Re: Parameter entity references in WF docs

Bert Bos wrote:
> 
> From: Tim Bray <tbray@textuality.com>
> 
> > 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, particularly for
> lightweight
> > standalone processing.  It seems like PE's are largely in the spec
> to
> > support authoring-end activities.
> >
> > Would it be reasonable to think about saying either that
> > (a) PErefs should not be used in the internal DTD subset, or
> > (b) PErefs should not be used in WF docs?

My vote would be to add nothing to the spec until you get a
1.0 version out.  At this point that "extremely simple 
dialect of SGML" rant is getting thin as the language 
gets fat.  Sometime soon, some people will get wise and 
will begin to question the full court press of articles 
based on wishful thinking about this.  

> There is a better solution, and even more in the spirit of XML:
> define XML in such a way that no parser ever needs to parse a DTD.

Stop the *spirit of XML* arguments.  Requirements were used in the 
beginning to stop and start decisions; now is not the time 
to switch over to *spritualism* based on private 
wants or speculations.  That sort of thing only works on 
the painfully naive and the membership of this particular 
working group is past that appeal.
 
> That would leave the burden of parsing DTDs (and PEs) to certain
> generic validators and generators. You don't need many of those, but
> you do need hundreds of normal parsers.
> 
> This is not just a wish. I think this is absolutely necessary.

Sort of a wishful thought then.

> Most
> parsers *will* treat XML as if it didn't need a DTD and it is better
> for interoperability if there is a spec for that format. 

That's one prophecy.  There aren't a dozen XML applications 
out there and the people making this prediction have the least 
experience with markup systems of anyone on the list.  Not 
good credentials. Here's another prophecy: many people without DTDs 
will quickly write them once they work out it isn't hard and it is
extremely 
useful.  

> Section 2.10
> (Required markup declaration) of the latest working draft already
> outlines the expected problems. That's OK for a draft, but eventually
> the spec needs to solve them.

The spec can't even tell us what the output of a parser is.

Requirements.  Justify based on the requirements.

len bullard

Received on Wednesday, 28 May 1997 21:03:37 UTC