Re: Sudden death: request for missing input

Dave Peterson writes:

> At 7:44 AM 4/30/97, Todd Freter wrote:
> >XML's ease of implementation and ease of use for hand-coders should be a
> >significant plank in any positioning and marketing of XML. Certainly XML's
> >ease of use should be more fundamental than "the customer is always right."
> "Ease of use" or "ease of implementation"?  "The customer" usually wants
> what he considers "easy to use", whether it's a good idea or not.

I guess I mean that it's not only easy for application developers to implement
XML but also for document creators to produce well-formed XML, even when they
are coding by hand. It just isn't that hard.

> When you argue for "ease of use" of XML, remember that the document creator
> as well as the document reader is a "user" of XML.  OMITTAG (which is
> essentially an error-recovery mechanism embodied in SGML to the point
> where if you can use it to recover, the trigger is not considered an
> error after all) was, I believe, introduced for "ease of use for
> hand-coders". Are you asking for something like that?  It sounds like you're
> arguing the other way.  Which may be a good idea, but the user might argue
> that it's not making XML easier to use.

Earlier, Paul Prescod argued that the goal with XML should be correct
documents, and I think that's right. And Tim's compelling point is that XML's
well-formedness characteristics are easy to achieve.

I'm not asking for anything like OMITTAG. You're right; I'm arguing the other
way. OMITTAG makes sense in a setting where validation is a constant concern,
but in XML validation is optional, and something like OMITTAG could be
problematic without a DTD's context in which to interpret markup. If in XML
you have neither a DTD nor reliable standards for well-formedness, you don't
have much.

I doubt that XML will lose much of its constituency if XML hand coders are
required to enter closing tags. I think those customers will rely on parsers,
"lint"-type tools or strength of character to insure well-formedness. If XML
proliferates as much in the next 18 months as some people predict, there will
be desirable markets for robust XML authoring tools. The other "ease of
use"-seeking customers will probably not be unserved. Look at what GRIF has
done already with their XML+CSS tool. It's just the beginning.


Received on Wednesday, 30 April 1997 11:43:36 UTC