I agree, i should take some time to fix it. And i do actually need some help with that.
Things i want to change:
)And maybe some more stuff.
And better authorship/management for uploaded styles (that’s a bummer for me).
It’s a site with some interesting but specific challenges. I’m a web developer so I can give a hand. How about moving this to its own topic?
Yes, that one of the other points i want to address. And yes, maybe it should be it’s own topic 
I am absolutely ready to help out. I can probably devote at least a few hours/week if needed.
How big of an overhaul are we considering?
Should this be considered a complete re-write from scratch? What type of requirements would we envision needing?
Is this intended as a straightforward repository of styles? If so, how deep should the social/community aspect be?
Please feel free to edit/add more - these are just brainstorming thoughts off the top of my head.
I obviously like the github idea 
it’d elimmimate the need for user accounts and uploads. Static site!
The only reason i can imagine so far for not to do it all on github was that it might exclude the less techinal users from contributing styles. (but then again… they are mostly linux users anyhow)
Other thing that might be missing with git as main upload thingy is extra metadata.
What do you mean “extra metadata?”
Metadata is all other data which is not the style itself (or isn’t include). eg. Author, Licence, Author contact information, Revision of the style, Changelog of what has been changed etc.
I’d drive all that information from a json file or something similar.
But how does it get into the json file ? Are we letting the uploaders fabricate their own json file ? Wich then has to be validated, but what when it’s wrong ? At that point it get’s more userfriendly when there is a webform wich validates the data.
By then, git is still a valid backend, but so is a regular filesystem