April 4

Hi everyone! My allergies have been awful this week so I haven't been able to do as much as I'd like, but still plenty of stuff to talk about.

ArtCentral (name might change later)

This project is new. Given Gumroad's recent pro-NFT stance, and a suggestion from an artist that a website for artists to display commission info and so forth isn't really around right now, I've decided to take on the mammoth project of making a full 'artist website', so to speak.

Theoretically, it shouldn't be massively difficult per se; there's an absolute ton of stuff to do, but most of it I can reasonably see how to do. The main irritation will be doing the frontend, not so much because it's difficult but because I find frontend work rather... frustrating. It's probably the one aspect of software development where standards - outside of e.g. HTML5 and CSS - are completely lacking, and everybody has their own half-baked hacked together solution for every problem (and seriously, can we consign jQuery to a trashcan pls). For more functional-oriented sites like ArtRetweeter, I can get away with a basic frontend, but for ArtCentral I'll need it to be both reasonably nice and also responsive to mobile users. I've done Android development before so it's plausible an app could happen one day, but that's thinking way too far ahead.

For now, I'm going to start with a kind of "commission and artwork portfolio website" with messaging, then expand to include a way for people to buy digital artwork. I'm planning for the latter not to be a file host, but simply a place where a user buys some pack, and then the website directs you to the link where that pack can be found.

Doing something like Gumroad, where the files are hosted directly, adds a lot of bandwidth costs and technical challenges - and I realised that many artists are already hosting their art packs on other sites, since for Patreon purposes they need somewhere for patrons to be able to download their rewards. Much cheaper for artists not to make ArtCentral another file host (and also easier to link everybody to one link than have a whole bunch of separate uploads), and it also means I avoid the problem of figuring out how to store many terabytes of data across multiple servers, figuring out how to connect servers to appropriate links, and so forth.

Technically speaking, there is zero security benefit to having a dedicated file host where e.g. only "registered users who purchased the pack" can download it. Once someone has downloaded it, they could re-upload it anywhere else, no matter what you do - either you allow other people to buy the file or you don't, pretty much.

I expect that this project will take many months, perhaps over a year, to get to a usable state. Due to its nature - and especially if purchases become available on it - I want to ensure that it is done properly, which means writing some decent test coverage from the outset and a bunch of other stuff. I don't yet have a Trello page for the todo list - I need to break it down into smaller steps first, as there's simply too many right now.