manga & my shell script definitely could not do that without a major refactor, so I used it as an excuse to rewrite it and learn some new stuff along the way.

The code is available on my git host if you'd like to check it out!

Building Docker Images

This is the first time I've ever built a project from scratch and wanted to have it run somewhere other than the cloud - this is a simple lightweight node script with no uptime requirements, so it's a perfect candidate to use to learn about containerising stuff!

I've used docker for years now running tools, programs and scripts other people have written, but I've never actually dockerized anything on my own since I thought it would be really daunting and difficult! In reality - making the actual image was really simple!

FROM node:20-alpine

WORKDIR /checker

COPY package.json ./
COPY pnpm-lock.yaml ./
COPY src ./src

LABEL org.opencontainers.image.source=https://git.pfy.ch/pfych/chapter-tracker
LABEL org.opencontainers.image.description="Ping a webhook when a new MangaDex chapter is released"
LABEL org.opencontainers.image.licenses=MIT

RUN npm install -g pnpm
RUN pnpm install
RUN pnpm run compile

CMD ["node", "./.out/build.js"]

The hard part came from the question "Now how do I run this on my NAS?".

Turns out Gitea (which is what I use to host my repos & runners) has a Container Registry component that allows you to publish images alongside a repo! I won't go into any of the actions I used to automatically build and publish the image, but it really was as simple as a few commands to publish a package to my gitea instance.

docker login git.pfy.ch
docker build -t git.pfy.ch/pfych/chapter-tracker:latest
docker push git.pfy.ch/pfych/chapter-tracker:latest

So easy!

And now I can just reference that from my compose file!

services:
  checker:
    image: git.pfy.ch/pfych/chapter-tracker:latest
    volumes:
      - "./config.json:/checker/config.json"
      - "./mangaHistory.json:/checker/mangaHistory.json" 

I'm definitely going to be more likely to write non-cloud scripts now that I've actually tried and succeeded in dockerizing a Node project. I'm really surprised it's taken me this long.

Before uploading any images to my site, I need to compress them. Images out of my camera are huge and I don't want people downloading 25mb+ images on random pages of my site. I use ImageMagick and a little helper function I wrote to do this.

I generate a sm, md and lg version of each image with the following commands:

# A thumbnail version of the image
convert originalImage.jpg -strip -interlace Plane -scale 10% -blur 0x2.5 -resize 1000% -resize 180 "sm/${1%.*}.jpg"

# A suprisingly good quality version of the image
convert originalImage.jpg -strip -interlace Plane -gaussian-blur 0.05 -quality 85% -resize 720 "md/${1%.*}.jpg"

# A nicer "HD" version of the image
convert originalImage.jpg -strip -quality 85% -interlace Plane -resize 1800 "lg/${1%.*}.jpg"

Once I've generated all the images, they're added to my sites storage in folders. My site generator then reads all the files in the sm folder, and generates a html for all the images. The html for each image looks something like this:

<img 
  src="https://assets.pfy.ch/sm/image.jpg" 
  finalSrc="https://assets.pfy.ch/md/image.jpg" 
  hdSrc="https://assets.pfy.ch/lg/image.jpg" 
/>

By default the sm version of the image will load, when the images is in view, I swap out the src with the value in the finalSrc attribute. This is done using an Intersection Observer.

const images = document.querySelectorAll('.galleryImage');

const loadImage = (image: HTMLImageElement) => {
  image.src = image.getAttribute('finalSrc') || '';
};

const handleIntersection = (entries: IntersectionObserverEntry[]) => {
  entries.forEach((entry) => {
    if (entry.intersectionRatio > 0) {
      loadImage(entry.target as HTMLImageElement);
    }
  });
};

const observer = new IntersectionObserver(handleIntersection, {
  root: null,
  rootMargin: '0px',
  threshold: 0.05,
});

images.forEach((img) => {
  observer.observe(img);

  img.addEventListener('click', (event) => {
    const image = event.target as HTMLImageElement;

    image.onerror = () => {
      console.warn('Could not get LG version of image!');
      image.src = image.getAttribute('finalSrc') || '';
    };

    image.src = image.getAttribute('hdSrc') || '';
  });
});

I would love to have the images popup or go fullscreen when they're clicked, but I didn't want to make it too complex.

Since they're just plain HTML images, you can just open them in a new tab or window. Which seems to work fine for most people who browse my photos!

If you haven't read the previous post, webmentions are a standard way of pinging a site to let them know that you've linked to them. My site has the following tag in it's <head>:

<link class="externalLink" target="_blank" href="https://api.pfy.ch/webmention" rel="webmention">

Posting to this URL with a JSON payload containing a source (somewhere on your site) and a destination (somewhere on my site) will add a little backlink to your post on my webpage acknowledging that it's been referred to elsewhere on the web! Obviously there are check's in place, but it really is that simple. As long as the source contains a direct link to destination it gets through!

Just like with my previous implementation - I still haven't gotten around to actually pinging other peoples webmention endpoints from my custom CMS... It shouldn't be too difficult, I'm just a bit lazy and weary of pinging people too much.

However, there are only so many times you can use document.createElement(); before you go insane.

How scripts work on my site

I write all my scripts in Typescript. Using Esbuild, I bundle these scripts into individual minified js files. My custom SSG tool handles hot reloads and can do all this on the fly, so there's no real friction in using Typescript over plain Javascript.

Every page on my site is defined as a Markdown file, using the front matter I can set the title, summary, publish date etc. but also include a list of scripts. By doing this, over a central script.js file (although I do have one), I can keep load times for my site low since pages will only load the scripts essential for them to function.

How I initially did interactive pages

When I first was adding JS & Interactive pages to my site, pages would look something like this:

---
title: example
scripts: old/example.ts
---

Hello World

<div class="example">

</div>

and the following script:

window.addEventListener('DOMContentLoaded', () => {
  const target: HTMLDivElement | null = [
    ...(document.getElementsByClassName("example") || [])
  ]?.[0]
  
  if (target) {
    const p = document.createElement("p");
    p.innerText = "Hello World"
    
    target.append(p)
  }
})

While this worked, the developer experience was quite tedious. Especially with larger components. I eventually wrote this function to somewhat assist with element creation:

export const e = <T extends keyof HTMLElementTagNameMap>(
  tag: T,
  attributes: Attributes<T> | null,
  children: (string | HTMLElement)[] | string | HTMLElement = [],
) => {
  const element = document.createElement(tag);
  Object.entries(attributes || {}).forEach(([key, value]) => {
    const castKey = key as keyof HTMLElementTagNameMap[T];
    element[castKey] = value;
  });

  if (Array.isArray(children)) {
    children.forEach((child) => {
      element.append(child);
    });
  } else {
    element.append(children);
  }

  return element;
};

Which turned the previous code into something like this:

import {e} from '../element';

window.addEventListener('DOMContentLoaded', () => {
  // fetch target ... 
  
  if (target) {
    target.append(e('p', {}, "Hello World"))
  }
})

This was great, and a big leap in dev QOL. But I really, really, really, wanted to use something as slick as JSX. Turns out - It's mostly just a tweak in my ts config. I could even re-use the utility function I had written before!

Using TSX

By setting the jsxFactory property in my config to the function e, when typescript/esbuild compiles a file - instead of inserting a React specific React.createElement(), it'll insert my function instead.

{
  "compilerOptions": {
    "jsx": "react",
    "jsxFactory": "jsx",
  },
}

Without setting jsx to react pretty much every linter will complain when using TSX. I also needed to create a type definition in my project root, otherwise I wouldn't get any typing. You technically get this for free if I installed React as a dependency but that's overkill.

