RK
=================
October 13, 1997
W3C founding member Murray Maloney discusses XML for meta data
By Lynda Radosevich
InfoWorld Electric
Murray Maloney, technical director at Grif, a Paris-based company
developing the Extensible Markup Language (XML), is a founding member of,
and still serves on, several World Wide Web Consortium (W3C) working
groups. In an interview with InfoWorld Senior Editor Lynda Radosevich,
Maloney clears up a few questions about using XML.
InfoWorld: Is one of the potential uses of XML to enable more sophisticated
searching than is available today?
Maloney: This is one of the most obvious and compelling benefits XML could
provide.
InfoWorld: To enable searches on specific categories, say on "Welsh
corgis," won't vertical industry segments have to agree on tags for
XML-based markup languages such as the Dog Breeder's Markup Language (DBML)?
Maloney: That is darned close to being completely accurate. The SGML
[Standard Generalized Markup Language]/XML pedants would tell you that you
don't have to agree on a set of tags [element types], but could instead
agree on a set of attribute names and values. In fact, the use of
attributes to provide deeper meaning is a characteristic of "architectural
forms." The good news is that there are quite a few good examples to draw
on from aerospace, automotive, telecommunications, computer systems, and
software.
InfoWorld: Will the W3C referee all these different vertical industry
languages?
Maloney: I hope not. Each industry needs to collaborate on the markup and
semantics that are appropriate to their needs. Across industries there will
be a need for some common markup and semantics. I expect that W3C may play
a role in some of that. Within organizations, business practices will guide
the process of developing the language elements.
InfoWorld: How will Web tools handle the proliferation of new markup
languages?
Maloney: There are plenty of SGML-based tools and systems already deployed.
These tools are equipped to handle any number of markup languages.
In general, Web tools will have to be equipped to read an XML document,
present it according to the rules specified by a style sheet (which may be
Cascading Style Sheets, Extensible Style Sheets, or some other style
sheet), traverse hypertext links a la Extensible Link Language, and be able
to execute programs [Java, ECMAscript, Visual Basic, etc.] that are
associated with the language.
There are tools that do these things today, but there has not been a common
standard. Suddenly, the will to standardize these capabilities has reached
critical mass, thanks in large part to the efforts of Jon Bosak [online
information-technology architect at Sun].
InfoWorld: Some SGML tools vendors say XML is good for publishing complex
documents on the Web, but it's not good for creating and managing them
because it does not support exceptions. What's your take on that?
Maloney: Nonsupport for exceptions [inclusions and exclusions] in XML
document-type definitions is a good thing. Exceptions would make the
language harder to work with, so they are one of the features of SGML that
haven't survived.
Exceptions make it easier to write document-type definitions. But the
inconvenience can easily be overcome by writing an SGML Document Type
Definition and transforming it into an XML equivalent with all of the
exceptions expanded in place.
This is only relevant for editing documents that adhere to a complex
document structure. More often than not, exceptions have been used to
enforce editorial guidelines or otherwise overcome constraints imposed by
some component of the SGML authoring or publishing system.
As you explore the world of XML and SGML, you discover a phenomena that I
characterize as "SGML pedantry." SGML pedants will be spending a lot of
time telling you, or anyone who will listen, that only "full SGML" can meet
your needs; that only SGML is an ISO standard. To some extent they may be
right. There are things that one can do with full SGML that will be very
difficult or impossible with XML or HTML. But so what? Laser printers in
the '80s couldn't do what high-end composition systems could do, but HP and
Apple did very well. We learned that having design tools didn't make
everyone a designer. Similarly, HTML has been very useful in teaching us
the limitations of a single markup language.
--- Rohit Khare /// Graduate Student /// UC Irvine Computer Science rohit@uci.edu /// Work: (714) 824-3100 /// Home: (714) 823-9705[Urgent? (617) 960-5131 still works to page me]