On Jan 5, 2006, at 9:08 AM, Dan Connolly wrote: > On Tue, 2006-01-03 at 11:32 -0600, Pat Hayes wrote: >> [...] So, here's an >> alternative proposal. We don't get rid of them by >> tweaking the definition: we leave the definition >> naive, think of it as a necessary condition on >> answer bindings, and we add a remark about answer >> sets, that servers are not obliged to deliver >> 'redundant' answers. > > Sounds familiar. That's the "redundancy optional" > design option that I suggested back on 4 Oct. > http://www.w3.org/2001/sw/DataAccess/issues#rdfSemantics > > I'm not sure why it's more attractive to others > now than it was then, but I still like it. [snip] Thinking more about this (and discussing with Kendall), I guess we won't lie in the road on this. I would like to it be marked that nailing down what redundancy is allowed (and when) is an important thing to have done and a strong contraint on a future group (i.e., I think this is a hole in the spec, but a livable one). Cheers, Bijan.Received on Thursday, 5 January 2006 15:34:48 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 1 October 2009 14:42:04 GMT