- From: Octav Chipara <ochipara@cse.unl.edu>
- Date: Wed, 15 Nov 2000 23:03:15 -0600
- To: Martin Gudgin <marting@develop.com>
- cc: XML Protocol Comments <xml-dist-app@w3.org>
On Tue, 14 Nov 2000, Martin Gudgin wrote: > I'm not at all sure why we are working at the character encoding level. Most > other W3C XML specs since XML-Namespaces have been defined in terms of the > Infoset, not the character encoding. I find it worrying that there is no > mention whatsoever of the Infoset in our requirements doc. > > I realise that in the 6xx cluster we are talking about protocol bindings but > I don't see any real need to specify a character-encoding. At the end of the > day people will send XML that conforms to the XP spec. Whether that XML is > encoded in UTF-8, UTF-16, ISO-10646, ISO-8859-x or some binary form > shouldn't matter. Either the other end will be able to decode it or it > won't. If it can't people will either stop calling that end point ( in which > case maybe the implementer will support more encodings ) or the sender will > try a different encoding. > > Maybe people will mainly use UTF-8, maybe people will use UTF-16, maybe they > will use a binary XML encoding we don't even have yet. I don't want XP to > restrict the encoding used. Specifying a 'character set of choice' seems > pointless. If we don't *mandate* an encoding ( and I don't think we should ) > then let people choose. > > Gudge > > I think that what you are proposing is very simple. However, having a UTF-8 agress Canonical XML which sets the standard for the simplest XML grammer. It is important to be consistent with other XML related proposals. - Octav
Received on Thursday, 16 November 2000 00:03:23 UTC