Archive of UserLand's first discussion group, started October 5, 1998.

Re: T. Nelson -- Doesn't understand XML?

Author:Paul Snively
Posted:11/10/1999; 11:42:46 AM
Topic:T. Nelson Critique of Embedded Markup
Msg #:12940 (In response to 12921)
Prev/Next:12939 / 12941

Dennis Peterson wrote:

If I have a text stream like "294839543", there are an infinite number of possible meanings for it. You could replace my tag with another giving it a completely different meaning.

Bad news: there are an infinite number of possible meanings for "This is a sentence written by Paul Snively," too. It's just that among a couple of billion people on planet Earth there's a reasonable expectation of a consensus interpretation--one sufficient to enable a process called "communication," in which side-effects of the interpretation are relatively predictable.

Sure, it's just a different interpretation of my number, but it might be a completely erroneous interpretation.

Erroneous according to whom? Don't make the classic error that knowledge representation researchers made early in their field of endeavor, attaching "intrinsic" meaning to human-readable symbols. If you tag something and someone else replaces that tag with , it makes no difference whatsoever as long as whatever is processing the tags is able to do something meaningful with the data tagged .

In other words, ultimately the meaning of XML tags is defined by the behavior of the system that interprets them. One implication of this is that it is entirely legitimate--and, in the real world, quite useful--to interpret an XML document containing one tag set, and recast the document using what may very well be an entirely different tag set, e.g. in order to normalize incoming data from external organizations for aggregation into a large repository. In other words, a very normal data-conversion process.

It could even be the case that my string stands for three different values, which are separated by XML tags, and that your tags will group them into two numbers instead. If we want to use XML to transmit reliably meaningful data, this doesn't seem like a good idea.

With all due respect, I think you miss the point: the meaning of an XML document is entirely up to whatever/whoever interprets it. Naturally, we wish for, hope for, a consensus interpretation of a wide variety of tagsets to evolve. But there may very well be an application that takes your three values, performs some computation on them that results in two values, tags those, and spits out the result. That may be extremely helpful to users of that particular application.

A better way to modify tags, IMHO, is to use XSL tranformations. You can leave the original XML in place if you want, and filter it through your own XSL document to apply your personal interpretation, with less risk of screwing things up by accident, and with the end user better able to separate the interpretation from the original meaning.

XSL is indeed useful for these kinds of straightforward transformations; I think more is needed, e.g. for the computational case described above. In any case, there's zero meaning to the idea of "the end user better able to separate the interpretation from the original meaning" in the absence of a consensus view as to what the original tags mean in the first place, and I'd not wish to rely on the human user providing such a mapping from symbols to meaning in any but the most trivial cases.

There are responses to this message:

This page was archived on 6/13/2001; 4:53:27 PM.

© Copyright 1998-2001 UserLand Software, Inc.