/// <reference lib="DOM" />

declare namespace JSX {
  type Element = HTMLElement;

  interface IntrinsicElements extends IntrinsicElementMap {}

  type IntrinsicElementMap = {
    [K in keyof HTMLElementTagNameMap]: {
      [k: string]: any;
    };
  };

  interface Component {
    (properties?: { [key: string]: any }, children?: Node[]): Node;
  }
}

I ended up renaming e to JSX for easier importing and now the above script looks like this:

import JSX from '../jsx' 

window.addEventListener('DOMContentLoaded', () => {
  // fetch target ... 
  
  if (target) {
    target.append(<p>Hello World</p>)
  }
})

I genuinely don't know why I didn't do this sooner - using JSX for the interactive parts of my site is so much nicer than what I was doing before.

import JSX from '../jsx';

const postsList = (
  <ul className="postsList">
    {...posts.map((post) => (
      <li>
        {post.emoji || '📝'}
        <a href={post.link.replace('https://pfy.ch', '')}>
          <span>
            <span className="date">
              {post.date.toLocaleDateString()} -{' '}
            </span>
            {post.title.replaceAll('&amp;', '&')}
          </span>
        </a>
      </li>
    ))}
  </ul>
);

vs

import { e } from '../element';

const postsList = e(
  'ul',
  { className: 'postsList' },
  posts.map((post) =>
    e('li', {}, [
      `${post.emoji || '📝'} `,
      e('a', { href: post.link.replace('https://pfy.ch', '') }, [
        e('span', {}, [
          e(
            'span',
            { className: 'date' },
            `${post.date.toLocaleDateString()} - `,
          ),
          post.title.replaceAll('&', '&'),
        ]),
      ]),
    ]),
  ),
);

A lot of them have moved to Bluesky from Twitter, and I heard it's pretty open with it's API so I thought I'd give it a try doing something similar, but with it instead of Mastodon.

There's an API endpoint for getting a thread at /xrpc/app.bsky.feed.getPostThread but it requires a uri parameter. This URI parameter is made up of my user did, the collection and the object id. Making the final pattern something like this:

at://<did>/<collection>/<id>

I know the collection for posts is app.bsky.feed.post from reading through their docs, and the object id is just the characters at the end of the post URL, but getting the user did was a bit harder.

The API endpoint /xrpc/app.bsky.actor.getProfile allows you to get a bunch of information about a user, including their did. It's just passed as a query parameter.

/xrpc/app.bsky.actor.getProfile?actor=pfy.ch

Which returns my did:

did:plc:lpxtxymhlfbs5foxba7mfewt

So - to get this posts comments I'll pass this AT-URI:

at://did:plc:lpxtxymhlfbs5foxba7mfewt/app.bsky.feed.post/3livnrzudt22s

and it returns all replies to that post!

I dont handle embeds, threads or links yet - only top level replies. But it should be pretty easy to add those if I decide to down the line. Overall it was surprisingly easy to get working & I'll likely start posting about my stuff here more often on Bluesky since they're now intertwined with my website. Really hoping that Bluesky doesnt get enshittified and close their API down in 5 years. 🤞

However, I stumbled upon this form thread where *mr.lite.touch* provides a solution!

1. Install the LTConfig plugin in Deluge
2. Run ip addr show in both your Deluge & Gluetun containers.
3. Both containers should have a tun0 interface with an IPv4 address that is the same. Note it down.
4. In Deluge, open Preferences > LTConfig
5. Find listen_interfaces.
6. Set it to <IP>:<PORT> where the IP is the IPv4 address of your tun0 interface & the Port is whatever was there before (42676 in my case)
7. Click "Apply" and "Ok"
8. Select all errored torrents and right-click > Update Tracker

From what I understand, this forces libTorrent to use a specific interface when listening for tracker connections. By default, libTorrent will listen on all interfaces, which I think is why it wasn't working before & would randomly throw these errors.

flowchart LR
    subgraph Container 
        Deluge --  tun0  --> Gluetun
    end
    
    Gluetun --  tun0  --> WAN

These changes immediately fix the issue for me. Hoping that it doesn't come back...

Rules

• Each player starts with 25,000 points
• 2-8 manzu tiles are removed (only 1 & 9 are used from the set), all other tiles are used
• Non-player seat winds count as Dora (optional). When drawn, the player can call "kita" and draw from the dead wall. Kita calls can be ronned, but do not award chankan. Players can still call riichi.
• Red 5's are used
• Chii is not allowed

Setup

1. Shuffle all the tiles
2. Create 4 walls of 13-14-13-14 length
3. Dealer rolls 2 dice and breaks the corresponding wall
4. Count 3 to the left of the dead wall and add those tiles
5. Draw as normal

If you would like to support me and what I do with this site and elsewhere online please click the following link to make a donation via Stripe, all purchases are in AUD and are completed via Stripe Checkout:

Choose your price!

You can also donate with Cryptocurrency:

• ETH:

  • 0xE1f27cbcE0576b114789660Ccc3C2B3fb932569D
    

• XMR:

  • 47Ysk63QvQ6GKQ4wF7fQkpAhhShjobAHtTdpkceSaqWUSB48XLzs5ycRcBqoktu2NVLXwMvsNkdWKgNa9x5TsbqNSs3YTBT
    

Thank you to the following people for either donating or reaching out with kind words!

• Maxime, regarding my Godot tutorial
• Jeff, regarding Samsung monitors
• Thomas, regarding Samsung monitors

• The Tiles
• Basic Hands
• On your turn
• Yaku & Scoring

The Tiles

The Sticks suit (Usually call "souzu"), The 1 of sticks is usually a bird.

The Pin suit (Usually called "pinzu")

The Number suit (Usually called "manzu")

Honor tiles, consisting of 4 winds (East, South, West & North) and 3 dragons (White, Green & Red)

Basic Hands

The goal of Mahjong is to create a hand with 14 titles. This hand contains 4 sets of 3 and a pair. Like so:

Sets of 3 can either be a sequence or three of a kind.

Three of a kind:

Sequence:

(You cannot make a sequence out of the honor tiles like dragons or winds!)

On your turn

Each turn begins with you drawing a tile from the wall, this draw should bring your hand to 14 titles. Once drawn, you can either:

1. Declare that the drawn tile won you the game (Tsumo)
2. Discard a tile from your hand, returning it to 13 tiles

When you discard a tile, other players have the option of "stealing" it.

Any player can take the tile to make 3 of a kind (Pon)
Only the player after you can take the tile to make a sequence (Chi)

If a player takes the tile, they must place the sequence or 3 of a kind to their right open for everyone to see.

This continues until somebody either draws a tile to complete their hand (Tsumo) or they "steal" the tile from someone to complete their hand (Ron).

You can steal for 3 of a kind from any player, but you can only take a sequence from the player before you.
You can steal for anything if it wins you the game.

Yaku & Scoring

Riichi Mahjong scoring can be a little confusing - especially for new casual players.

You start with 25,000 points, the game ends after each player has been dealer once & if someone has 30,000 points, or if someone has negative points.

The value of a hand is determined by the amount of Yaku & Fu. As a beginner you should use a scoring calculator, but it's important to know some basic Yakus' since you'll need at least 1 in your hand to score points.

Here's a few easy to remember Yakus:

• Menzenchin tsumohou (Closed hand)
  • Not stealing any tiles, and drawing your winning tile.
• Riichi (Closed Hand)
  • Announce you are 1 tile from winning with a closed hand. You must discard all draws unless they win you the game. You can win from Drawing or "stealing" (Ron)
• Tanyao (Can be open or closed)
  • Win without 1, 9 or Honor tiles in your hand.

• Yakuhai (Can be open or closed)
  • Have at least 1 set of dragons, Round winds, or Seat winds in your hand.

• Toitoi (Can be open or closed)
  • All 3 of a kind & 1 pair

If you won by drawing (Tsumo) the point value of your hand is taken from everyone's points, If you won by stealing (Ron) the point value of your hand is taken from that players points. If you think someone is about to win make sure you dont discard the tile they need!

It takes the following plain text: 119p 19s 19m 1234567z
Then renders out the following:

Currently, it just parses my markdown files on my site and renders out the tiles at runtime, but I'll eventually make it an NPM package if I get around to it.

Interactive demo

Update the input below to view your hand. Use MPSZ notation.

Code Sandbox

I've put a small amount of the code into a sandbox for you to poke around with. I actually discovered a few bugs with my sandbox implementation, so I'll eventually need to fix those... This code is in plain JS and not Typescript - which the main implementation is. Another limitation of my sandbox sadly.

