XML

[Alan Kennedy]

Lulu of the Lotus-Eaters wrote:

> I just read a couple dozen 7-bit ASCII documents on c.l.py before
> getting to Martin's 7-bit ASCII document arguing against the prevelance
> of 7-bit ASCII documents (and I haven't even opened my favorite HTML
> viewer for the day yet).

But I think we've all got to admit that we're non-typical computer users here in
c.l.p.

1. Usenet: Yes it is true that Usenet messages are 7-bit ASCII. But look at how
much parsing and wrapping has to go on in order to make 8-bit characters (áéíóú)
travel safely through the 7-bit transports that may be encountered on Usenet.
Now, I'm not saying that MIME, or anything like it, should be replaced with XML
(although some people do [1]). But I think it's much tougher to construct and
parse MIME than XML, even with the great "batteries" that have been "included"
with Python. And MIME is one of the simpler "protocol"s that are encoded in
plain text.

And it would be interesting to know how many readers of c.l.p. read it through a
HTML interface such as Google Groups, because the interface is so much richer
than something like Netscape Messenger, for example.

2. Email: I can only speak about my own experience, which is that whatever
servers I maintain send far more HTML everyday than they do emails, by a factor
of between 100 and 1,000. And I personally read many more web pages than I send
and receive emails (except for the dratted SPAM >:-( the worst of which is
always in HTML.....)

One more point: although .mbox files are textual, most of the world's email is
stored in MS Outlook binary *.pst files. Or in some proprietary format on
(hot|yahoo|*)mail.com servers, and viewed through a HTML interface. I know
*dozens* of non-IT people whose only email address is a hotmail one.

3. Programmers hard disks: I think it's always going to be the case that
programmers will have more "plain text" on their disks than non-programmers (go
on, admit it, you enjoy this stuff, don't you ;-) I can imagine that the average
Windows users disk would contain very few text files (*.ini, *.bat, etc).

4. Semantics vs. format: The OP was concerned with finding good use cases for
XML, where it is a natural fit for the problem at hand. He seemed to be
dissatisfied with XML both as a format, and as a method for building semantic
object models. My thinking is that it is the process of mapping format ->
semantics that is the key issue: The easier that is, the more likely it is to be
usable and to catch on. I think XML will continue to be grow in usage (not
always used appropriately) because the format is trivial for people to
understand (at a minimum requirements level), there is a common data model which
most people can get their heads around, and the mapping from one to the other is
easily dealt with through usage of the plethora of high-quality, standards
compliant XML processing software that is out there.

Refs.

[1]XML MIME Transformation Protocol (XMTP)
http://www.openhealth.org/xmtp/

-- 
alan kennedy
-----------------------------------------------------
check http headers here: http://xhaus.com/headers
email alan:              http://xhaus.com/mailto/alan