pcplantdb

[Chad Knepp <pyg@galatea.org>]

Richard Morris writes:
 > <snip>
 >
 > >  >       15 Re:The American view of the world
 > >  >       15 RE: Major histocompatibility complex (MHC)
 > >  >       15 Biogas digester plans
 > > 
 > > Wow interesting stats!
 > > 
 > Yep I think it helps to get a feel of the sort of questions
 > people ar interested in.
 > 
 > Simply generated. I've been saving permaculture posts in a folder
 > since 2001 and
 > 
 > grep "^Subject" permaculture\ digest | sed  "s/Re: //" | sed 
 > "s/\[permaculture\] //" | sed "s/Subject: //" | sort | uniq -c | sort 
 > -rn | more
 > 
 > gives the list. The unix commands grep, sed, sort, uniq, cut
 > give a great db query language.

I also have hopes/plans that HG will be manipulatable by
non-standard/custom clients.  Having a well defined API will allow all
kinds of roll-your-own end-user exploring/tinkering.

 > >  > With a bit of thought and a lot of discussion
 > >  > we can try to put these into some sort of graph structure.
 > >  > With each node as a specific topic.
 > > 
 > > Also think of the relationships themselves as atomic elements, like
 > > for example, the kind of relationship worms have with soil and the
 > > relationship worms have with compost/vermiculture.  The relationship
 > > itself is information, not just a link. 
 > 
 > Indeed, it would be possible to give relationships a class.
 > Theres several systems out there which does this.

Yep.  I've been thinking about the sorts of things you could do with
the meta-information created by just looking at the relationships and
not the elements themselves.... Hmmm, relationships of relationships.

 >  > I also think that things can
 > > have the relationship of being similar to other things (a set) such as
 > > animals or annual garden plants... sort of categories.  I think this
 > > relationship would be helpfull in generalizing some relationships such
 > > as chicken --> annual garden plants is largely the same as chicken -->
 > > cabbage and doesn't require just duplicating the same thing in chicken
 > > --> broccoli, chicken --> kale, etc.
 > 
 > First though was that you have really got
 >      chicken -> brassica
 > and the other links are really specialisations of the above.
 > Theres a very good case for data at the genus and family level.
 > 
 > However
 >      chicken --> annual garden plants
 > does not work so simpily as annual garden plants cross over many families.

Very true.  I thought about it after that a chicken --> tomato is very
different than chicken --> kale.  

 > > 
 > >  > After a lot of work organising we could probably come up with
 > >  > a network representing much of permaculture.
 > >  > A touch graph type interface could make it fairly quick to
 > >  > navigate to the particular topic of interest.
 > > 
 > > Speaking of TouchGraph, have you looked at it? 
 > 
 > Yep see
 > 
 > http://www.ibiblio.org/pfaf/pcplantdb/PlantFamilyTree/PlantFamilies.html
 > 
 > which is a touch graph map of the plant kingdom.

Excellent!  Nice proof of concept.  Everybody check this out if you
haven't already.  

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