Re: INPUT enhancements - was Re: SELECT NEST proposal

nemo/Joel N. Weber II (
Thu, 13 Mar 1997 00:46:37 -0500 (EST)

Date: Thu, 13 Mar 1997 00:46:37 -0500 (EST)
Message-Id: <>
From: "nemo/Joel N. Weber II" <>
In-reply-to: <> (message from Simon Cox on Wed,
Subject: Re: INPUT enhancements - was Re: SELECT NEST proposal

   Date: Wed, 12 Mar 1997 10:44:28 -0500
   From: Simon Cox <>
   Organization: CSIRO Exploration & Mining
   X-Mailer: Mozilla 3.0 (Win95; I)
   MIME-Version: 1.0
   References: <>
   Content-Type: text/plain; charset=us-ascii
   Content-Transfer-Encoding: 7bit
   X-Mailing-List: <> archive/latest/7596
   Precedence: list

   I suggest <INPUT TYPE="range" NAME=string>
   but there are quite a few attributes that are needed, at least:  
   1.  variable type - float, int, ?alpha

I can't think of any places where alpha would be useful.

I'm also not sure if float is nessisary.

Of course, if someone sees counterexamples, I'm willing to reconsider that

So maybe we can leave that attribute off?

   2.  limits - min & max, and

I agree this is needed.

   3.  granularity/quanta

I'm not sure if this is needed.  But I think it is.  Does anyone
have an example of where it's useful?

   4.  selector type - slider|dial - though perhaps this is a client issue

If we can do that with CSS, I think that would be prefered.

We'll probably want to add pseudoclasses related to sliders at that point.

   5.  selector scaling - linear (default), logarithmic, ?reciprocal,
   ?square etc

I'm not sure this is nessisary.

The Richter (sp?) Scale used for measuring earthquakes is logarithmic.
We tell people it was a 7.2 earthquake, and we hope that they're clueful
enough to understand that.  Most people don't realize that's ten times
stronger than a 6.2, but that's their problem.

The stereo volume is analogous.  I have a radio I'm listening to right
now with a 0-10 volume control.  I've never considered whether or not
it's logarithmic, and I don't really care.  All I know is that I keep
it below 5 unless I'm planning to listen to it someplace other than my

So I favor not providing any logatrithmic features.

   6.  discontinuous ranges?


10% of the functionality will do the job in 99% of the cases.  I'm not
convinced that offering every concieveable possibility is useful.

For this to become really useful, you have to convince browser writers
that they want to support it.

   Introducing these all as additional attibutes for INPUT begins to
   overload it 
   somewhat.  Suggestions?

Clean addition of attributes is something I don't object to.  My
feeling is that I want every attribute to have exactly one meaning,
regardless of what tag it's on.  For example, the CLASS attribute
can be handled by the same code for nearly every tag in existance.

   (NB  I think that "alt" may not be valid within INPUT, but I _always_
   to give alt's for images, and I think this may have been an oversight?)

Why would alt be useful for INPUT?