wac

Yet another place I hadn’t seen yet. 🙂 This is a bit of a novel given what’s been said so far, so bear with me. I’m the other developer besides error404 on WordPress-Pg. My work is continuing on the WordPress-Pg 1.5 port at this time. It’s been a fair bit of work to say the least and the progress is lots of small victories.

The talk of going to some sort of large generic object persistance layer or database abstraction tool is most likely not going to work out in the long run. Such things tend to bloat and slow down your program. Given WordPress’s necessarily interactive nature such slowdowns would be bad. I imagine many are thinking like Dr. Venkman, “I’m fuzzy on the whole good/bad thing. What do you mean ‘bad’?” I’ve already observed some performance issues with the number of queries and the load generated on modest servers, to add more would really grind everything down to a halt. It doesn’t help that PHP isn’t exactly the fastest language on the block.

That said, I think there is room for database abstraction, if for no other reason so the code streams can come together at some point to satisfy many users like fmayhar (and myself) who think that vendor lock-in is bad. As an added benefit when things like “Getting the most recent N posts and the number of approved comments for each” can become a function call that looks at some global “what database are we using” setting, they can be optimized on certain databases that only need one roundtrip to get that particular bunch of information (subqueries are good) as opposed to N+1 query roundtrips. I certainly plan on coding in that direction after the 1.5 port is more or less together, although I’m uncertain how receptive the existing WordPress maintainers will be. Maybe they’ll take my patches if I send them some beer… 😉