All Demos

• Intractable 3D HTML cube
• Fake 3D grass mesh in CSS
• Drawer over content
• 3D Transform on image with shadow
• Simple dot following cursor

You can check the page out here

By default – Javascript is disabled and not evaluated until you enable it. This can be seen in this demo which will only set the background to pink once you enable JS and interact with the editor. Be warned though, once Javascript has been enabled – if a script loops endlessly you may lock up your browser. Don't run anything here you wouldn't run yourself in the console!

Saving to URL was an unexpected issue for this project, Most URLs can only be about 2000 characters long¹! This was overcome by using a GZIP implementation to first compress the data before inserting it into the URL parameter. This obviously doesn't solve every use case, however the editor will now warn you when trying to load a code snippet that is too long from the URL.

I used the following packages to help with development:

• sass - Compile SCSS into CSS
• pako - GZip Implementation
• prism - Web native syntax highlighting

Some of the main hurdles I encountered while working on this project were:

• There are no children inside a <textarea /> element, to modify styles of written content you must first mirror the content to another element. Syncing the display & scroll state of these two elements to create a seamless editing experience was a bit of a challenge. However once they were synced up, prism did a majority of the heavy lifting.
• Compressing the code into a URL safe string was quite difficult. Initially I implemented an LZ77 compressor which worked quite well with plain strings however it really only lowered string size by 5-10%. I was told to look into GZip, this implementation seemed much too complex to do myself, so I opted into using an existing solution. The most popular library I could find called pako only worked with UInt8Buffer's which I had never worked with before. It was interesting to work with the string -> UInt8Buffer and back conversion process.

An unexpect side effect from this project is I now have inline code blocks & previews using plain iFrames! I don't know if this is a good or bad thing yet haha.

The apartment was an old red-brick, renovated housing commission. By renovated I mean; the owner threw a new lick of paint on the walls and replaced half the tiles in the bathroom before going: "Ok, lets get a Tennant in!".

• Some of the lights didn't work, and were bare LEDs soldered in a fixture (so I couldn't replace them).
• The bathroom ceiling was dropped to add an exhaust fan (that exhausted back into the bathroom through a hole in the ceiling)
• There was a solid mould problem
• My neighbour above me LOVED to fight his furniture at 1am

This was my first unit living alone, and it was my own place. I took care of it. I decorated it with books, hung skateboard decks off the walls, had a dope sound system and TV hooked up to my PC. I really tried to make it my own. I lived there for almost 3 years with rent only being raised once - $290 a week to $320 a week (with no lease). It was a pretty good setup.

I made so many memories in that unit. Both good and bad. But I'm not upset I'm not there anymore. I mean sure, at the time locking the door for the last time ever was a little sad. But I don't think I'd ever go back.

It's the small things you don't think about.

• I can open my blinds without worrying about someone breaking into my unit
• I can wake up to the sun
• I have a kitchen I enjoy cooking in
• I have space to store my stuff

These are all things my last unit didn't have that I didn't think I needed. I had a dope place on my own for cheap, so what if I cant open the blinds or get sunlight? My kitchen has space for a microwave?! Why would I need a good cook-top?

Eventually though, I did end up moving. I started to feel more comfortable moving into a place with someone else and splitting the bills. Maybe I don't need my own cave? It's possible to have my own space while also sharing part of it with someone else?

Today I caught myself cleaning. Like just instinctively, not like picking shit up but I mean like I was scrubbing tiles. I never would have done this at the old place, because I think deep down I really didn't actually like that place - So I didn't take care of it. This new unit has sun, space, places to put my shit, a bathroom that doesn't spawn mould like it's life depends on it. It's genuinely improved my mental moving, even if it is an extra $5 more rent...

I used to joke with people that I take better care of other peoples place than I do of my own, and I don't think that applies any more. I actually care about this space, it makes me feel comfortable. I don't unconsciously flee from it on weekends or during time off, and its been really nice to recognise all this now that I've moved.

To preface this rant, I feel the need to clarify that I absolutely love working with Serverless and AWS professionally. Over the course of more than 2 years, I have deployed numerous applications across various industries, gaining valuable experience. Serverless has allowed us and our clients to scale and grow into multi-million dollar ARR businesses without ever having to worry about infrastructure.

This said however, I still don't like to recommend this stack to people. I always say something along the lines of:

"Oh Serverless is great if you're being paid to learn it."

Hopefully, in this post I can describe some of my pain points with Amazons official offerings around Serverless, & Why I'll be using SAM over the third party Serverless Framework for most of my personal projects from here on out.

Serverless Framework?

If I'm being completely honest, deploying to a serverless environment is a complete shit-show. There's heaps of ways to do it, they're all slightly different, but they all compile down to one thing in the end: Cloudformation.

In almost all my projects whether it be at work, or personally, I used the Serverless Framework. It abstracts writing cloudformation in relation to Lambdas' and makes it easier to focus on writing code. However, Serverless Framework is not AWS Native, and I only ever deploy to AWS. Also, a lot of the functionality I take advantage of in Serverless Framework is based on community plugins which can break at any time with zero warning. I want my infrastructure, which is critical to my work, to be backed by a large body & stable, not something maintained by someone in their free time.

Serverless also seems like they're trying to push their Dashboard and extra tooling, which I don't need.

Serverless Framework L's

• Lots of functionality I use is based entirely on community plugins which can break
• Serverless Framework tries to support all cloud providers when I only need one
• Serverless Framework has their own motives and goals with the library to sell their dashboard

Serverless Framework W's

• Easy to use straight up plain Cloudformation if required
• Lots of community support & plugins
• Can easily start focusing on your applications code

CDK?

CDK instantly gets written off by me just by glancing at its docs. It's an extreme abstraction of Cloudformation. Where you write Typescript, run a magic CLI command, and it spits out Cloudformation files. Any sort of black box transformer seems like an issue waiting to happen, especially when it's tied to a completely different language.

I don't want another layer that can potentially break when deploying my infrastructure. That's what I'm trying to avoid & one of the reasons I want to swap from the Serverless Framework. No matter the abstraction you will have to manually write and manage your cloudformation at some point. Adding extreme abstractions on top of it only delays the inevitable.

CDK also requires the use of AWS SAM to run locally, so I don't really understand at a glance why I'd want to use CDK over SAM? Personally, I haven't used CDK at all. I jumped straight into SAM, but from the brief foray into CDKs docs I don't feel like I'd ever use it.

SAM

AWS SAM is so close to being what I want it to be. It's an extremely light abstraction of Cloudformation at a glance, and it allows for extremely easy local testing & deployment. However, it is not without its downsides.

SAM Requires the use of Docker, which the Serverless Framework does not. This adds an extra hurdle to initial development when getting started. I'm not speaking for everybody here, but personally I find docker (on macOS¹) a bit of a nightmare to set up. Do I install docker desktop? or docker engine? Do I install it through brew or do I use the pkg on the website? These are all things that as a developer who's done it before are trivial questions, but I can see these being asked by coworkers who have never used docker before. This isn't a slam on SAM, but it is a pain point I do foresee briefly in the future if professionally the company I worked for migrated from Serverless Framework to SAM.

Another major pain point with SAM for me was getting DynamoDB to work. In a perfect world DynamoDB should just work right out of the box and get stood up if your template file contains table definitions. Serverless Framework had² a plugin which did this. Instead, Amazon recommend you run the DynamoDB .jar locally, or stand up a docker container yourself. The docker steps also add an extra layer of confusion for new developers since it recommends adding your apps image (what?) to the docker-compose and linking them. If you're just running a SAM app locally all you need to do is create a virtual docker network and put Dynamo's container on it, then start sam with an extra flag:

sam local start-api --docker-network <network-name>

This is not mentioned in the docs, and no error is thrown on SAMs end about this. It's up to the developer to notice there's no connection, realise it's due to SAM actually being in a docker container, and then knowing how to put both containers on the same network...

This stackoverflow answer is how I found out about this.

Oh, by the way there's no way to seed table definitions from SAM into this container now that it's stood up. Here's the gross script I whipped up to seed from a template file into the container. There is probably a better way, but I'd rather focus on writing my actual applications code. (SAM should do this if there are table definitions!!!!)

