The XML spec in XML: missing tags

Peter Murray-Rust peter at
Wed Feb 25 19:28:20 GMT 1998

At 10:47 25/02/98 -0500, Eve L. Maler wrote:
>Oh, you want *documentation*, do you??  Well, the DTD was hard to write; it
Not ME. I was weaned on 5-hole paper tape. Variables should be no longer
than 1 character. 

>should be hard to understand. :-)

Yes. I strip comments from FORTRAN programs as it is good for the soul and
saves cards.

I must have dreamed it, but someone posted a month or two back that
documentation was a *required* part of a DTD :-)

>Seriously, I keep saying that I'll release the reference documentation Real
>Soon Now, and in fact I'm hoping to be able to spend a few hours tidying it
>up and releasing it later this week.  (There's also a minor DTD update in
>the pipe.)

Great. Seriously - although it wasn't perhaps intended, rec.xml is a
splendid vehicle for people to cut their teeth on - it's got structure,
uses normalisation, has a good variety of elementTypes but also uses some
in a generic manner. The only thing it doesn't use is entities. I have
tweaked my SAXSplit jiffy to do produce entities for div1, etc.

And - an argument for preserving comments in document structure - there is
some splendid archaeology inside...
><termdef> is a really odd way to do term definitions, for my money, but
>that's what the users wanted. :-)  It captures an "inline" definition of a

*Users*?? DTD by committee?? gulp.

>term, and because of the mixed content model, it can't even ensure that a
><term> is present to identify the actual term being defined.  Likewise, it
>can't ensure that the definition captured functions as a "standalone"
>sentence or set of sentences.  I suspect that the cut-off sentence was more
>in the spirit of poetic license.

Fair enough. The approach I am taking to terminology is based on MARTIF
(ISO12200 and ISO12620) - MARTIF itself having strong TEI roots. So I shall
use some simple heuristics to transform termdefs to my termEntry's
><term> is occasionally used legitimately without a <termdef> wrapper; it's
>marking a term being used in a special way, without an accompanying

Yes. I shall abstract these.

>Gee, maybe I should just collect all the questions and do the documentation
>as a Q&A...

Not a bad idea. I certainly don't want you to go to a lot of trouble. One
line sentences for each elementType are probably OK, plus any hardcoded
semantics (e.g. what the target of IDREFs may/maynot be. [I have a set of
simple tools in JUMBO that allow you to browse documents, so you find all
elementTypes, their allowed children, attributes, attribute values, etc.
and can then display the actual location in the document. You can then make
a pretty good guess at what they mean.]


Peter Murray-Rust, Director Virtual School of Molecular Sciences, domestic
net connection
VSMS, Virtual Hyperglossary

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as:
To (un)subscribe, mailto:majordomo at the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list