Good news! I made more progress than i thought i did!
So, I'm building this app with very loosely coupled- essentially independent- components that are tied together with a larger framework. Two of the earliest components I worked on were a thing that parses usertags, hashtags and links from a shortpost, and another that works as an image uploader, verifier and thumbnailer. And when I dug into infra stuff I was mid-way though marrying those components within the framework to create an actual functional microblogging service. Neato. Next up after that is threads and replies.
The other thing I was working on was a captcha. The easy way to do this is to generate a capcha and then assume its OK if ocrad fails to read it. This seems... inadequate. So I'm playing with the idea of doing an extra call to Textract to verify that the captcha is indeed very hard to read mechanically. But that is very, expensive. Both as an api call and financially.
What I'll probably end up doing is building the captcha images using some wierd old fonts I found on a 90's shareware disk. If it ends up being a problem I'll add utf8 codepoints. I'll have to build an emojii picker for that, but I'll eventually need one anyway. Absolute worst case scenario I'll switch to google captcha.
I'll also be working on user-account specific stuff like account visibility (public, protected, private), user and hashtag subscriptions, feeds, and content flagging.
Thanks again for stopping by. Please tell your friends about us and don't forget to DONATE for a beta invite.
~Sofie ❤️