# Convert the Template YAML into JSON that create-table expects
# Dump this to a temp file so its easy to "while read" later
yq -o=json '.Resources | filter(.Type == "AWS::DynamoDB::Table")' ../template.yaml \
  | jq \
    '.[] | {
      TableName: "local\(.Properties.TableName[1][1])",
      KeySchema: .Properties.KeySchema,
      AttributeDefinitions: .Properties.AttributeDefinitions,
      GlobalSecondaryIndexes: .Properties.GlobalSecondaryIndexes,
      BillingMode: "PAY_PER_REQUEST"
    }' \
  | jq -c '.' > tables-seed.txt

# Remove Null GSI if it exists since create-table will crash if a value is null
sed -i '' 's/,"GlobalSecondaryIndexes":null//g' tables-seed.txt

# Run create table against each line in the seed data file
while read -r line; do
   aws dynamodb create-table --no-cli-pager --endpoint-url http://localhost:8000 --cli-input-json "$line"
done < tables-seed.txt

My final pain point was SAMs built in ESBuild integration. I love ESBuild. It's made my life 10x easier when working with Typescript. However, I was unable to get the built-in auto-build working when deploying.

This is how I bundle manually:

esbuild src/handler.ts \
  --target=es2020 \
  --platform=node \
  --external:aws-sdk \
  --sourcemap=linked \
  --outfile=.build/handler.js \
  --bundle

This is how SAM can do auto bundling:

Resources:
  HelloWorldFunction:
    Type: AWS::Serverless::Function
    Properties:
      # ... Function properties
    Metadata:
      BuildMethod: esbuild
      BuildProperties:
        Format: esm
        Minify: false
        OutExtension:
          - .js=.mjs
        Target: "es2020"
        Sourcemap: true
        External:
          - aws-sdk
        EntryPoints:
          - handler.ts

I have no idea why, but in one of my projects, the SAM bundle is 21.25MB while the manually bundled one is only 1.3MB. The SAM bundle also doesn't work in a Lambda... Which is entirely the point. Luckily its relativity easy to use your own bundler, but you lose out on live reloading. I'll definitely come back to this one day, it's likely a poorly documented feature or there's something wrong with my config.

AWS SAM L's

• Abstracted docker container management, not an issue until you want to use DynamoDB
• DynamoDB is not handled out of the box with SAM
• Cannot seed from a SAM template into a DynamoDB instance
• Documentation & Community support is lacking
• Struggled to get an API gateway stood up
  • This was confusing:
    HttpMethod: '*'
Cors: "'*'" (Quotes in quotes AND inconsistent!)
• Connectors seem cool but are an abstraction that makes things confusing and seem like they'd become a nightmare to maintain, easier to write IAM roles.
• Built in ESBuild just straight up didn't work when I used it, worked when I manually built with ESBuild.

AWS SAM W's

• Running API Gateway locally rocks
• ESBuild if it worked would be amazing
• Not extremely abstracted from the underlying cloudformation
• Easy to use existing tooling or own tools alongside, or in replacement of, SAMs tooling.

Less complaining!

Even though I've just spent the last while ranting about pain points in SAM, I can see its potential. It's almost there. I want to use it over the serverless framework. The main pain points I talked about were huge hurdles for me getting started which I only jumped over because I was being paid to jump over them. But, now that I know how to do it properly; I'll likely be using SAM for all my personally projects going forward.

The benefits of a serverless stack for me far outweigh the cons if I don't factor in the time learning how to properly use it. Serverless is amazing, and I do see it as the future of cloud computing, but the documentation and new developer journey is abysmal and extremely discouraging.

Recommendations for new Developers

• Use SAM, or plain Cloudformation.
• Use plain IAM roles to define permissions:

MyFunction:
  Type: AWS::Serverless::Function
  Properties:
    # ...Properties
    Policies:
      - Version: '2012-10-17'
        Statement:
          - Effect: Allow
            Action:
              - dynamodb:Query
              - dynamodb:Scan
              - dynamodb:GetItem
              - dynamodb:PutItem
              - dynamodb:UpdateItem
              - dynamodb:DeleteItem
              - dynamodb:BatchGetItem
            Resource:
              - !Sub ${myTable.Arn}
              - !Sub ${myTable.Arn}/index/*

• Include your stage name (prod/dev) in table names and names of other infrastructure items.
• Learn how to properly take advantage of DynamoDB (I'll make a post about this one day!).
  • Properly index your data.
  • Use multiple tables.
  • Be mindful of Read/Write units.
• If using SAM set up a custom API Gateway API, it's worth it for when you do eventually need it:

ApiGatewayApi:
  Type: AWS::Serverless::Api
  Properties:
    StageName: !Ref Stage
    Cors: "'*'"
    MethodSettings:
      - ResourcePath: "/*"
        HttpMethod: "*"
        ThrottlingRateLimit: 100
        ThrottlingBurstLimit: 500

• Use Express & Typescript with a single handler for most APIs:³

/** handler.ts */
import express from 'express'
import serverless from 'serverless-http';

const app = express() // Theres waaayyyyy more stuff here usualy.
export const handler = serverless(app, {})

app.get('/my-function', (request, response) => {})
app.get('/my-function/sub', (request, response) => {})

MyFunction:
  Type: AWS::Serverless::Function
  Properties:
    # ... Properties 
    Events:
      Root:
        Type: Api
        Properties:
          Path: /my-function
          Method: any
          RestApiId:
            Ref: ApiGatewayApi
      Sub:
        Type: Api
        Properties:
          Path: /my-function/{any+}
          Method: any
          RestApiId:
            Ref: ApiGatewayApi

• Ensure everything runs locally as well as once deployed!
  • If you've split your infrastructure into separate stacks it should be easy to deploy to a temporary testing stage.
  • If working with multiple developers it is a horrible idea to test by deploying to dev or prod.
• Set up GitHub actions, Bitbucket Pipelines etc. to deploy when pushing to dev or prod, keep it predictable!
• If it seems too complicated it probably is. Serverless (sadly) currently has lots of ways of doing one thing. Research everything!

I've written a newer post about this here. This article is still mostly relevant regarding the implementation, but the code is written using SAM - which I no-longer use.

I recently learned about the W3 Spec for Webmentions!

This spec is described by W3 as:

[...] a simple way to notify any URL when you mention it on your site. From the receiver's perspective, it's a way to request notifications when other sites mention it.

I've always wanted to write an implementation for a spec and this one did not seem extremely complex, plus I had a long weekend coming up, so I thought it'd make for a fun challenge!

The Plan

1. Write a serverless implementation of webmentions.
2. It should use AWS Cloudformation to stand everything up automatically
3. It should be cheap to run indefinitely
4. It should be testable locally

Breaking a receiver down, there are really only 3 endpoints required:

1. Submit a webmention
2. Get the status of a webmention
3. Query webmentions

The basic flow of our function to receive webmentions is as follows:

sequenceDiagram
    participant User
    participant Lambda
    participant DynamoDB
    
    User ->> Lambda: Send webmention request
    Lambda ->> Lambda: Confirm valid request
    Lambda ->> DynamoDB: Create status object
    DynamoDB ->> Lambda: Return status object identifier
    Lambda ->> User: Return 201 & correct headers
    
    Lambda ->> Lambda: Confirm mention exists
    Lambda ->> DynamoDB: Create webmention object
    Lambda ->> DynamoDB: Update status object
    
    User ->> Lambda: Query webmention
    Lambda ->> DynamoDB: Request webmention
    DynamoDB ->> Lambda: Return webmention
    Lambda ->> User: Return webmention

Cloud?! Lambda?! What's the cost?!

I've chosen using DynamoDB & Lambda on AWS since when used correctly they're extremely cheap and efficient. Both are charged based on usage so if you're running a small site like mine you can expect the bill to come out at basically nothing. Lets do some quick math.

DynamoDB

DynamoDB charges based on read & write units. With queries against an index costing 1 read unit and writes costing 1 unit per index on the table. The downside of Dynamo is that if you expect your data structure to be complex or have intricate relationships - it can end up costing much, much, more than a standard database.

