Re: Parameter entities vs. GI name groups

[Tim is getting an extra copy of this because I haven't yet seen
the results of the vote on list behavior.]

Tim Bray wrote:
| At 02:07 PM 19/06/97 -0400, B. Tommie Usdin wrote:
| >It seems to me that is really a political question, or perhaps a
| >theological one.  
| 
| Except for this sentence, I agree with the rest of Tommie's posting.
| I want to lose PE's and am willing to accept that for people who need
| to live in a complex-DTD world, they'll probably have to use Full SGML.

Ah, so XML is not only something that obdurate B.S.'s can code
parsers for in a week, it's something that's fundamentally unscaleable.
Now I understand.

| But Eve is correct, I think, in saying that namegroups in declarations 
| do replace one or two common PE usages.

Not in functionality or utility.

| My willingness to make the the trade-off is because despite really a lot
| of work by myself and (even more) Michael, the PE section of the XML-lang
| spec is blatantly hideous compared to the rest of it.  It is hard to

So rewrite it.  These are macros, right?  CS grads understand macros, right?

| explain, hard to understand, and hard to implement.  It needs a supporting
| section in an appendix to try to "explain" it (always a bad sign). It is a 
| psychological barrier to the acceptance of descriptive markup.

?  There are no p.e.s in an instance except as marked section keywords
(if still allowed there).  We're not talking about markup but the
specification of markup.  

| Partly because this because the current PE reference replacement rules 
| are arguably B.A.D. (broken as designed) - Michael and I came up, I think,
| with a significant innovation in specification tactics in the use
| of the %-operator, and the result is still very very complicated.

I find the "%-operator" almost prevents me from reading the BNF.  Your
argument here is that you can't explain a perfectly clear concept 
clearly, so you want to drop it.  This is not a suitable argument to
make to the WG whose time you are uptaking.



Regards,
  Terry Allen    Electronic Publishing Consultant    tallen[at]sonic.net
                   http://www.sonic.net/~tallen/
    Davenport and DocBook:  http://www.ora.com/davenport/index.html
          T.A. at Passage Systems:  terry.allen[at]passage.com

Received on Thursday, 19 June 1997 20:50:36 UTC