Re: rel=type or rel=profile, issue 92

> Excellent, then unless else someone comes forward with a formal 
> objection to using link rel=type to signify ldp resource types, 
> we're decided (since Henry was formally objecting on the other side 
> of this issue).

Sandro, I'm not sure that we know that.  But easy enough to test.

When we originally introduced rel=type it was output-only.  This hid the 
apparently different ways that people intended to use it ... as purely a 
data type, or as an object type (by which I mean both RDF-type and 
interaction model).  It was only recently that this bifurcation of thought 
was articulated within the WG.  Having corresponded some with Mark Baker 
in the service of addressing his LC comments (one of which specifically 
appeared to decry exposing the type via a header because of the potential 
for clients to make assumptions about the interface ... what I think we 
mean by interaction model), he seems to have caught this distinction 
during his review.

FWIW to others, since by Henry's own statements this won't matter to him: 
I contacted the type=rel RFC author last week to get a more pointed read 
of his intent, since the text (all prose) certainly can be read different 
ways.  His response was that his intent was it meant (paraphrasing) 
something closer to rdf:type (than to that plus interaction model).  He's 
also the author of the Prefer header RFC, and I sent him a courtesy copy 
of the draft registration so he can alert us early if we're going off the 
rails in his opinion.


Best Regards, John

Voice US 845-435-9470  BluePages
Tivoli OSLC Lead - Show me the Scenario