Re: Parameter entity references in WF docs

Paul Grosso wrote:
> 
> At 09:19 1997 05 28 -0700, Tim Bray wrote:
> >I'm now putting PE reference handling in Lark, and it's turning into
> >quite a lot of code - since these things are just for DTD's, and Lark
> >is nonvalidating anyhow, I'm wondering if the extra processing and code
> >size required are in the spirit of XML,

[snip]
 
> I, for one, am getting the uneasy feeling that XML is starting to
> suffer from "over-indulging":  it's putting on too much weight.

I for two am agreeing.
 
> Of course, everyone wants it to be powerful enough to solve his/her
> problem, but then it risks being too big to be the ubiquitous solution
> we want it to be.  

Then use SGML.  That has been the standing answer to the request 
for more functionality since James Clark first used it during 
the character set debates.

> (I had the same feeling with various SGML Open
> technical resolutions:  I started out thinking something simple would
> give us a good cost/benefit ratio, and then features kept getting added
> until I feared we were reducing the chance of having it widely implemented.)

Mission creep:  a committee should work on what is on the agenda at 
the break and never work on what it proposes over lunch.

> I say this not from an implementor's point of view, but as one who just
> wants XML to succeed:  SGML and HyTime "do it all" but get stuck in a
> small niche, and HTML grabs at a very small part of the problem, but
> pulls in a huge market.  

Time and conditions change with requirements.  What worked the first 
time may not work again.  XML will succeed if there is a definite 
mission it can fulfill.  As "SGML On The Web", it has a mission and 
we generally know what that is.  As "The Fix for HTML" it is 
proposed for a mission most are not sure is worth risking existing 
investments in.  As the "We Can Do This Better Than WG8 and Bury 
That Awful HyTime Thing", it is an intellectual heist and some 
what like listening to Pat Boone sing "Good Golly Miss Molly".

> I'm beginning to lean (even more so that I
> may have at first) toward trimming what we can.  Though I reserve the
> right to be convinced otherwise, I would currently support ideas to
> simplify XML by restricting the use of PErefs.

I lean toward a solid set of requirements to which both the ERB and the 
list abide through the first version of XML 1.0.  After a 
period of implementation and market feedback (give it six months) 
the next version can be proposed, drafted, and worked.

Right now, it is becoming  mess of private experimentation just 
as the press is calling it something it cannot be until implemented,
fielded and tested.

Len Bullard

Received on Wednesday, 28 May 1997 21:16:44 UTC