- From: John Arwe <johnarwe@us.ibm.com>
- Date: Sun, 30 Sep 2007 16:25:19 -0400
- To: public-sml@w3.org
- Message-ID: <OFFADE89AA.0BC5A7C2-ON85257366.006084E5-85257366.0070550D@us.ibm.com>
(from bug 5100 just opened) SML 4.1.1.2 Null Ref change "...not to search the reference for reference scheme information items." to "...not to attempt to recognize any reference schemes in it." I made that change in the bug because it pretty clearly preserves the original intent. The larger question is whether or not we think that original intent is correctly stated. It says not to examine null refs for ref scheme data, but later it says " The question of whether a null reference is resolved or not is undefined; it is an ill-formed question." Note that the quoted text talks specifically about resolution, not about parsing the reference element for ref scheme data. Those could, perhaps should, be viewed as separate sequential processes. In which case we need to assert whether or not it is an ill-formed question to ask "which ref schemes does a null ref contain?". I think it would be consistent with our discussions to date to assert that processing of null refs should be proscribed either at parsing for schemes or at actual resolution once a given scheme has been identified in the reference. Having flipped a coin, I'll propose we make the assertion that null references must not be parsed/matched for ref scheme compliance (or more generally, "must not attempt to recognize a null ref as being compliant w/ any ref scheme(s)"). Best Regards, John Street address: 2455 South Road, Poughkeepsie, NY USA 12601 Voice: 1+845-435-9470 Fax: 1+845-432-9787
Received on Sunday, 30 September 2007 20:25:38 UTC