W3C home > Mailing lists > Public > public-html@w3.org > January 2009

Re: Who is the Intended Audience of the Markup Spec Proposal?

From: Ian Hickson <ian@hixie.ch>
Date: Tue, 27 Jan 2009 04:06:46 +0000 (UTC)
To: "Michael(tm) Smith" <mike@w3.org>
Cc: Lachlan Hunt <lachlan.hunt@lachy.id.au>, public-html <public-html@w3.org>
Message-ID: <Pine.LNX.4.62.0901270345190.952@hixie.dreamhostps.com>

On Tue, 27 Jan 2009, Michael(tm) Smith wrote:
>
> I well understand the concern about negative side effects that having 
> one official schema can have. But the fact that the "Content model" 
> sections of the spec currently use a particular schema formalism is not 
> an essential feature of the spec.

I've nothing against formalisms being used non-normatively, since they 
sometimes can be useful for making things clearer than prose. However, 
normatively they are only useful, IMHO, if they are comprehensive, and a 
comprehensive formalism would be incomprehensible if used to describe 
HTML5's rules.

Right now, the draft in question has both examples of the formalisms being 
incomprensible as well as exampels of the spec not being comprehensive.

An example of incomprehensible in the current text is:

browsing-context-or-keyword = ()|([^_].*)|(_[bB][lL][aA][nN][kK])|(_[sS][eE][lL][fF])|(_[pP][aA][rR][eE][nN][tT])|(_[tT][oO][pP])

This is not going to help most of your target audience understand what 
they can do with target="". I wrote the corresponding prose and it still 
took me a few seconds to work out whether that was right or not. (It's 
not, as far as I can tell. "()" shouldn't be a valid alternative, unless 
I'm misunderstanding the formalism.)

An example of not being comprehensive is the "assertions" section of 
several elements (e.g. <header>) or the definitions of the mediaquery or 
date productions.

I think being comprehensive in this instance would be harmful for the 
reasons Henri gave, so I don't recommend going down that route. Moreover, 
if the document is normative and uses formalisms, then the current 
language used will likely be found to be unsuitable, and thus I would 
recommend finding a better language to use (or developing one specifically 
for this purpose).

I think it'd be fine to be non-normative and still use incomplete 
formalisms. However, that would mean not publishing as a REC track 
document, as discussed by Maciej. I think that's fine too. (The second and 
third clauses of the audience section would have to change, since such a 
document wouldn't have conformance statements.)

(The other option I see, namely being normative and not using formalisms 
while still addressing the same audience, is what the HTML5 spec is 
already doing, so doing that seems like a duplication of effort.)

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Tuesday, 27 January 2009 04:07:22 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:28 GMT