- From: Graham Klyne <gk@ninebynine.org>
- Date: Fri, 03 Oct 2003 16:44:33 +0100
- To: Francois Yergeau <FYergeau@alis.com>
- Cc: w3c-i18n-ig@w3.org, w3c-rdfcore-wg@w3.org
At 09:41 03/10/03 -0400, Francois Yergeau wrote: >Graham Klyne wrote: > > I disagree with this... I don't think we should be > > prescribing what to do > > with bad inputs ... it's up to an implementation how it > > wishes to deal with error situations. > >Like XML says you MUST crash and burn if you encounter input that's not >well-formed? Well, yes, if you refer to XML's definition of "Fatal Error". I don't particularly agree with the way XML spec tells applications what to do here. (I'd say it's primary role is to define what constitutes valid XML, maybe define certain kinds of conformance, say what documents may be considered equivalent, and maybe point out the problems of trying to process documents that don't meet some basic conformance requirements.) But in this case, if NFC is a SHOULD, I don't think it can be regarded as any worse than what XML calls an error, from which an application is permitted (not required) to recover. I could support something like a "security considerations" note that points out the potential dangers of accepting non-NFC text. #g ------------ Graham Klyne GK@NineByNine.org
Received on Friday, 3 October 2003 13:30:11 UTC