- From: John Cowan <cowan@mercury.ccil.org>
- Date: Mon, 18 Feb 2013 12:31:55 -0500
- To: "Henry S. Thompson" <ht@inf.ed.ac.uk>
- Cc: Paul Grosso <paul@paulgrosso.name>, core <public-xml-core-wg@w3.org>
Henry S. Thompson scripsit: > The XML Core Working Group notes with concern the proposed [1] > removal from DOM Core of the xmlEncoding, xmlStandalone and > xmlVersion members of the Document interface. Sorry to be stroppy about it, but after considerable thought I find that I cannot associate myself with this complaint. Since the very early days of its development, the Infoset has never attempted to prescribe to its clients which information items or properties they should or should not represent. It simply provides names for the abstract things that an XML document represents at the most superficial level of meaning. > The XML Infoset spec [2] says that standalone, version, and the > character encoding scheme are part of the information content of > an XML document. They are relevant to application processing in > general and essential for any application which attempts faithful > reserialization. If "faithful" means "verbatim", then the whole of the Infoset does not suffice for faithful reserialization. Any other definition of "faithful" is going to be circular, since it involves specifying which properties are preserved and which are not preserved in terms of the Infoset itself. For most practical purposes, reserialization need not regard the original settings of these properties but may make its own choices. When the Infoset was designed, the DOM was the only API for accessing XML documents in a tree-structured fashion, and it was a matter of considerable concern what it did and did not specify. There have been alternative APIs for many years now, and one of the most common criticisms of the DOM has been its over-generality, which leads to awkwardness. DOM4 is plainly aimed at the browser environmen specifically, and as such has better focus than earlier DOM versions, a trend I basically support. So I propose that we say nothing, and allow the WAWG and WHATWG to restructure the DOM as they see fit. The wider XML context will not suffer for it. -- Even a refrigerator can conform to the XML John Cowan Infoset, as long as it has a door sticker cowan@ccil.org saying "No information items inside". http://www.ccil.org/~cowan --Eve Maler
Received on Monday, 18 February 2013 17:32:27 UTC