pcplantdb

[Chad Knepp <pyg@galatea.org>]

Bear K writes:
 > Howdy all,
 > 
 > Don't worry, I'm still around. :-)
 > A couple of things on the front-end (web) front...
 > Chad, would it be possible to wrap the plant descriptions in a few more 
 > tags? It would be nice to be able to format it a bit more. I suppose I 
 > should write a mockup for you. For the results something like:
 > <div id="results">
 > <div id="found">Found 1020 matches for 'food'</div>
 > <ul class="plant">
 > <li class="family"><a 
 > href="/cgi-bin/eden?family=Compositae">Compositae</a></li>
 > <li class="binomial"><a href="/cgi-bin/eden?plant=6731">Taraxacum 
 > officinale </a></li>
 > <li class="common">Dandelion</li>
 > <li class="description">
 > <p>Description:  A deciduous perennial that grows to 1.0 meters (3.3 
 > feet) high by 0.5 meters (1.65 feet) [snip]</p>
 > </li>
 > </ul>
 > <ul class="plant">
 > ....
 > </ul>
 > </div>

Totally.  This was what I was thinking myself and pretty much exactly
how I would have done it.  The next release will have this.

I also agree with what John S. said about this becoming a UI issue and
we need to have a lot more discussion about it.  I'll repeat that I'm
pretty open to implementing anything I can (within reason) that folks
suggest.  I'm not really sure what questions to ask in this discussion
so I'll leave it to others to bring it up.  

 > It would really allow a lot more flexibility on how things get 
 > displayed.
 > Something similar for the plant description. As Richard mentioned, 
 > maybe when we're up to it, output as XML, and run it through XSLT to 
 > get the markup above.
 > 
 > Also, on tags like the plant uses, should we perhaps use a 
 > (tooltip-like) popup to hold the description, instead of a link to a 
 > new page? There are some pure CSS methods, and a few javascript methods 
 > as well to accomplish this. 

Another idea I was thinking as well.  <acronym/> for css I
think... this will of course get us into browser support.  Not all
browsers support css or javascript.  What's our bottom line.  

 > Could we also use anchors for the references [13]? Maybe there's a
 > regex-pression that can find brackets (find left bracket, replace
 > with [<a href="#, add [0-9] up to right bracket, add ">", add [0-9]
 > up to right bracket, replace right bracket with </a>] )...thats
 > using 2 instances of the text between the brackets, not sure if
 > regex can do that...  I'm assuming we need that because the
 > references are part of a SQL text field, and not dynamically
 > generated.

This is a fascinating point of discussion.  The references [0-9] are
actually part of the text authored by Ken Fern and as plain text have
no markup.  As you can imagine, this is a problem.  These fields in
particular are the basis for me wanting to XML'ize the dataset;
effectively giving them markup.  The only reason the references
actually show up in the description page is that long ago I wrote a
perl script that went through all the text in the dataset and created
a references to plant index table.  This is, of course, lame (static,
etc.).  I would much rather have some kind of markup like:

   edges of woods[21], alluvial thickets and gravels[43].

To

   edges of woods[<a link="reference=21">21</a>], alluvial thickets
   and gravels[<a link="reference=43">43</a>].

This can be translated by Eden to whatever the client wants or just
left raw and allow the client to XML -> whatever it wants.  This is
actually what happens internally (except for references aren't marked
up yet in the dataset) and is ouput in text or html depending on the
client.  You can see this in the text client by specifying links="on"
or links="off".  The html (cgi) version always generates html <a
href/> type links.

 > In other news, I've looked at TouchGraph some more, and it does look 
 > very nice. I'm not a Java programmer, I've tried to import the source 
 > into Apple's Project Builder, but I can't seem to get the directory 
 > structure set up right to compile. Java seems readable, and I _may_ be 
 > able to tweak things well enough to be useful, if I can get this thing 
 > set up correctly. I wonder what relationship data we can draw from the 
 > database as it's currently set up.
 > 
 > Lastly, I got a response from ibiblio, they say I should log in using 
 > "pyg" and "change2me", doesn't seem quite right...

Sending you something offline...

 > Cheers from the rainy coast,
 > Bear

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