Profiling and Packaging XML

[Thomas B. Passin]

From: Simon St.Laurent <simonstl at simonstl.com>


>It seems like everyone has their own list of favorite features in XML, as
>well as a list of features whose inclusion in XML they consider heretical.
>(The W3C, for instance, doesn't seem very happy about processing
>instructions, based on some statements in things like the style sheet PI
>Rec.)
>
>Since the SGML declaration was first against the wall (mostly an excellent
>thing), we don't have any way whatsoever to indicate that features are
>required, optional, or prohibited.  We don't even have a way to indicate
>that documents might, for instance, require validating parsers to avoid
>losing the 99.9% of their content that happens to reside in external
>resources.  (I've done a lot of work on this in XML Processing Description
>Language, XPDL, at http://purl.oclc.org/NET/xpdl, but figure my work will
>show up in some other project at best.)
>

>It sounds like the W3C is pondering XML packaging in round III of the XML
>Activity.  Might this be a good time to discuss what kinds of information
>we'd like to see in those packages?  Does profiling seem like a useful
>tool?  It seems like it might allow developers to specify rules like 'no
>external entities', 'no PIs', and 'no namespaces' for particular
>applications, and let us all have our opinions but not necessarily inflict
>them on everyone else's parsing.


One way to specify the presence or absence of optional features would be to
insert an element just after the root element that functions somewhat like the
xslt  <xsl:output> element.  An xslt processor determines what output format is
desired based on this element, which is optional.  Similarly, an xml processor
could look at a tag, perhaps <features-xml:required
external-unparsed-entities="no">, and learn what features are needed to process
the document.  There would be a namespace for features:xml, and a processor
would be free to disregard the features attributes.

It seems like a PI could also provide these services just as well.  Either way,
a processor that does not understand the instructions would just provide full
processing as usual.

Of course, there are lots of ways to accomplish the goal of specifying features
needed by a document.  The BIG question is, should such a mechanism become part
of the XML standard or not?  We have several precedents to look at.  Namespaces
and associating stylesheets were accomplished without adding them to the XML
standard itself.  I suggest that this is the way to go - have another small
standard for specifying optional features  This standard MAY be implemented by
xml processors.

>
>(I suspect that everyone will actually continue to use XML 1.0 parsers
>rather than write their own, but I don't think that reduces the need for
>documenting such profiles.)
>
>Phew!  A week away from XML-dev.  Lots of messages, almost all of them
>interesting.
>
>Simon St.Laurent
>XML: A Primer, 2nd Ed.
>Building XML Applications
>Inside XML DTDs: Scientific and Technical
>Sharing Bandwidth / Cookies
>http://www.simonstl.com
>

Tom Passin


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To unsubscribe, mailto:majordomo at ic.ac.uk the following message;
unsubscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)