Re: Enhancements to Forms - another loss??

Simon Cox (simon@ned.dem.csiro.au)
Wed, 12 Feb 1997 08:43:14 -0500


Message-ID: <3301C8F2.1B52@ned.dem.csiro.au>
Date: Wed, 12 Feb 1997 08:43:14 -0500
From: Simon Cox <simon@ned.dem.csiro.au>
To: "Chuck D'Antonio" <c_dantonio@harvard.edu>
CC: www-html@w3.org
Subject: Re: Enhancements to Forms - another loss??

Chuck D'Antonio wrote:
> 
> I'd suggest working these into the CSS framework rather than extending
> the space of form tags to include specific interface widgets.  This is
> partly an impulsive reaction since I'm strongly in the "HTML is not a
> presentation language" camp; 

I guess I would generally join you in that camp.  However, I 
am not so sure that your conclusion follows from this.  
Looking at my three enhancements

1. I don't see that structured (nested) menus imposes any 
particular presentation expectations additional to what is  
implicit in the existing menu syntax - its just that selecting 
_some_ items from a list will reveal a further list, rather 
than actually selecting the first chosen;

2. For sure, sliders and dials are presentational widgets, 
but they should just be a browser's implementation of a 
"choose integer|float from range (min,max[,granularity])" op. 
to complement "select item from list", which is already provided;

3. The current single-click imagemap is a restricted case of 
multi-click imagemaps.  This one certainly makes no sense in 
anything other than a "conventional" graphical browser, but again, 
since the restricted case is already widely implemented and 
used, then a small generalisation is not too offensive.  

> When I think of what HTML forms lack for generic interface development
> my first thought is "validation" and then I think about widgets that'd
> be great to have.  It's interesting that you think of them as needed
> less than other widgets, though I guess it makes sense since sliders
> and dials, etc. can perform some rudimentary validation 

I would contend that there is *substantial* implicit validation 
provided by all the form elements *except* text-entry - you can 
only choose from the choices presented!  When creating forms, 
I try to avoid text entry fields in forms in an effort to prevent 
clients making impossible requests (or making typos).  Isn't this 
one of the main strengths of GUI's in general!  (In this context, 
a choose-number-from-range operation is a *gross* ommision from the 
current forms implementation - maybe this is another example of how 
math/science use and influence on the net has waned ...)  

> I don't know where members of the W3 are in terms of developing more
> robust forms cababilities for web applications, but I'd think we could
> get something going from the outside that might make a lot of sense to
> them.  Let's give it a shot.

Let's keep talking this through.  

> Simon Cox wrote:
> >In particular I'm thinking of
> >
> >1.  cascading/nested/pull-right menus - to handle those
> >long lists of choices in a more structured way;
> >
> >2.  sliders and dials - to allow selection of a numeric value
> >from a range;
> >
> >3.  multi-click imagemaps - to allow selection of a line or
> >area as well as points;
> >
> >[4.  some client-side validity testing?  maybe less immediate,
> >but ...]

-- 
__________________________________________________
Dr Simon Cox - Australian Geodynamics Cooperative Research Centre
CSIRO Exploration & Mining, PO Box 437, Nedlands, WA 6009 Australia
T:  +61 9 389 8421   F:  +61 9 389 1906   simon@ned.dem.csiro.au
http://www.ned.dem.csiro.au/unrestricted/people/CoxSimon/