Amazon currently has the cost of read and write units priced at the following:

Then if we break down how many units we use per function in our application:

We can calculate the cost as the following:

Lambda

AWS Lambda is a serverless compute platform where you can just run code. Amazon will provision servers globally for you, and these servers will only run while your function is active.

We run our lambda with 512mb of memory³, and with a timeout of 15 seconds. AWS charge per million requests and per second the function runs based on the memory used.

Assuming worst case scenario and our function runs for 15 seconds we're paying 0.0000003245 per execution. In reality the cost is less than this, especially when querying for webmentions & status objects since its runs at almost sub second speeds.

Total stack cost

Combining both our DynamoDB & Lambda costs we can figure out at what point our bill with become "excessive" (more than a dollar).

This means that It'll take around 1,540,880 webmention creation requests before our bill exceeds a dollar! We can use AWS Budgets & API Gateway Rate limiting to stop execution before we hit this, which I'd highly recommend. Nothing is worse than an unexpected bill at the end of the month.

Implementation

Our final stack should consist of the following:

• 1 Lambda: Running express with 3 entry points.
• 2 Tables: One with a single index and the other with two indexes.

Tables

The status table is extremely simple, since it only really needs to store basic information temporarily. It would be possible to add a TTL to records, however for now I plan to manually clean this out over time.

The table consists of a single index: id. Which will be used by other services to query the status of webmentions.

statusTable:
  Type: AWS::DynamoDB::Table
  Properties:
    TableName: webmentions-status-table 
    AttributeDefinitions:
      - AttributeName: id
        AttributeType: S
    KeySchema:
      - AttributeName: id
        KeyType: HASH
    BillingMode: PAY_PER_REQUEST

The mention table has a few extra definitions, most importantly it's target global secondary index (GSI).

We use id as the main unique index, however we do not query it. Instead, our functions query based on the webmention target. By indexing this field, we can get Webmentions for a page at incredibly fast speeds.

mentionTable:
  Type: AWS::DynamoDB::Table
  Properties:
    TableName: webmentions-mention-table 
    AttributeDefinitions:
      - AttributeName: id
        AttributeType: S
      - AttributeName: target
        AttributeType: S
    KeySchema:
      - AttributeName: id
        KeyType: HASH
    GlobalSecondaryIndexes:
      - IndexName: target-index
        KeySchema:
          - AttributeName: target
            KeyType: HASH
        Projection:
          ProjectionType: ALL
    BillingMode: PAY_PER_REQUEST

Lambda

We then define our Lambda & its corresponding API Gateway. This is pretty chunky since we also need to give our function permission to access the tables we've just defined.

  ApiGatewayApi:
    Type: AWS::Serverless::Api
    Properties:
      StageName: !Ref Stage
      Cors: "'*'"
      MethodSettings:
        - ResourcePath: "/*"
          HttpMethod: "*"
          ThrottlingRateLimit: 100
          ThrottlingBurstLimit: 500

  Webmention:
    Type: AWS::Serverless::Function
    Properties:
      MemorySize: 512
      Timeout: 15
      CodeUri: .build
      Handler: handler.handler
      Runtime: nodejs14.x
      Architectures:
        - x86_64
      Policies:
        - Version: '2012-10-17'
          Statement:
            - Effect: Allow
              Action:
                - dynamodb:Query
                - dynamodb:Scan
                - dynamodb:GetItem
                - dynamodb:PutItem
                - dynamodb:UpdateItem
                - dynamodb:DeleteItem
                - dynamodb:BatchGetItem
              Resource:
                - !Sub ${statusTable.Arn}
                - !Sub ${statusTable.Arn}/index/*
                - !Sub ${mentionTable.Arn}
                - !Sub ${mentionTable.Arn}/index/*
      Events:
        Root:
          Type: Api
          Properties:
            Path: /web-mentions
            Method: any
            RestApiId:
              Ref: ApiGatewayApi
        Sub:
          Type: Api
          Properties:
            Path: /web-mentions/{any+}
            Method: any
            RestApiId:
              Ref: ApiGatewayApi

A minor complaint with AWS SAM is that it's built in esbuild bundler for functions does not behave as expected, to get around this we manually bundle ourselves.

Luckily esbuild works pretty much on its own and doesn't require very much external config like webpack or other bundlers.

esbuild src/handler.ts \
  --target=es2020 \
  --platform=node \
  --external:aws-sdk \
  --sourcemap=linked \
  --outfile=.build/handler.js \
  --bundle

Express

Getting express to work on a Lambda is super easy once you know what you're doing⁴.

We've set up our API Gateway to forward requests at /web-metions or /web-mentions/{any+} through to our Lambda.

import express, { Express, NextFunction, Request, Response } from 'express';
import {
  APIGatewayEventRequestContext,
  APIGatewayProxyEvent,
} from 'aws-lambda';
import serverless from 'serverless-http';

export type RequestContext = Request & {
  context: APIGatewayEventRequestContext;
};

export const createApp = (): Express => {
  const app = express();
  app.use(express.urlencoded({ extended: true }));
  app.use(express.json());
  app.use(cors(corsOptions));
  app.options('*', cors(corsOptions));
  app.use((req: RequestContext, res: Response, next: NextFunction) => {
    console.log(`Request: ${req.method} ${req.originalUrl}`);
    next();
  });

  return app;
};

export const createHandler = (app: Express) =>
  serverless(app, {
    request(request: RequestContext, event: APIGatewayProxyEvent) {
      request.context = event.requestContext;
    },
  });

const app = createApp();
export const handler = createHandler(app);

Once we've configured the app we can now use express as we normally would!

app.post(
  '/web-mentions',
  async (request: RequestContext, response: Response) => {},
);

app.get(
  '/web-mentions/status/:id',
  async (request: RequestContext, response: Response) => {},
);

app.post(
  '/web-mentions/query',
  async (request: RequestContext, response: Response) => {},
);

Deploying

AWS SAM provides single command deploy, I've wrapped both build and deploy up into scripts in the projects package.json.

export REGION='ap-southeast-2' # Our target AWS region
export PROFILE='pfych-aws' # Our target AWS profile as configured in AWS-CLI
npm run build
npm run deploy:dev

This process takes a few minutes on first deploy but should be sub minute on any subsequent deploys as long as you don't change the SAM Template.

In closing

It was a fun challenge to implement Webmentions in a serverless manner. It wasn't extremely difficult since the spec is well-defined. I did skip over web mention updates & deletions, but I'll implement them at another time. Serverless definitely has warts & almost all of it is in its tooling & documentation. I hope that with time this can improve.

My next steps is to implement sending Webmentions from my CMS when I make new posts, but I haven't had too much time to work on personal projects recently.

The project is available on Github. It currently is deployed with the Serverless Framework but there is a pending PR to use AWS SAM instead since it has less 3rd party dependencies.

This article was written as a supplement to a small talk I gave in the office where I work about scripting. It sort of stands on its own but there will be references to internal practices and tools which might not make sense to some readers.

Preface

We constantly refer to "Bash" around the office in reference to our scripts but what actually is Bash and can we use it more effectively?

Bash is a Unix Shell, that runs inside a Terminal instance.

flowchart LR
    Computer --> Terminal
    Terminal --> Shell[Unix Shell]
    Shell --> Kernel["Kernel (Operating System)"]

Your Terminal is a piece of software on your computer that can interact with your selected Unix Shell. Some of the common terminal tools we see and use are:

• Terminal.app
• iTerm2
• The terminal in our IDEs

The Terminal program handles input, output and errors between you and your shell.

The most common shell is Bash which is shorthand for its full name, "Bourne Again shell". Bourne Again since it's intended to be used as a replacement for the Bourne Shell (sh)¹. Bash can run scripts written for the Bourne Shell but not vice versa

MacOS no longer ships with Bash as the default shell and instead uses ZSH. ZSH can run Bourne Shell scripts, but it cannot run Bash scripts.

So there are 3 main Unix Shells we interact with:

• Bourne Shell (often just called shell or sh)
• Bash (Technically GNU Bash)
• ZSH

So which shell do our scripts use?!

All scripts we use are prepended with the file extension .sh, however this file extension is not what is used to determine the shell!

Enter, the shebang!

