Re: unclear point in html 4.01 spec - radio buttons

On Mon, 27 Sep 1999, Ambrose Li wrote:

> However, the standard does not specify what the user agent should
> do if the HTML does not specify a default "on" button, i.e., when
> all the buttons defaults to "off".

There is much confusion around this in HTML specifications.
See http://www.hut.fi/u/jkorpela/forms/choices.html#app and
http://ppewww.ph.gla.ac.uk/%7eflavell/www/testradio.html

As regards to authors, the only sound principle is to provide
a default for any set of radio buttons. But as regards to
specifications, the situation is more complex. The simplest
approach would be to require that a default be specified in the
markup. (This of course cannot be made as a formal syntactic
requirement in the DTD but only in the prose part.) What about
error processing? The HTML specifications generally do not impose
requirements on error handling. This need not be an exception.
In fact, recommending or even mandating some particular processing
could be taken as a semi-official permission to continue bad practice
in authoring. It would perhaps be appropriate to make an informal note
(or warning) that browsers differ in this area.

Unfortunately, HTML 4.01 seem to have taken the opposite position in the
sister problem, SELECT with no option preselected:
http://www.w3.org/TR/html40/appendix/changes.html#h-A.1.1.12
On the other hand I cannot see any change in 17.6.1, and I cannot
see what "17.3.2" refers to (except that in HTML source it has HREF
http://www.w3.org/TR/html40/interact/forms.html#h-17.6.1 !). Perhaps
http://www.w3.org/TR/html40/interact/forms.html#h-17.13.2 which says
(differently from HTML 4.0):
"When no options are selected, the control is not successful and neither
the name nor any values are submitted to the server when the form is
submitted."
 
> Right now, Netscape actually will display all-off buttons, but
> lynx will arbitrary display the first button as "on".

It's not arbitrary; it complies with the requirement in RFC 1866
(HTML 2.0).
 
> And it's disheartening that there are actually sites using radio
> buttons where the "all-off" case is in fact valid input :(

And specifications which contain such examples. :-(
http://www.w3.org/TR/html40/interact/forms.html#h-17.4.2

-- 
Yucca, http://www.hut.fi/u/jkorpela/ or http://yucca.hut.fi/yucca.html

Received on Tuesday, 28 September 1999 06:59:24 UTC