- From: Dan Connolly <connolly@w3.org>
- Date: 22 Apr 2002 17:27:19 -0500
- To: Pat Hayes <phayes@ai.uwf.edu>
- Cc: w3c-rdfcore-wg@w3.org
On Mon, 2002-04-22 at 16:58, Pat Hayes wrote: [...] > In practice, this assumption may need be modified somewhat. [...] > OK ?? no. > This would be the only mention of 'unasserted' triples in the > document, and the whole issue of what counts as an unasserted or dark > triple would be relegated to some other domain of consideration, > which might be called the operational deployment of RDF in some > larger context. i.e. it's a "hook". My engineering experience says: don't put hooks in until you have tested at least two ways to exploit the hook and be sure it's of the right size and shape. W3C process calls for implementation of whatever features are in a spec before granting Proposed Recommendation status. We have (or at least: I have) become very aware of the trick where, when groups can't come to consensus, they put in a hook so that everybody can be happy; but then they never test the hook, and it just becomes an interoperability nightmare, with various implementors ascribing various meaning to it. If the hook stays in, I'll be sure to look in our request for Proposed Rec status for evidence of interoperable implementations. I'd really rather just take it out until we're more clear about how to use the hook. -- Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Monday, 22 April 2002 18:27:15 UTC