Dave Crocker <dhc2(_at_)dcrocker(_dot_)net> writes:
ps. I am strongly hopeful about long-term use of XML. My concerns are
only about the stage of market development, and as I said, it is
progressing nicely.
Just to throw out another random data point from someone who tends to rant
against XML, my primary objection to XML is as an original source format.
I find it very author-hostile due to the verbose tags, the need for rather
a lot of markup in the document to take full advantage of the capabilities
of the language, and the use of ASCII text for the tags as well as the
content (which makes it hard to visually distinguish between content and
markup). HTML has all of the same problems. The editors tend to be large
black-box sorts of applications without much real control over the result
and heavily slanted towards people who adore GUIs and compose documents
with lots of pointing and clicking.
What would bring me around to accepting the idea of XML as a standardized
document format is better converters from markup languages that are either
author-friendly (something like POD, SDF, or texinfo) or old enough that
people already know them quite well (like LaTeX or *roff). I've seen a
lot more things like that recently than I had before, which is hopefully a
sign that the use of XML is expanding beyond the people who, due to
whatever strange mental quirk :), actually like SGML syntax. Once those
converters are pretty common and reasonably solid, I think XML will be a
much easier sell.
It's certainly pretty clear to me that manually paginated ASCII text with
embedded "page headers" that pop up in the middle of the content when
being read with a pager isn't ideal as an archival format.
The best thing people who are in favor of XML for RFCs could do in my view
is to write a few simple converters from much simpler languages that could
handle the average RFC. (Even the very simple capabilities of POD are
adequate for quite a few RFCs.) Then the people who are just trying to
describe a protocol and don't want the document markup getting in their
way can use something simple and the people who are trying to communicate
complex information that needs a lot of semantic markup can use something
with more expressive power.
--
Russ Allbery (rra(_at_)stanford(_dot_)edu)
<http://www.eyrie.org/~eagle/>