There's scope for a few interesting papers about the publication process -
both the good parts and the bad parts.
Quite a number of the W3C specs are authored using a DTD called "xmlspec" -
because it originated for the XML specification itself. Keeping control over
this has been a bit of a struggle, there are now quite a few versions and
variants. Many of these are managed using DTD customization techniques, but
I think there has also been some forking. Similarly, there's a common XSLT
stylesheet to translate into HTML, but this also has variants - generally
managed using xsl:import, but I think there's been some forking as well. The
variations are needed for the requirements of each spec - the XSLT spec for
example has extensions to handle the way element syntax summaries are
presented; but managing the variety is a continual challenge.
The XSLT spec is managed as part of the family of "QT" specs (XQuery and
XSLT), which have a single build process. There are mechanisms to maintain
cross-references between the specs, for example, and a joint database of
references to external specifications so that when we want to refer to a new
Unicode version, in theory it can be done by changing one central document.
There's also a lot of technology for managing the syntax rules (this affects
XPath and XQuery rather than XSLT) - the rules in the specs are generated
from the same source as the W3C syntax-checking applets, and the build
process checks examples against the grammar. For the next version of XSLT
I've also added tools that will generate SVG diagrams from source XML
descriptions of trees. Similarly I've recently added mechanisms in the build
for the Functions and operators spec to check the all the examples against a
reference implementation. So there's a fair bit of technology deployed here.
The build process is a collection of Ant scripts.
If there was a clever tool and source format in use (other
than typing plain HTML with vi!) I'd like to know... of
course if it's not a W3C secret!.. as I have sort of the same
publication work to do.
Have you considered docbook? The xmlspec system is a bit specialized.
Regards,
Michael Kay
http://www.saxonica.com/
http://twitter.com/michaelhkay
--~------------------------------------------------------------------
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
To unsubscribe, go to: http://lists.mulberrytech.com/xsl-list/
or e-mail: <mailto:xsl-list-unsubscribe(_at_)lists(_dot_)mulberrytech(_dot_)com>
--~--