Re: HTML forms

Dear James,

> I understand that.  There might be substantial benefits from
> reconsidering those opinions.  Within the IETF, public debate
> is assured on almost all controversial matters.  The W3C,
> however, constrains meaningful debate to those willing and able
> to pay US$50,000 per year.  

That is not true, on a variety of counts. I'll name two.

First, membership has two levels: full and affiliate. For
more details, please refer to:

Second, people who demonstrate an interest in W3C work and
are constructive contributors to W3C work may become invited
experts. Invited experts provide critical resources - time
and effort. They do not pay fees. 

If you review the staff comment to your original note which
was amended at your request, you can read the opinion of the
working group. The URI for the Staff comment on the July
1999 Form Upload W3C Note is:

The critical bits:
Addendum - 3rd March 2000

The participants of the HTML Activity have reviewed the
proposal for possible use in their work on forms. 

Device Upload: The HTML Working Group's position

The following statement is from the chair of the HTML
working group:

HTML is intended to be a device-independent markup; this
means that any proposal that suggests markup that includes
the word "device", especially if there is no fall-back
mechanism, should ring alarm bells.

This is the fundamental problem with the device upload
submission (without going into detailed criticism).

To take an example from the submission

<INPUT name="picture1" type="file" device="camera"

What should happen if the user doesn't have a camera, but
does have a photograph and a scanner?

What if there is a camera, but not connected to the
computer, so that the user has to take a picture, download
it, and then upload it from a file?

These differences shouldn't be visible in the markup, but
should be abstracted out into the *intent*, so that a user
agent can then offer whatever is available on the client
machine in question.

In other words, the markup should state that an *image* (or
a movie, or a sound, or whatever) is required, not a camera,
and the user agent, seeing that an image needs to be
supplied can then offer whatever facilities are available
for delivering an image (scanner, camera, filestore, TV
card, ...)

In fact this is exactly how the markup works currently in
HTML 4 ( For instance, you can

     <INPUT name="picture1" type="file" accept="image/*">

It should then be up to the user agent to offer to let you
use the camera or scanner, as available, or the filestore.

Apparently the browser manufacturers have mistaken the
attribute value "file" to mean "something from the
filestore", rather than "something to be packaged and sent
as a file". A possible solution to this could be a
submission in the form of a note on recommendations to
browser manufacturers on how to implement the user interface
to file upload.


James, you are clearly passionate about your work. It is
also important that you listen to the feedback you have
received over the last six months. People have provided you
with a thoughtful technical evaluation of your proposal.
Disagreement is not a justification for disrepectful
treatment or false accusations.


Janet Daly

Received on Thursday, 30 March 2000 23:51:27 UTC