September 20

Hi everyone! All ArtCentral updates again, just logging what I've been doing. Feel free to ask if there's something you're curious about or anything ^^

Site appearance customisation

A feature for subscribed users will be to change the site theme/colours in a highly customisable way, so I've been refactoring the CSS to use variables. That'll make it a lot easier to implement this feature later, since everything will be using the same variable for a given colour and so forth.

Overlays: Hiding NSFW galleries and showing loading screens

I've implemented a simple div structure to do this. For a bit of fun see here: https://twitter.com/antsstyle/status/1567825097783746561. It's both for hiding individual artworks, in galleries with only a few NSFW pieces, and for hiding whole galleries behind an NSFW prompt for more NSFW-oriented artists.

Making nice "spinning icon" loading overlays was more or less the same task, so that should be easily doable on most pages where it is needed now. I've done it for galleries, for other pages I'll do it on a case by case basis depending on how long they're likely to take to load.

User profile banners

I've modified the user profile page to include banners, but I haven't yet integrated a way of uploading custom ones or cropping them to size, that's still on the todo list.

Two factor authentication

Two factor authentication is done. I thought it would be harder to implement than it was, but it's all nice and set up now.

JPEG compression tests

As part of planning the costs of the website, I've been experimenting with different levels of JPEG compression to determine what's acceptable. Artstation uses 90%, other sites don't say what they use and I can't be bothered to try and guess that, so I've been doing some mass testing in the cloud using the same resizing functions I made to do the resizing for the site.

The reason for this is that financially, CDN costs are by far the biggest cost of a website like this, and images are almost all of that. Reducing file size by say, 50%, almost halves website costs. As a result striking a balance between compression and quality is critical, even for an art website.

From my tests, I've concluded that 80% is fine. It's very hard to tell the difference between 80 and 90% compression in most cases, and even then it's only visible if you are really looking for it and examining it up close. The file size increase from 80 -> 90 compression quality is around 45%, which is a BIG deal especially for a smaller website. If the site got huge, CDN costs get a lot lower per GB due to economies of scale and then the compression quality could change, but for now 80% is honestly very reasonable (and without being told it's 80%, nobody would ever know for the most part).

Financial planning

While nothing is decided yet, I've been making some spreadsheets to try and simulate the costs of the website as user numbers increase. The main concern is figuring out what subscription price the website would need, and what percentage of users would need to be subscribed, in order for the website to be sustainable.

As you can probably imagine, this is incredibly complex to try and estimate. Users don't all follow the same number of artists or view the same number of images/galleries each month, each artwork is a different filesize... the list of variables goes on. I've got some fairly solid estimates to work with - at least as solid as estimates can really be in this scenario, I think - but I still need to do more analysis.

For now that's still a work in progress, but I know that the percentage of subscribers is never going to be high. If I can consistently get 1% (or if I'm lucky, 2%) of users to subscribe, that'd be good, but I know that's a high bar to reach even though it might not initially sound like one.