- From: Henrik Frystyk Nielsen <henrikn@microsoft.com>
- Date: Thu, 5 Dec 2002 16:46:00 -0800
- To: <noah_mendelsohn@us.ibm.com>
- Cc: "Martin Gudgin" <mgudgin@microsoft.com>, <xml-dist-app@w3.org>
I didn't mean to make any general characterization of possible infosets. My focus was merely the serialization of a SOAP message infoset defined by "application/soap+xml", which IMO, doesn't include items not present in the SOAP message. I think the discomfort I have with introducing text to the effect indicated below is that it gives the impression that there is inherent flexibility in the serialization which I don't think there is, but as I said, maybe I am missing something. >So, you're implying that there is no infoset that serializes >as having, for example, an element declaration? No way to >say: this attribute is of type ID? That stuff just can't be >inferred from the infoset, so you always need additional >information to decide to serialize such things. I don't think >that's how I would have built Infoset, but I suppose it's a >plausible conclusion. If everyone else agrees, I can live >with it. I >remain nervous that at best we're missing an opportunity to >make crystal clear something that now needs quite careful >reading, Still, if nobody else shares my concern, let's drop >it. Thanks. Henrik
Received on Thursday, 5 December 2002 19:46:47 UTC