I've decided that there's no immediate need to refactor the contest system as I had initially planned. I used to think the inefficiency of the database logic caused the poor performance of the site, and that it was fast sometimes because of caching. The actual truth is that all the slowness is due to occasional extremely high network latency, so nothing I do can guarantee that the site never takes 10 seconds to load (except moving to a dedicated server, which is actually too expensive).

I don't see any major issues with the Judge and I think it has all the piecen an online judge would be expected to have, with one exception --- support for non-admin problem setters and contest setters (as in SPOJ, which I consider the "industry standard", although we did have a comment system before they did, I think).

How many [non-admins] would like to become problem setters and contest setters, and how often do you see yourself adding problems?

After I add this functionality (if there is sufficient interest), I plan to basically stop active development on the Judge (barring some major breakthrough such as a significant expansion of the user base).

I've decided that there's no immediate need to refactor the contest system as I had initially planned. I used to think the inefficiency of the database logic caused the poor performance of the site, and that it was fast sometimes because of caching. The actual truth is that all the slowness is due to occasional extremely high network latency, so nothing I do can guarantee that the site never takes 10 seconds to load (except moving to a dedicated server, which is actually too expensive).

I don't see any major issues with the Judge and I think it has all the piecen an online judge would be expected to have, with one exception --- support for non-admin problem setters and contest setters (as in SPOJ, which I consider the "industry standard", although we did have a comment system before they did, I think).

How many [non-admins] would like to become problem setters and contest setters, and how often do you see yourself adding problems?

After I add this functionality (if there is sufficient interest), I plan to basically stop active development on the Judge (barring some major breakthrough such as a significant expansion of the user base).