I’ve decided to start working on Priblog again. Basically I’m getting sick of how utterly slow Pyblosxom becomes under load. I still like Pyblosxom more than I like solutions such as MovableType or WordPress. But it’s not optimal when you’re running your server on a 200MHz K6.
So with Priblog what I’m doing is a compromise. It still is all based on text files sitting on the hard disk, but each file is an XML file. I’ll eventually write a simple converter or something like that too. Anyway, the major difference is that Priblog will cache most entries as files. This means that instead of having entires magically appear after entering them, you’ll need to run a little script to regenerate some files.
Hitting most of the pages will still be a CGI type interaction, as it doesn’t warrant storing all the possible combinations of files. However, notably the RSS file will now be cached. This is a HUGE improvement over the current PyBlosxom setup which needs to crawl the directory tree in order to find the files to make the RSS. It’s one of the major reasons why my RSS feed is so slow.
Another thing I’ve wanted is the ability to put an entry in multiple categories. It seems like I’ll be able to do that with little work. A change from my original idea is thta Priblog no longer requires pridb as a database. This should free up resources on low end computers. We’ll see how much hacking I get done on this. Hopefully airport time later this week will prove useful to me.