W3C home > Mailing lists > Public > xml-editor@w3.org > July to September 1999

Re: Omission in XML 1.0 AttValue syntax definition

From: C. M. Sperberg-McQueen <cmsmcq@acm.org>
Date: Sat, 28 Aug 1999 11:08:32 -0500
Message-Id: <4.1.19990828104101.00a5be20@tigger.cc.uic.edu>
Message-Id: <4.1.19990828104101.00a5be20@tigger.cc.uic.edu>
To: Bengt Richter <bokr@accessone.com>, xml-editor@w3.org
At 00:20 99/08/28 -0700, you wrote:
>	There seems to be a missing line following
><excerpt>
>[10]   AttValue :=    '"' ([^<&"] | Reference)* '"' 
>                   |  "'" ([^<&'] | Reference)* "'"
></excerpt>
>
>	namely,
><missing-line>
>                   |  Nmtoken
></missing-line>
>
>	Or am I dreaming? Certainly this omission would seem
>to exclude a large class of useful, concise, and easy to read
>expressions.

With respect, I believe you are dreaming.

The expressions excluded are indeed easy to read, but it appears 
that they are not easy to write or recognize correctly.  Experience 
in SGML systems, including HTML, shows that typical users -- and 
many implementers of tools which read marked up text -- have
rather imprecise notions of what is and what is not a name character,
and thus are either frequently unsure about whether a value needs
to be quoted or not, or else seldom unsure but frequently wrong.

The simple rule "quote the attribute value -- always" is a 
simplification for those people and systems involved in creating
documents, rather than for those involved with reading documents.
It was not introduced out of a false belief that unquoted nmtokens
are hard to parse; it was introduced out of the conviction that
the SGML rules are hard to explain, without bringing any corresponding
benefit.  The saving of two bytes would be worthwhile only if it
caused no confusion to users.

>	If this omission really was intentional, how might this
>be amended in the next revision?

First, I think you will need to arrange to have me killed, since allowing 
unquoted attribute values is one of those things which will only happen 
over my dead body.

More seriously, you have done already what is necessary to make it
possible to introduce the change you suggest; when the time comes for
a version 1.1 or 2.0 of XML, the Working Group responsible will, in 
the nature of things, consider proposals for changes to XML syntax,
including those recorded in the archives of the xml-editor list,
to which you addressed your note.  I hope that that Working Group will
reject the change, since I think it not a good idea, but the decision will
rest with that WG, not with me.

>	Thank you in advance for your consideration,

Thank you for your comment; I am sorry I am unable to be more enthusiastic
about your suggestion.

C. M. Sperberg-McQueen
Received on Saturday, 28 August 1999 12:07:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:29 GMT