Re: Auto fill for form fields

Adam Jack (ajack@corp.micrognosis.com)
Thu, 15 Feb 1996 12:33:36 -0500


Message-Id: <31236E70.4F8B@corp.micrognosis.com>
Date: Thu, 15 Feb 1996 12:33:36 -0500
From: Adam Jack <ajack@corp.micrognosis.com>
To: "Scott E. Preece" <preece@predator.urbana.mcd.mot.com>
Cc: JHTaylor@videodiscovery.com, www-html@w3.org
Subject: Re: Auto fill for form fields

Scott E. Preece wrote:

> 
> If you don't extend the standard to define names for the items that
> can be auto-filled,
> how does the form author indicate to the browser which
> auto-fill value goes in which slot? 

IMHO it is the User Agent that should make this choice. It is a lot
closer to the user who has, or has not, configured a value for that
purpose of a per "field" basis. Placing the control in the hands of
the form provider is restrictive.

> Are you guessing based on the name of the field and the surrounding
> text? 

Yes & No. I was matching the variable "NAME" against a local database
of values & aliases. E.g. "firstname" "givenname" were aliased. (It
could learn an alias.)

> This sounds a little less predictable than I would like, 

I don't disagree -- however it is not that unpredictable. It benefits
hugely, however, in being available with all todays forms.

> Just to take the tax example, your scheme works if the form slots are
> labeled "name", "phone", etc., but fails if the slots are named 
> "line3", "line31", etc., since the semantics of the name might vary
> from year to year.  The value on line31 this year might be on line 33
> next year.

Good point. How would you suggest this were done with "autovalue"?
I guess a mapping would need to be made to "phone" etc. If this is
the case then all we are talking about is that I am overloading
the "NAME" tag. 

I agree it is overloading -- but it isn't severe. The FORM author
choses a name that allows automatic processing. One could define a
standard that all authors ought use -- but I really wonder whether
that would work in practicality.

> I don't think you're wrong, but I do think it would be safer to
> provide a base capability in the form of a set of defined names.
> A user agent would still be free to implement a learning/guessing
> scheme to augment that set.

If this is to be the case and the use is within the "NAME" tag then
there is the risk of clashing w/ an existing form that has named a
field "name" but actually  has no relevance to givenname.

I will admit -- making a new TAG, say "AUTOVALUE" and defining a
set of values seems cleanest. Unfortunately I think it will miss
its target purpose and not be manageable. I am sorry if I am
arguing through pessimism however I believe that is a fair arguement 
in this case. This is a world where proprietary extensions, such
as netscape's, thrive and clean standards, such as X.400, flounder.

Adam
--
+1-203-730-5437 | http://www.micrognosis.com/~ajack/index.html
ajack@corp.micrognosis.com -> ajack@netcom.com ->  ajack@?.???