Re: [Gmod-ajax] Simile::Timeline

[Mitch S. <mit...@be...>]

Chris Mungall wrote:
> All that is required is a trivial amount of data exposure as RDF and  
> minimal coordination on use of URIs (much of what is already in  
> place). I think some work still needs done on the tools side of  
> things, but provision of the datasets will help complete the virtuous  
> circle.
>   
I'm interested in the URI coordination issue.  If we end up with (say) a 
wiki page per feature, how should the URI for the wiki page relate to 
the URI for the feature?  I'm tempted to have them be the same URI and 
serve different content using HTTP content negotiation.  Does anyone 
here have good or bad experience with this?

I understand that in some circles it's considered broken, but these 
people are trying to decide whether the server should send back HTML or 
xhtml, which is a fine distinction in the best of circumstances.  I'm 
hoping that accept: application/x-das-features+xml or accept: 
application/rdf+xml will be easy enough to distinguish from accept: 
application/html.  But my experience here is limited.  Are there any 
browsers that include both RDF and HTML in their accept: headers?

I'm inclined to say that the actual resource (e.g., feature) is the same 
in all those cases, and RDF vs. x-das-features+xml vs 
application/html+xml is just a different representation of the 
resource.  But I'd certainly like to hear any other opinions on this point.

I imagine that the URI coordination on the ontology side is pretty well 
implemented, but for feature URIs I'm not sure where things stand.  Does 
LSID have any relevance to the kinds of URI coordination we want to do?  
What's the state of implementation with LSID anyway?

Mitch