pcplantdb

[Bear Kaufmann <bear@ursine-design.com>]

Hi all,

Those are some major issues.

Is this an artifact of the substring searching algorithm?
Can a better scoring method be implemented?
Latin name match: high
Common name: high
Attributes: mid
String search: lower

We might also want to consider a word index table. Split all the text 
into words (on space, period, hyphen, etc), insert into the table, and 
increment a word count field each time (counts could be useful). That 
table might also be useful in fine tuning search results. We could store 
search terms, find what's not matching etc.
I know PHP offers a method to test the similarity of 2 strings, and also 
convert a string to a simplified phonetic representation. The phonetic 
could help find misspelled words, I'm also wondering if MySQL offers a 
similarity search to use in 'similar word' suggestions.

Regarding the page appearance, I really haven't looked at dev a whole 
lot in terms of design.
If I could get a templating language to work in that'd be nice. Either 
some PHP/Python crossover, a TAL templating system (native Python), or 
maybe Django. A little more Model/View/Controller like (Code <---> Content).
If Chad can process the variables for me, and pass them to the template, 
I can handle layout/HTML, and it's a lot nicer then writing HTML inside 
of Python.

First impressions are important. So maybe we should wait to wrap those 
issues up.

Cheers,
Bear