Re: [Gnu-arch-users] Re: documentation as info

[Tom Lord]

    > From: Colin Walters <address@hidden>

    > On Tue, 2003-09-16 at 19:09, Tom Lord wrote:

    > > Are you really at a loss to find many of the available critiques of
    > > XML? 

    > Oh there are plenty - but I don't think there are any that hold any
    > water.

heh.

    > Let's take your documentation format, for instance.
 
    > 1) How do you specify what character set it's in?

That's not an issue for this format, at the moment.


    > XML solves this problem; XML documents declare the character set
    > in use at the top of the document, and default to UTF-8 (which
    > is well-known since it's a standard).

This is hardly an excuse for the entire mess that is XML.

    > 2) How can you specify characters outside the range of what you're able
    > to type?

    > XML solves this problem using entities.  This is standardized, and all
    > XML parsers grok them.

Probably -- but I have far less confidence that all - or even _any_ -
XML parsers actually implement XML.

    > 3) How can I verify the structure of the document?

    > XML solves this with XML Schemas and DTDs.  

It's funny you should say that.   Would you care to point out a Schema
validator with a correct regexp matcher?

And that cuts deep.   Because as these things are deployed with
incorrect matchers -- what's the consequence?   It's analogous to the 
way that HTML parsers have to put up with so much bogus HTML -- only,
wasn't XML supposed to fix that very problem?

Moreover, the entire idea of a declarative validator is entirely
suspect.   The complexity of implementation of such a validator is
right up there with a programmatic validator, the latter of which
would permit more flexible and more useful validation specs.

    > And moreover, there are *tons* of tools to transform XML documents to
    > other formats.   Transforming your nonstandard ".doc" format involves
    > telling people "Oh, you have to get my scheme implementation, and then
    > this software..."...

This is so extraordinarily far off-topic and non-responsive and
straw-mannish that it's remarkable.  Have I ever _once_ said that,
instead of XML, people should be embracing my ".doc" format?  Even
_once_?  What the hell are you talking about?

For various pragmatic reasons, my .doc format is currently the right
thing to be using in software that I maintain and distribute.  My .doc
format exhibits some design ideas that I think are good ideas in a doc
format, especially for technical documentation, especially of C code.
I think those are about the strongest claims I've made.


    > These three are just a few examples of Real Problems that XML solves.

Sorry, but those are not Real Problems except in the
castles-in-the-sky imaginations of W3C.   Nor are the purported
solutions widely and verifiably implemented with accuracy -- which
just creates a whole new set of problems.

-t