- From: Graham Klyne <gk@ninebynine.org>
- Date: Thu, 02 Oct 2003 20:10:49 +0100
- To: Francois Yergeau <FYergeau@alis.com>, "'Brian McBride'" <bwm@hplb.hpl.hp.com>, Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Cc: w3c-i18n-ig@w3.org, w3c-rdfcore-wg@w3.org
At 12:55 02/10/03 -0400, Francois Yergeau wrote: >The last part, however, is not consistent with the first. If data SHOULD be >normalized, then implementations SHOULD NOT accept it when not normalized >(but may, if "the full implications must be understood and carefully weighed >before choosing a different course" [RFC2119] is fulfilled) and SHOULD issue >a warning in such circumstances. 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. Particularly since, in this case, I see no reason why different RDF implementations that choose to accept and process such non-NFC data as valid would arrive at different conclusions. #g ------------ Graham Klyne GK@NineByNine.org
Received on Friday, 3 October 2003 04:58:08 UTC