#!/usr/bin/env bash

This line sits at the top of all of our scripts and tells our Terminal program what Unix Shell to run the script against.

Most Unix old heads will push for you to use either:

• #!/bin/sh
• #!/usr/bin/env sh

Since it's the most widely supported and come pre-installed on most unix/unix-like systems: MacOS, BSD, GNU Linux, etc.

MacOS & BSD do not ship with most GNU utils. MacOS ships with GNU Bash, but not anything else. causing confusion when using tools like sed or date which have BSD and GNU versions.

The shebang tells your Terminal what program to run the following file under. This can be anything as long as it's a valid path to a binary.

Syntax Examples

The following are some handy snippets that help explain the Syntax of Shell Scripts. Bash, ZSH and Shell all have extremely similar syntax. There are no types, and everything is considered "text".

Binaries on your system can pass data via stdin and stdout, stdin can be referenced in your scripts with numbered variables:

#!/usr/bin/env bash
echo "$1"

> ./example.sh "hello world"

hello world

a space is considered a separator, ensure you use quotes or escape the space if your input contains a space.

> ./example.sh hello world

hello 

All 3 shells offer operators for moving your binaries stdout into the stdin of another binary.

Pipe (`|`)

The Pipe (|) operator directly connects the output of one program to the input another

Arrow Brace (`>`)

The Arrow Brace (>) writes the output of a binary to file

> echo "hello" > ./world.txt
> cat ./world.txt

hello

Double Arrow Brace (`>>`)

The Double Arrow Brace (>>) appends the output of a binary to file

> echo "hello" > ./world.txt
> cat ./world.txt

hello
hello

Subshells `$()`

Subshells fork the current process and runs the inner content inside another process, it does not have access to any of its parents environment.
The stdout of the subshell is "returned".

MY_VAR=$(echo "Hello World")
echo "$MY_VAR"

Variable Expansion `${}`

This one's a bit complex, it allows operations to be run against a variable before the line is executed.

> MY_VAR=""
> echo ${MY_VAR:-"Missing!"}

Missing!

We use the :- operator here to substitute a default value if the left hand is missing. This is executed before the full line (with echo) is run.

There are a few operators:

• :- Substitution with a default value
• := Substitution with default assignment
• :+ Substitution for actual value
• :? Substitution with value check

These are explained in further detail in other guides².

Single Quotes (`''`)

This represents a plain string, you cannot use variables inside these.

> MY_VAR='HELLO WORLD'
> echo '$MY_VAR'

$MY_VAR

Double quotes (`""`)

Fancy pants string, you can use variables inside them.

> MY_VAR="HELLO WORLD"
> echo "$MY_VAR"

HELLO WORLD

You should always wrap variables in double quotes when using them as input to a binary.

> MY_VAR="HELLO WORLD"
> echo $MY_VAR

HELLO
Error: unknown command WORLD

Useful binaries & builtins

A lot of Unix-like systems ship with these tools, however be weary of slight syntactical differences between the BSD/GNU versions of these tools.

cat

cat is commonly used to output the contents of a file

> cat hello.txt

Hello world!
This is a second line
and a third line

grep

grep lets you easily search and filter stdout or files.

> cat hello.txt | grep "world"

Hello world!

> grep "world" -f hello.txt

Hello world!

You can also pass a Regular Expression to grep. Linux Grep & macOS Grep have minor differences in how they handle Regex, it's almost never an issue, but it's worth keeping in mind. (-o flag)

cat .zsh_history | grep -E '[A-z]([0-9]*)'

for

for allows for easy iteration over files or arrays.

for file in *.png; do
  echo "$file"
done;

Bash & ZSH have support for arrays:

declare -a ARRAY_VAR=("foo" "bar" "baz") 

for variable in "${ARRAY_VAR[@]}"; do
  echo "$variable"
done

if

if has different syntax than expected depending on your shell.

Bourne Shell (SH/BASH/ZSH)

You can pass flags to if to handle comparisons³.

A="FOO"
B="BAR"

if $A -eq $B; then
  echo "A is equal to B"
elif $A -ne $B; then
  echo "A is not equal to B"
done

Bash & ZSH

Bash 2.0.2 introduced the [[]] "extend test command" which allows for more standard syntax when doing comparisons.

A="FOO"
B="BAR"

if [[ "$A" == "$B" ]]; then
  echo "A is equal to B"
elif [[ "$A" != "$B" ]]; then
  echo "A is not equal to B"
done

You can also use the original flags inside the square brackets:

A="FOO"
B="BAR"

if [[ "$A" -eq "$B" ]]; then
  echo "A is equal to B"
elif [[ "$A" -ne "$B" ]]; then
  echo "A is not equal to B"
done

Sed

Sed allows you to edit input or a file with a regular expression. Sed also has differences across Unix-like systems⁴.

echo "Hello world" | sed 's/world/mars/g'

On BSD sed you need to supply -i when editing a file, this is not mandatory on GNU sed.

sed -i .bak -e "s/world/mars/g" hello.txt

JQ

JQ is not a built-in tool, but we use it often enough that it's worth mentioning. JQ allows you to parse JSON objects and offers almost all the utilities JS offers for working with Objects: find, map, reduce, filter etc. It's quite extensive and worth reading their docs⁵ about its capabilities.

I spent maybe 75% of my time there split between GiGO, LeisureLand and Hey in Akihabara playing games ranging from Dodonpachi & IIDX to games like Kancolle Arcade (I have a new addiction).

While some people might see this as a waste of a trip I really enjoyed it. Australia doesn't have very many good arcades and the good arcades lack a lot of popular games. It was amazing to be able to play without worrying about a queue of people behind you waiting to play since there were so many cabinets.

The Imperial Palace & Greenery

On my first day in Tokyo I decided to wander around the centre of the city and see what's around. I later learned that I had accidentally lined my trip up with Golden Week, which explained why everywhere was insanely busy. I wasn't able to get into the Palace, so I mostly took photos of the outside.

Day out with friends

I linked up with a friend who happened to be traveling around Japan at the same time I was in Tokyo. We went out to the Science Museum, Skytree and the Aquarium. While it was crazy busy at all the touristy destinations, it was really fun.

Asakusa

I had a day by myself and didn't feel like spending another whole day in Akihabara playing games. I caught the train up to Asakusa to see what was around.

Late nights in Kameido & Ikebukuro

My Hotel was in Kameido which was really close to Akihabara, so I'd always come back super late at night. I cant remember which of these photos are from Kameido and which ones are from Ikebukuro since I had loaded this camera with CineStill 800T specifically for taking photos at night, and I was relatively sporadic with when I'd decide to take photos.

I'd never heard about Ikebukuro but some friends I met up with live nearby and said the Game Centres in the area were chill so I spent a bit of time up there.

Kengo Kuma Trek

Before going to Japan I had picked up a random book at a book store about Kengo Kuma, it had pictures and plans of a lot of their works, and it was all super interesting, so I made it a goal of mine to try check out some buildings that were nearby.

Sadly the Gateway station was covered in Scaffolding, so I couldn't get a good look at the facade, but I could still appreciate the interesting roof from the inside.

I also walked to Sunny Hills bakery and bought some cakes. I was offered a free cake and tea with my purchase and had a great chat with staff about how I knew about their shop and what I thought of the cakes. They were amazing and I highly recommend them!

Harajuku

I didn't really plan to go to Harajuku or Shibuya on this trip, I don't speak Japanese, and I was traveling alone, so it didn't feel like it'd be the best way to experience these areas. However, I did want to visit Meiji Jingu, so I stopped in Harajuku on the way. There was a surprisingly large amount of graffiti in this area which was pretty surprising.

Meiji Jingu

I visited Meiji Jingu on one of my last days in Tokyo, it was raining a bit so the area was very quiet. I ended up sitting on a bench in the shrine for a while just enjoying the view. It was very cozy.

IIDX Retrospective

I've been playing IIDX for almost 3 years now and have played actively through 4 styles¹. Throughout this, I have tracked almost all of my scores through various means. I wanted to wait until I got Kaiden² before I did anything with this data but with Australia just getting its first Lightning Model³ & Me clearing 9th dan⁴ the weekend beforehand I thought it'd be nice to see some stats about my Pre Lightning Model play history.

