[eluser]Jelmer[/eluser]
@pickupman
Regretably I still haven't tested the English version of it myself. Will probably have a German version tested before ever getting round to that (I'm working on a website with a Dutch and German frontend). I'd love to hear about it if you get to testing the English implementation.
My sense of it is that MATCH..AGAINST/fulltext search will work great on huge datasets and sites with many large articles. But for smaller websites(< 100 pages)/datasets (< 1000 items), the 50% threshold will be a huge problem.
LIKE will probably never be very efficient as I believe it doesn't use an index or anything (like fulltext search does). Which means it needs to do string comparissons on each cell in the columns you're testing against. That can only mean horrible performance, especially on bigger datasets.
Which, in the end, was how I got to the point of writing my own indexer. It's a lot weaker then MATCH..AGAINST but will work great for small to medium datasets (like a small-business website or something like that). I'm not sure if it's the best thing in this case though.
@metaprinter
I'd suggest only allowing productname search (using LIKE) and giving customers selection options within categories which can limit the SELECT output with simple WHERE statements. But if you want your search to cover the entire website and not just the products, then my library might indeed be an option for you.
