[Note The post by Eliot and followup by David Megginson both contain a couple corrections per (what I understand to be intended in) Eliot's re-post w/ a correction for the explicit "name" attribute -rcc].
Subject: Architectures, Schemas, and XML: Proposed Amendment to ISO/IEC 10744:1997 From: "W. Eliot Kimber" <eliot@isogen.com> Date: 1997/12/12 Message-ID: <349190F6.D0C6102D@isogen.com> Newsgroups: comp.text.sgml --------------------------------------------------------------- At last week's WG4 (SGML) standards meeting, James Clark and I put together a one-page proposed amendment to ISO/IEC 10744:1997 (HyTime) that provides a PI-based syntax for declaring the use of SGML architectures (schemas). The proposal has been submitted for immediate ballot, which should be completed in the next three or four months. This amendment is an implementation of the various "PI for architectures" proposals made recently (e.g., David Megginson's paper presented at the XML Developer's day). The design provides more meaningful (and obvious) names for the architecture configuration attributes. You can find the full text of the amendment at http://www.ornl.gov/sgml/wg8/document/1957.htm A typical architecture use declaration within an XML document would look like this: <?XML version="1.0" ?> <?IS10744:arch name="isobase" public-id="-//ISOGEN International Corp.//NOTATION ISOGEN Base Architecture//EN" dtd-system-id="http://www.isogen.com/archs/isobase/isobase.dtd" ?> <mydoc isobase="isogen-document"/> The minimal declaration simply provides an architecture (schema) name: <?XML version="1.0" ?> <?IS10744:arch name="isobase" ?> <mydoc isobase="isogen-document"/> However, you would normally want to point to at least the DTD declarations, if not the public name for the schema, just so it's clear what the architecture name really refers to. Of course, in the case of well-known or widely-used schemas the name may be sufficient (e.g., RDF, HyTime, etc.). By default, the architecture name (the name following "IS10744:arch") is the name of the attribute used to map elements to elements in the architecture. The attributes shown in the example are interpreted as follows: public-id The globally-unique name for the architecture as an abstract concept, that is, as as set of rules that govern documents. These rules could be defined in whole or in part using any schema mechanism, including something like the XML-Data proposal. dtd-system-id The system ID for the DTD-syntax declarations for the architecture. The architecture DTD declarations can be used by architecture-aware processors to perform syntactic validation of the document according to its architectural mapping. For example, the SP parser does this today. As this amendment simply provides an alternative syntax for an existing facility of the standard and does not change the functionality in any way, we do not anticipate any opposition to its approval. ---------------- Date: Sat, 13 Dec 1997 17:58:55 -0500 Message-Id: <199712132258.RAA00384@unready.microstar.com> From: David Megginson <ak117@freenet.carleton.ca> To: xml-dev Mailing List <xml-dev@ic.ac.uk> Subject: XML Architectural Forms I don't remember seeing an announcement here (apologies if I'm mistaken), but Eliot Kimber and James Clark have announced on comp.text.sgml a proposed amendment to ISO 10744 that will make it possible to use Architectural Forms in XML. You can find the text of the amendment at the following URL: http://www.ornl.gov/sgml/wg8/document/1957.htm Here's Eliot's example of a simple, well-formed XML document that uses the base architecture "isobase": <?XML version="1.0" ?> <?IS10744:arch name="isobase" ?> <mydoc isobase="isogen-document"/> This is very exciting, because if accepted, the amendment will make it possible to solve the XML namespace problem with an International Standard, instead of forcing the W3C to throw together a consortium standard. Base architectures also provide a simple and elegant solution to multiple inheritance; for example, here's Eliot's example modified to implement _two_ base architectures: <?XML version="1.0" ?> <?IS10744:arch name="isobase" ?> <?IS10744:arch name="mslbase" ?> <mydoc isobase="isogen-document" mslbase="microstar-document"/> The element <mydoc> corresponds to <isogen-document> in the isobase namespace and to <microstar-document> in the mslbase namespace at the same time. Even more interesting is the ability to embed the architectural attributes in a DTD, so that they do not appear in the document instance at all. For example, you can create an external DTD like this: <?IS10744:arch name="isobase" ?> <?IS10744:arch name="mslbase" ?> <!ELEMENT mydoc EMPTY> <!ATTLIST mydoc isobase NMTOKEN #FIXED "isogen-document" mslbase NMTOKEN #FIXED "microstar-document"> Now, every XML document that uses this DTD will implement the two architectures automatically, with no additional markup required: <?XML version="1.0" ?> <!DOCTYPE mydoc SYSTEM "mydoc.dtd"> <mydoc/> Authors won't even have to know that they're using architectural forms. Congratulations are due to Eliot and James for taking the time to start this process. David -- David Megginson ak117@freenet.carleton.ca Microstar Software Ltd. dmeggins@microstar.com http://home.sprynet.com/sprynet/dmeggins/ xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@ic.ac.uk Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ To (un)subscribe, mailto:majordomo@ic.ac.uk the following message; (un)subscribe xml-dev To subscribe to the digests, mailto:majordomo@ic.ac.uk the following message; subscribe xml-dev-digest List coordinator, Henry Rzepa (mailto:rzepa@ic.ac.uk)