Re: Parameter entities vs. GI name groups

Paul Prescod wrote:
> 
> len bullard wrote:
> > Good.  I suspect parameter entities will be reintroduced
> > at a later date, but given the resistance to DTDs in WebLand,
> > let them discover why.
> 
> Rather, they will most likely give up on DTDs as too verbose and
> repetitive. They could standardize, rather, on Bray/Netscape-style
> "schemata".
> 
>  Paul Prescod

That one scares the hell out of me too, but I don't think 
it will be because of lack of PEs.  I think that has been 
on some minds since SGML On The Web Day 0.  It definitely fits the 
comments of some folks from the HTML community for some time.
Really, I don't care about that.  It is a long game and 
the ball will swap sides a lot.  Since I don't build browsers, 
I don't care who wins.  But I do care about the quality of 
the tools I get.  I think a season without PEs is fine.  
Here's another one that will raise hackles:  since the 
first of this decade when I got my hands on my first real 
SGML hypertext system, I've never had to use a general 
entity for *transcrewsion*.  Links did the job.  For print, 
different story.

Yet, when it came time to validate HTML, it sure was the 
cheap way to do it, particularly with that free parser 
ARCSGML.... I mean SGMLS.

If they give up on DTDs, let's give up on them and move on.
I don't think they will.  XML is spreading beyond them.  
I am reading a VRML group looking at DTDs seriously as one 
way to think about visualization for document sites.  Really.
I'm not sure MS and NS have quite as much clout as 
is regularly assumed.  Java is throwing a pretty big 
wrench into that.  A framework is not a browser.

I really hate to work with an implicit threat in the 
conversation.  If one gives up DTDs, it gives up one of 
the major competitive advantages of SGML over it's 
competitor.  If the competitor gives it up, they give 
up their business because the other will cut their throats with it.
Standardizing around a DTD is the sure and fast way 
to gain market share built over a stable, validatible 
standard.  If they don't understand that, they really are idiots.

But giving up PEs for a year to six months, nope. It 
gives us an opportunity to think about the kind of 
application extensions that might really be a classThang.

len

Received on Friday, 20 June 1997 22:23:48 UTC