It's weekend again, it's also time for me to make progress on updating my site. This time I've completed alot of things that's on my todo list, I would say this will be another major site update. However, it will be backend-wise, nothing much unless I do some cosmectic updates (which is also on my todo list @@)

Anyway, the biggest problem I encountered right now is building my own search engine that search through my weblog entries with ability to sort results by relevancy/date/topic/category. There are few easy ways to do this, one is using MySQL built-in "Full-Text Search" function, which also includes boolean full-text searches using IN BOOLEAN MODE modifier (eg using + - etc signs). Another way to do this is using a simple sql query with LIKE %string% clause.

Everything sounds so sweet and nice before I realize that my blog entries contains bbcode tags. The content of the weblog entries are rendered on-the-fly. So if you searched a keyword that's used for a bbcode tag, then the return result will not be accurate. MySQL also requires pre-indexing the data before you can use the full-text search, that means you'll need extra database storage for indexes and slower performance too.

So, I want to do a search on my rendered weblog entries instead of entries in raw format. That's possible, but that will probably be done with regex in gayness loops in PHP rather than MySQL (well, at least that's the way I can think of). If not then I'll have to stripped the bbcode tags and build my own indexing system. That's kinda redundant and I won't implement it with a complicated relevancy matching algo, so what's the point eh ?

I read some articles/guide/how-to but neither of them mentioned or have solution to the problem I mentioned. So I went to a few other people's weblog/forums to test this. The test results surprised me. Just try yourself in those "homemade" personal website/blogs and compare the ones that uses systems like phpbb/blogger.com. Use search query like "[b ]" or "<b>" and you'll know what I mean.

For now, I'll just see what I can come out with. If anybody has experience in this or knows great resources please let me know *scratching head... *

Some readings:
MySQL Full-Text Search Functions
Solano Systems - MySQL Relevancy Search