pcplantdb

[Chad Knepp <pyg@galatea.org>]

Lawrence F. London, Jr. writes:
 > 
 > Tried another eleagnus search, plain, with ", with ' - all returned the 
 > same results
 > 
 > Did search on eleagnus + fruitlandi
 > and got:
 > http://permaculture.info/cgi-bin/eden?search=eleagnus+%2B+fruitlandi&button=Search+PIW
 >   Found 239 matches for 'eleagnus fruitlandi'
 > Podocarpaceae Prumnopitys andina Plum-fruited Yew
 > Podocarpaceae Podocarpus salignus Willowleaf podocarp
 > Rutaceae Citrofortunella microcarpa Calamondin orange
 > Cactaceae Opuntia ficus-indica Prickly pear
 > Vitaceae Vitis berlandieri Spanish grape
 > Compositae Anacyclus pyrethrum Mount Atlas daisy
 > ... and so forth with no mention of any plant with eleagnus in its name 
 > until the next page
 > which contained:
 > Elaeagnaceae Elaeagnus multiflora  Goumi
 > --- no more for several pages until  Elaeagnaceae Elaeagnus multiflora  
 > Goumi reappears
 > --- then no more for a few pages until eleagnus begins poipping up with a 
 > little more frequence;
 > but nothing like the results you get for eleagnus alone.

Well, I can think of a couple of reasons.  Elaeagnus has an 'a' in it
that I forget a lot myself.  Eden does accommodate 'alternative'
spelling but gives better results with accurate input.  I also don't
think 'fruitlandi' is in the current dataset.  Hmmm... something to
add?  

Not really sure why those particular plants score so highly.  I think
it would be good to have an optional display of what parts a
particular search matched for analyzing query performance.  I'm
already missing 'scores' to see if these are really that much higher
than the various Elaeagnaceae I would expect it to match.  Maybe
everything is score 1 which is a very low level of certainty.

-- 
Chad Knepp
python -c 'import base64;print base64.decodestring("cHlnQGdhbGF0ZWEub3Jn")'