[Treebase-devel] query predicate vocabulary

[Hilmar L. <hl...@ne...>]

(Sorry for cross-posting - this does concern both treebase and phylows  
groups though)

Rutger et al,

I assume that the TreeBASE terms and predicate vocabulary terms  
spreadsheet [1] is the basis for the treebase.owl [2] ontology in the  
TreeBASE codebase? If so, I was happy to see the tb. prefix removed  
from the property labels in the OWL file. If not, what is the relation  
between the two? (BTW neither vocabulary artifact is linked from the  
TreeBASE API documentation [3] - shouldn't they, or at least one of  
them, possibly the OWL file?)

One of the outcomes of the Phylogenetics Standards working meeting at  
the 2010 TDWG conference [4] is that we need to move forward on a  
standard PhyloWS query predicate vocabulary. Obviously, the one  
created for TreeBASE would be a template for that, and ultimately  
TreeBASE could import that vocabulary, and add its own custom  
predicates (similarly as it now imports Dublin Core and then adds its  
own predicates). I think this approach (standard ontology that  
individual data providers import into theirs and then add to) would  
also allow data providers to add annotations indicating which  
predicates they actually support.

Are there concerns or considerations that could make this a bad idea?

Assuming for a moment that it's not a bad idea, here are some initial  
thoughts I had when looking at the treebase.owl file as the presumed  
starting point.
1) The pattern for constructing the label uses a dot to delimit  
"words" (e.g., identifier.tree), whereas normally the pattern I've  
seen uses CamelCase (which would yield treeIdentifier). For the  
standard vocabulary I'd rather stick with common conventions, so what  
were the reasons or examples that motivated the dot pattern?
2) None of the properties seems to have a definition. I think in the  
standard one we should aim for all properties to have good  
definitions. Do others agree that this is worthwhile, or do you think  
this wouldn't gain anything. (There are initial definitions in the  
spreadsheet, for example.)
3) The spreadsheet has additional information that looks actually  
useful, such as the Xpath expression for NeXML. Wouldn't that be worth  
retaining to?

Nico - I don't know whether you're subscribed to the phylows group -  
if not, you may want to if you want to stay in the loop on this.

	-hilmar

[1] https://spreadsheets.google.com/pub?key=0Av8UW3JvZsgcckwtLU83cHloUjhGY25uRzUtb2ZBbHc&hl=en&single=true&gid=0&output=html
[2] http://treebase.svn.sourceforge.net/viewvc/treebase/trunk/treebase-core/src/main/resources/treebase.owl
[3] https://sourceforge.net/apps/mediawiki/treebase/index.php?title=API
[4] http://wiki.tdwg.org/twiki/bin/view/Phylogenetics/WorkingMeeting2010#Further_develop_thePhyloreferenc
-- 
===========================================================
: Hilmar Lapp  -:- Durham, NC -:- informatics.nescent.org :
===========================================================