[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: Validation - a history.

• From: "Len Bullard" <cbullard@h...>
• To: "'Gavin Thomas Nicol'" <gtn@r...>, "'XML List Developers'" <xml-dev@l...>
• Date: Fri, 2 Mar 2007 09:58:06 -0600

From: Gavin Thomas Nicol [mailto:gtn@r...] 

>One side of that was that people found DTD's to be useful only at  
>certain points in the process... during the production process a  
>document is not always/seldom 'valid'. Market forces made them  
>optional ;-)

Experience did.  True.  Even prior to IADS when working with the Context
system at GE on the US Navy CASS program, I learned that we had to turn off
validation until BigBindTime which for that system was bitmap production.

1. Syntax valid (aka, well-formed):  no typing mistakes given a spell
checker.

2. Logically valid:  my code has no bugs.

3. System valid:  your code and my code have no bugs.

4. Requirements valid:  the user sees no bugs.

But we all knew that.

>I think that there were actually few, if any 'true' SGML parsers. The  
>reality was that a significant number of SGML *subset* parsers  
>existing that shared a fairly common set of features.

Mostly true.  The concepts of 'Unicorn-goodness' etc. were just coming along
when the web burst.  OTOH, the cost of parsers was mostly a burden a few
years prior to the web.  By the time HTML was being noticed, it wasn't an
issue.  It has to be remembered that when ISO 8879 was released,
incompatibilities in computer systems passing files around were much worse
than the time that the desktops made the web-as-we-think-we-remember-it
possible.  SGML was a feature-ridden mostly because to guarantee that list
above, a lot more checking of agreement was needed.

>Very true, though it is useful to see everyone's perspectives on the  
>battle for supremacy.

No disagreement on that.  But there is a lot about that speech that reminds
me of the officers who sat in the Pentagon in the 1950s planning future
aircraft and claiming that since we would never again fight a conventional
war, we'd never need guns on jets, just rockets to take down bombers.  They
kept defending that decision well into the 1960s while the guys on the
ground crews were trying to install guns in the field to take down MIGs that
had them.

Dogfighting never goes out of style.  So it is with items 1 through 4 above.
We just keep shifting the places in the process where the guns are mounted.

len

• References:
  • Re: Validation - a history.
    • From: Gavin Thomas Nicol <gtn@r...>