[DOC-SIG] What I don't like about SGML

[Paul Prescod]

Guido van Rossum wrote:
> 
> > The fact that HTML authors ignore SGML rules
> > is a sad commentary on the Web, not on HTML.
> 
> I think you have a lot to learn.  The customer is always right.  

This has very little to do with the customers. The customers want to get
it right. The browser vendors have consistently undermined them in this
effort by silently accepting wrong documents without even providing a
*mode* that would give a hint that their documents are broken. 

Most HTML users are *amazed* when they are told how many mistakes there
are in their documents. They will typically respond: "But I saw that
broken construct on Netscape's web site!"

> I  would never say anything like that of Python users.

I didn't say anything bad about HTML users. I said something bad about
the Web as an information system. As a parser author, you know as well
as I do how broken it is in the area of document consistency.

If Python didn't give error messages when scripts were broken then the
state of the Python source base would similarly be a horrible mess. SGML
defined a concept of validity to prevent systems from getting into this
state. Browser vendors ignored that and have put themselves in a living
hell of ad hoc parser writing (so much so that they have inconvenienced
everybody by making XML over-strict to compensate).

Fredrik said:
> You mean "HTML editors", don't you?  

I think the browsers are more culpable. We've known for several years
that most HTML authors would not use editors and would use browsers as
their primary validation tool. You're right that HTML editors often
create bad HTML, however. They share the blame.

> You cannot really blaim the millions
> of people that have done so for not reading a standardization document
> that's not even freely available on the Web...

The HTML spec is on the web and most documents do not conform with it.
It has been a relatively "standalone" spec. for quite a while.

Fred said:
>  Roll over and play dead.  But not gracefully.  Even I think that the 
> "what Grail should do" aspect of this is irrelevant; Grail is
> targetted primarily toward the Web that's deployed, and only
> secondarily to make my life tolerable. 

Could you clarify what you are saying here? Are you arguing that Grail
already has to handle any tag in any place so the concept of standarding
them in a DTD is a waste of time? That even bothering to have a concept
of "correct HTML" isn't worth the effort? If the web is going to be an
information system then documents must conform to some minimum standard
of consistency. SGML seems to me the best tool to describe that
standard. Grail was just an example tool that makes up part of the
information system.

 Paul Prescod

_______________
DOC-SIG  - SIG for the Python Documentation Project

send messages to: doc-sig@python.org
administrivia to: doc-sig-request@python.org
_______________