All the styles

My first style was Heroic Verse, looking back now it wasn't the best style all around & the events were kinda mid; But it was my first. I have the AC⁵ poster framed up in my flat which I managed to pick up off another local player here in Sydney.

The overall trend leans towards assuming each style that comes out will lead me to play more. However, Resident is the first style where I've been confident enough to play level 11s & 12s⁶. If I continue grinding at this rate I could expect to potentially mind-block on certain charts and overall play less, or make the transition to playing BMS⁷ instead. Regardless, it will be interesting to see future data around this.

Plays between dan course clears

I have all my dan course clear dates listed on my website. Using those dates it's possible to extrapolate plays between clear dates. The data shown on this page is song plays not credits. Total money spent is marked further down...

As expected this number basically grows logarithmically & I expect it to continue to grow like this.

I am surprised how large the jump is to 7th is as well as the jump to 9th. Or More-so surprised by the lack of massive jumps between the other course clears. 7th Dan has "The Safari" which is notoriously tricky, and 9th Dan has the first 12 in it. I played 9th during Resident so the chart was Todestrieb; Which had chord streams I struggled to follow.

Using the power of "eyeballing" (I don't know my math too well); We can extrapolate plays to subsequent dan clears. I'm going to be assuming 10th, Chuuden & Kaiden will all be large jumps like the ones for 7th and 9th. With the Jump between 10th & Chuuden being slightly smaller.

I assume that 9th to 10th will be a larger jump since it's the first "All 12s" course. 10th to Chuuden and then to Kaiden will still be exponential jumps but not as large as the jump between 9th and 10th.

Regardless I'm very excited to look back and see how wrong I was!

Most played charts

This one is just for fun, there's not much to extrapolate here other than that I'm a scratchlet.

Money spent

For this section I'm going to make the assumption that a "play" is 3.5 charts. This takes into account Extra Stage as well as the few Premium free credits played in Resident. It's $2 AUD per play.

By Style

By Dan

This data is generated based on plays between clears. Meaning we can extrapolate the most expensive dan clears by seeing how many credits it took between each play to clear it.

As expected 7th & 9th dan are the most expensive so far.

Where from here?

I'm going to continue to track my scores as best as I can, there is very much some data missing from here. Especially if I played 1 song more than once during a credit & then didn't export my CSV right after. I stayed pretty on top of this but there's very much some data missing. I've played Black by X-Cross Fade way more than my data says I did.

It'll be extremely interesting to see if my predictions are correct. I'll definitely create another one of these when I get subsequent dan clears.

Thanks

There's a lot of people I've met either directly through IIDX or through rhythm games in general. Without IIDX I would not have met some of my closest friends. I think its cringe to do this sort of thanks at the end of this post, but I very rarely talk about my friends on my website. So without further ado:

To Purikura Sydney for even getting an IIDX cab & to Chris for convincing them to buy it well before I started to play, as well as for being extremely friendly.

To Isaac & Oswald for asking the random white dude watching them play if they'd like a go. Without their friendly attitude and first few pointers I never would have touched the game.

To Martin for being extremely helpful during my first few months playing, teaching me proper tech & for allowing the Australian IIDX community to play on his cab for free during events like Birthdays & BPL, we all very much appreciate the hospitality.

To Amal for helping me get into BMS & configure it correctly and to Lucas as well for helping me and others out with assorted BMS/IIDX tools & clients.

To the entire remaining Australian IIDX community & the English BMS community. They've all been extremely helpful & friendly over the last almost four years, and I'm extremely thankful to have gotten to know them.

Most importantly, to Zkldi & Aixee, Without their amazing tools, it would have been impossible to track my plays as thoroughly as I did. Especially to Zkldi since 99% of these scores were tracked via Tachi, which is an amazing tool I've watched grow since I started playing.

There are many more people I've met through the arcade in general who I'm glad to call my friend, I'm sure they know that I'm grateful for them without me mentioning them by name here.

A dream inside a dream, but who is the dreamer?

I wouldn't call myself a lucid dreamer, in fact I'm quite the opposite. My memory of late has been terrible, and I can hardly remember anything I dream about unless prompted. However, dream spaces, or at least mine; are constructed of things that exist within my reality. Perceiving these things can trigger déjà vu back to these spaces I've dreamt about in the past.

Schrödinger's space

These spaces can be split into two categories, subconscious spaces and conscious spaces. The most interesting ones are the spaces created by the subconscious. Spaces that, while present in them dreaming, are recognizable & recurring. But when awake, are instantly forgettable. With your conscious recalling them when primed by something that existed within them; A person, object, or item.

*Once perceived it's no longer a dream space.*

Once a subconscious space is perceived enough when awake, I'd consider the space to be different. It's now a space constructed in the conscious mind; No longer an amalgamation of separate thoughts & memories, but instead its own significant constructed place to be thrown into memory banks to be churned out into something new in a dream.

This space is no longer a sum of its parts, but a part itself. Something that future dream spaces can pick from, creating hybrids of places that seem similar but may not be.

flowchart LR
    Experiences --> PlaceA["Place A"]
    Experiences --> PlaceB["Place B"]
    Experiences --> PlaceC["Place C"]
    
    subgraph Memory
        PlaceA
        PlaceB
        PlaceC
        NewSpace[New Space]
    end

    subgraph Subconscious
        PlaceA --> DreamState[Dream Space]
        PlaceB --> DreamState
        PlaceC --> DreamState
        NewSpace[New Space] --> DreamState
    end

    subgraph Conscious
        DreamState -- Recall the dream space --> PerceivingMind[Perceiving Mind]
        PerceivingMind -- Recalling becomes memory --> NewSpace
    end

Once remembered, the place no longer exists as you perceived it.

This page is a little barren ...

I'm Pfych (she/her), a full time developer & hobbyist something or other in the Sydney region.

This site is full of personal projects, thoughts, photos and ideas that I've been working on for the past few years. I'm always looking for new things to do, so if you're interested in something, feel free to reach out!

• Professional Experience (A formal & boring)
• Games I'm into
• Music I'm into
• Recent projects I've worked on
  • GitHub
  • GitTea (Self-hosted, mostly private)
• Photo's I've taken

Part of my current job involves me working with Amazon Web Services infrastructure (basically fancy servers). By taking advantage of Amazons free tiers, it is possible to host your own static webpage entirely for free with very little programming know-how!

This guide will assume you know nothing about Amazon or about HTML. We'll start by building a simple HTML page, and then we'll deploy it to Amazon manually. This is all you really need! However, afterwards, I'll walk you through setting up Budgets in AWS Cost Management to avoid any bills over a dollar (just in case!).

How is it free?

Amazon Web Services (AWS) is behind an extremely large chunk of the internet as a whole and is one of Amazons main methods of revenue production, this allows Amazon to benefit from economies of scale. Basically, they have so many large clients that they are able to offer extremely generous free tiers to smaller users in the hopes that eventually they'll scale up and start paying, or they'll recommend AWS to their employer or company.

Since we're just making a personal website, this generous free tier will let us host whatever we want for very little!

Let's break down the potential costs really quick:

For storage, you get the following for free every month:

• 5 GB of data
• 20,000 requests
• 2,000 uploads
• 100 GB transfer limit

After going over 5 GB, you're charged $0.023 per GB over. For a simple personal site, you'll never go over this storage limit.

20,000 requests refers to any files your site uses, if you just have a single HTML page that means 20,000 people can view your page monthly. However, once you exceed 20,000 requests you're charged $0.0004 per 1000 extra requests (yes, that's the right amount of zeros!).

We can add caching to lower the amount of requests made, caching basically saves frequently accessed files by users to avoid people "double getting" a file. This makes the site feel faster to them and saves us money!

Amazons Caching service is free up to 1 TB and 10,000,000 requests. Once you exceed 10,000,000 requests, you pay $0.0075 per 10,000 requests after.

Later in this guide, we'll set up a Budget on Amazon to alert us via email if any of these values are about to go out of the free tier.

