Re: HTML 3.0 FORM

Dave Raggett (dsr@hplb.hpl.hp.com)
Fri, 16 Dec 94 16:25:41 GMT


From: Dave Raggett <dsr@hplb.hpl.hp.com>
Message-Id: <9412161625.AA22046@dragget.hpl.hp.com>
Subject: Re: HTML 3.0 FORM
To: michaelj@relay.relay.com
Date: Fri, 16 Dec 94 16:25:41 GMT
Cc: www-html@www0.cern.ch

>  <!ELEMENT FORM - - %body.content -(FORM) +(INPUT|SELECT|TEXTAREA)>

> If I understand this right, this means that INPUT, SELECT and TEXTAREA would
> be allowed *anywhere* inside a FORM, which would seem to be undesirable to me

SGML doesn't really allow us to specify what is needed. The HTML+ approach
as you point out allowed form elements outside the context of a FORM element.
That works fine if you can make default assumptions as to how to deal with
the form, e.g. send it to the same URL as that used to retrieve the document.
Following Dan Connolly in HTML 2.0, I switched with HTML 3.0 to the inclusion
mechanism, which is rather a blunt instrument. Whilst it is possible to stop
people from embedding INPUT in a SELECT element etc. it is still possible to
place such fields in unexpected places.

> Would it be better to revert to the HTML+ way of doing things

Perhaps you are right ...

Can we define what the default action is for a missing FORM element?
--
 Dave Raggett <dsr@w3.org> tel: +44 272 228046 fax: +44 272 228003
  Hewlett Packard Laboratories, Filton Road, Bristol BS12 6QZ, United Kingdom