Re: Last call comment process

> Or maybe you do... danc-01 might turn out to be more significant
> than I thought at first...

But Dan what should *I* do with this guy giving me grief over on rdf-comments!

He's picking holes in my exact choice of words - I am pretty sure he knows 
what they mean; well, at least good enough - but probe anything deep enough 
and you can hit some philosphical problems.

I really just don't want to have to think about it anymore, and get back to 
the serious stuff over in webont!

How do I get rid of him? How much am I allowed to offend him?

I think he goes by the name of connolly.


===

More seriously, I am glad these two are on the last call issue list.

So - this connolly fellow does not like my exact choice of words for defining 
RDF literals, nor the way I have defined graphs, and graph equality.

On the first one I think he could accept the words that we've got, although 
not very happily - what he really wants (untyped literals without a language 
really, really are Unicode strings) he can't have (Unicode strings in RDF 
Graphs are URIrefs !?).

My fear is that we could tweak the words one way or another, but it won't 
actually improve anything. The fact is that, particularly with XSD in the 
soup, there are interesting questions we can't answer. e.g. is "1"^^xsd:int 
the same as "0" when read as a sequence of integers (XML Schema user type 
derived by list). Thats what I learnt for my sins

0 = empty-set
n+1 = n union { n }

Any system that did this, would have a formally defensible position and not 
very many users.

I do think we've gone far enough this round of rdf core spec writing, in terms 
of getting the formalisms clear, and nailing everything down. connolly seems 
to want us to go further; there's probably a postponable issue lurking in 
here.

Jeremy

Received on Tuesday, 28 January 2003 16:38:03 UTC