Making a basic HTML site

This won't be a guide on how HTML works, or a guide walking you through designing your own website. HTML can be super fun, and I really recommend using Mozilla's MDN for explanations on what everything does. Or if you're a complete beginner, check out HTML for People which is a great resource for non-technical people.

Using any text editor, preferably something like vscode if you're a beginner. Create a file called index.html inside its own folder on your computer somewhere safe.

Inside this file, paste the following:

<html lang="en-US">
	<head>
	    <meta charset="utf-8" />
	    <meta name="viewport" content="width=device-width" />
	    <title>My Personal Page!</title>
	</head>
	<body>
		<h1>This is a heading!</h1>
		<p>This is a paragraph</p>
	</body>
</html>

There's some super basic stuff in here, and hopefully you can sort of figure out just by looking at what all of these do (don't worry about the meta stuff for now!)

If you save this file, and then open your File explorer and double-click on it, it should open a super simple webpage in your browser of choice!

If you want to learn more about HTML, check out Mozilla's beginners guide, once you're comfortable with HTML, you can start making it look pretty and custom by using CSS, Mozilla also has a beginner guide for this available!

Creating an AWS account

Creating an AWS account is super quick!

1. Navigate to the AWS homepage
2. Click "Create an AWS Account" in the top right of the page
3. Enter your email address and pick an account name
4. Click verify email address and enter the code sent through to you
5. Pick a secure password! I highly recommend making this 16+ characters long with numbers and special characters. Make sure to write this down somewhere!
6. Once this is done, select "Personal account" and fill out your details
7. Enter your credit card details
8. Enter your mobile number again to receive a text confirmation code
9. Select the Basic Support free tier

You should now be re-prompted to log in.

1. Select "Root user" and enter your email address.
2. You'll then be prompted to enter your password.

Once this is done, you'll now be on the AWS Console. In the top right to the left of your name you should see a list of regions. Select the region closest to you.

Deploying your website manually

You could use a toolkit like Serverless to automatically build a CloudFormation and manage your website, however that is out of scope for this guide. We'll be deploying manually for now.

Deploying to S3

Navigate to S3 in the Amazon dashboard and click "Create Bucket". You can name the bucket whatever you want, but it must unique globally (meaning nobody else is using this name). Make sure the region is the one closest to you.

1. Uncheck "Block all public access" and then tick that you acknowledge the risks.
2. Leave everything as is and click "Create bucket"

Next, navigate to your bucket by clicking its name.

1. Click on Permissions
2. Scroll down to Bucket Policy
3. Click edit
4. Paste the following (and replace "your-bucket-name-here" with your buckets name):

{
    "Version": "2008-10-17",
    "Statement": [
        {
            "Sid": "PublicReadGetObject",
            "Effect": "Allow",
            "Principal": "*",
            "Action": "s3:GetObject",
            "Resource": "arn:aws:s3:::your-bucket-name-here/*"
        }
    ]
}

Once this is saved:

1. Navigate to "Objects"
2. Click upload.
3. Click add files
4. Select all your site's content.
5. Scroll down and click Upload.

Once the upload is complete, close the upload popup by clicking close in the top right.

1. Navigate to "Properties"
2. Scroll to the bottom.
3. Click "Edit" on static web hosting.
4. Enable static web hosting.
5. Inside the index document input type: index.html
   1. Add an error document if you've made one, this will be your 404 page!
6. Click save changes.

If you scroll down to the bottom again, you should now see a URL, this is your site however it is not caching, We'll set that up now.

Setting up CloudFront

From the Amazon dashboard home, navigate to the service called CloudFront. You may have to search for it using the top search bar.

1. Click "Create CloudFront distribution"
2. Select your S3 bucket from the origin domain dropdown.
3. Scroll down to settings
4. Set "Default root object" to index.html
5. Click "Create distribution"

This will take a few minutes to deploy. Once deployed, you can navigate to the CloudFront domain displayed in the distribution to view your site! If you make any changes to the site and want to clear the CloudFront cache, you can optionally do the following:

1. Navigate to "Invalidation"
2. Create new Invalidation
3. Type /*
4. Click create

CloudFront will automatically invalidate eventually, so this is optional.

Setting up an Amazon budget

To avoid any surprise bills, you can set up an Amazon Budget. Go back to the Amazon Dashboard and Navigate to "AWS Cost Explorer", you may need to search for this.

1. In the left sidebar, near the top, click Budgets.
2. Click "Create a budget"
3. Select "Use a Template" and then "Monthly Cost Budget"
4. Give this budget a name
5. Set the max you ever want to spend (I set this to a dollar).
6. Under email recipients, put any emails you want to be alerted if you're about to hit this budget.
7. Click "Create Budget".

You'll now receive an email at 85% of the budget and a 100% of the budget.

Domains

Amazon offers a service called Route 53 which can manage your DNS as well as buy a new or manage an existing domain directly. This service costs $0.50 a month per domain, but makes connecting URLs to Amazon services and managing SSL certificates crazy easy. I will not be going in depth on this since there are plenty of guides online.

In closing

You're all set up with a static webpage! Do note that this site is static, meaning that all the code and content is pre-written. You cannot use a service like WordPress or Ghost to manage blog posts or pages. If you wish to stay free, I recommend avoiding services like EC2 & Amplify, as they can quickly get out of hand if you're unsure what you're doing.

I use a command line tool called Pandoc to convert my posts into HTML, which I write and store as Markdown in Obsidian¹. I'll eventually release the scripts I use to automatically convert my Markdown notes to this site.

Eventually I plan on making a guide soon for how to host an API for free using AWS Lambda, I'm aware that using a service like GitHub Pages will also allow you to host a static site for free but by keeping everything under AWS it makes connecting domains, SSL and eventually linking up an API all extremely easy (and extremely cheap).

If you have any questions, please feel free to email me. However, please do some research on your own before reaching out! Mozilla's MDN is a great resource, and so is Stack Overflow if you have any generic questions.

Guides & Articles

• Nix is good actually?
• Splitting .mc2 memory cards
• Setting up a Brother HL-L2400DW on Arch
• Setting up gamemoderun correctly
• The Images Page
• Re-enabling Webmentions
• BlueSky Comments
• JSX Clientside - using JSX on a static site
• Using Cloudfront as a reverse proxy
• Control a HomeBridge Accessory with Moonraker
• 4K Desktop Streaming from a 1080p Wayland Host
• PipeWire & Steam Remote Play
• Packaging for Arch Linux
• OPL, SMB & Mode 1
• Setting up func_godot
• Create a Google sheet from a template
• Facebook Oauth in Cognito without Amplify CLI or UI
• Hosting a static site on AWS
• Bash talk notes
• Serverless Please
• SAA-C03 Digitised notes
• SAP-C02 Digitised notes
• Professional Experience & Work History

Mini Projects

• Mangadex Chapter Checker - Check if new chapters have released for tracked Manga
• Rockbox Scrobbler - Rockbox playback log to Lastfm
• Riichi Mahjong Scoring Calculator - Simple scoring calculator for Riichi Mahjong
• Experiments - Experiments done on my commute home from work
• Gitadora to Kamaitachi - Importing your Gitadora scores to Kamaitachi
• Sandbox - Web based programming environment
• Render Tables - Render tables from JS Data
• Webmention Receiver - Serverless Webmention receiver
• BMS Scope of Work - SOW for a BMS related project
• IIDX Beginners Guide - Beginners guide for IIDX, animated with CSS
• IIDX Retrospective - Retrospective of game progress, charts done with CSS

The Show

I now have a mad respect for show photographers, snapping pics of artists while they do their thing is hard af.

Train Graf

Before the show I had about 15 exposures left on a roll of UltraMax & I had zero idea how to get through it. While waiting for my train, freight came through. Did my best to shoot while it rolled through.

The Photos

Almost all of these came out good, I'll be including them in chronological order with only (non-cool) blurry messes removed. None of these images have been edited.

Ultramax (400 ISO)

I had a few shots left in the camera during the opening DJ set.

Porta (800 ISO)

Finally whipped out this film, I was super worried they'd all come out shit since I'd never shot with it before but I'm insanely happy with the outcome.

Thanks / Updates

Shout-outs to the group I was with before and after the show, your company (even if brief) was mad.