- From: Sandro Hawke <sandro@w3.org>
- Date: Thu, 28 Aug 2008 09:28:55 -0400
- To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
- cc: bparsia@cs.man.ac.uk, public-owl-wg@w3.org
> This sounds like a good idea to me. Me too. BTW I raised issue-109 on a w3c staff list yesterday and it's turned into a huge discussion. Not a simple topic. :-( Hopefully we'll end up with some useful information to share. - Sandro > peter > > > > From: Bijan Parsia <bparsia@cs.man.ac.uk> > Subject: XML Syntax: Attributes > Date: Thu, 28 Aug 2008 13:36:16 +0100 > > > > > I propose that we put the attributes into "no namespace" rather than > > into whatever namespace we pick. > > > > Rationale: Unprefixed attributes do not inherit the default > > namespace. This is a deliberate design decision of XML namespaces. The > > idea is that bare attribute names which are *only* used in a specific > > vocabulary are sufficiently disambiguated by their parent > > element. Attributes that are intended to *cross* vocabularies (such as > > xml:lang or xlink attributes) *are* encouraged to be put into a > > namespace (for obvious reasons). > > > > The advantage of keeping the attributes namespaceless is that, with a > > default namespace declaration, OWL/XML will not need *any* prefixes for > > elements and attributes. Frankly, this is a huge authoring and reading > > win. It also means one can *just* use QNames for uri abbreviation > > (assuming we allow QNames in attribute content...subject of a future > > post!). > > > > My understanding is that no namespace attributes are the preferred > > design in this case anyway. We don't intend for these attributes to be > > used anywhere but in our vocabulary and it's pretty clear that no one > > would want them :) > > > > Cheers, > > Bijan. > > > >
Received on Thursday, 28 August 2008 13:30:27 UTC