W3C home > Mailing lists > Public > www-html-editor@w3.org > October to December 1997

Re: move "SGML limited support" info to appendix

From: Dan Connolly <connolly@w3.org>
Date: Sun, 26 Oct 1997 21:04:05 -0600
Message-ID: <345404A5.6CE6@w3.org>
To: "Ian B. Jacobs" <ij@w3.org>
CC: www-html-editor@w3.org
This message contains NO editing instructions.

Ian B. Jacobs wrote:
> >
> > Changing "should" to "must" is a change in the language.
> First of all, YOU NEEDN'T YELL.
> Second, you are wrong. Consult
> http://www.w3.org/MarkUp/Group/9710/WD-html40-971024/intro/sgmltut.html:
> a) The header reads:
>      Processing Instructions <?your user agent should hide these>
> b) A few sentences later we have:
>      User agents must not render SGML processing instructions
>      (e.g., <?full volume>).
> The spec was inconsistent, but I did not make a careless mistake.

I see; I apologize for my reaction.

I need to research this a bit...

Is it convenient for you to tell me when the "must not render"
sentence went in there?

Aha... I can find this out myself with cvs annotate.
It looks like most of the section was written by Dave
on 29-Jul-97, and you touched it up on 18-Oct-97; in

1.21         (dsr      29-Jul-97): <h4>Processing Instructions <?your
user agent should hide these></h4>
1.21         (dsr      29-Jul-97): 
1.39         (ian      18-Oct-97): <P>User agents must not render SGML
processing instructions (e.g.,
1.39         (ian      18-Oct-97): &lt;?full volume&gt;).

Let's see what was there before 1.39...

There was no "must" in that section in Dave's 1.21.

Lemme check the log (I wish I had taken time to track issues
this carefully all along. But at least the history is
all there. I love CVS!). Hmmm.. "John Burgers comments"
Too bad that's not linked to his email. Ah for better
tools... Anyway... checking the archve...

Hmm... he quotes the "must not render" sentence in
his message of 16 Oct 1997, but 1.39 wasn't written
until on 18Oct. Oh! I see... the "must not render"
sentence was not in that section.

more checking...

1.1          (ian      25-Jun-97): document. Similary, user agents must
not render SGML processing
1.1          (ian      25-Jun-97): instructions (e.g., &lt;?full
1.1          (ian      25-Jun-97): 

So... it's been there since 25 Jun. Dave probably should
have noticed it when he re-wrote the processing info on
29 Jun, but he didn't.

Well... what's the bottom line?

The spec is inconsistent, and has been since 29 June.
It wasn't your fault, but rather an unfortunate sequence
of events. Still... it shows that the whole process is
somewhat lacking.

So... what do I do about the inconsistency? I sort
of have to guess what the intent of the WG is.

I'm more comfortable putting this issue in notes.html
with the rest of historical bugs etc.

But DO NOT take that as an editing instruction; this
is a substantive issue and I'll send it to you via

Received on Sunday, 26 October 1997 22:02:38 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:08:19 UTC