Wow things in Tabulas are starting to stabilize a bit; the bugs are getting spottier and spottier ... I feel that the development is starting to plateau a bit ... and with all the recoding I did with some of the older stuff, Tabulas is looking like a very solid product.

Of course, one thing really troubles me; there was a few users who would report that they could 'access' other people's accounts randomly. I'm not sure how this would work since each user is generated a unique md5 hash that is stored in a cookie and that links to their userid; the only way they could access other accounts would be if the site generated two similiar session hashes (but there is code preventing that). No one's reported it for a while ... but I hacked up a bunch of extra "tracking" features that I've been using to see how that's happening.

I spent a lot of time reworking the communities features. It's starting to come together; a lot of the functionality is there, but the UI is really messy and it's throw together haphazardly. Hopefully as soon as I'm done with all the community features (soon) I'll be able to fix up the UI of the control panel.

I added RSS feeds to all communities ... I'm really hoping I can introduce Tabulas members to the beauty of RSS. I really think RSS is lovely and so convenient. Hopefully my love for RSS will be pushed onto others ;D.

The next "big" problem (besides the styling issues that plague Tabulas) is tackling the RSS parser. In my haste to get features done, I didn't write a proper PHP RSS parser. Each time I generate a RSS feed for a user, I hit the database 10 times. For each community, 5 times. Since a lot of entries are posted to multiple communities, you can see why this is a problem (each entry posted to 5 communities means that the RSS generator hits the database 35 times).

Ideally the RSS generator would throw the whole RSS feed into memory, pop up the oldest entry, and hit the database once and add that lovely entry in and pop the RSS file back out to the hard drive.

Well, right now I'm just trying to get it work. As long as I'm aware of the problems and I know how to fix them, I should be ok. Since I had implemented a bunch of features that reduce the work on the database from the display end of Tabulas (e.g. each user's Tabulas), the database can handle all those queries fine ...

There is also now a separate page for each user that lists all entries posted to communities they are a member of, kind of like your friends page (you can access yours off of /~username/mycommunity.html). Originally I was thinking of having all communities come out on your friends page (ala Livejournal), but ultimately decided that was a bad idea. To me, communities represent interests ... while your friends page should be restricted to keeping up with your personal friends. Since many Tabulas members seem to be in a lot of communities, an information overload was probably going to happen for a lot of journals.