Re: draft-holtman-http-safe-00.txt

> Quoting from section 5.2 of [2] (draft-ietf-html-i18n-05.txt):
> 
>    The best solution is to use the "multipart/form-data" media type
>    ^^^^^^^^^^^^^^^^^
>    described in [RFC1867] with the POST method of form submission.
>                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> 
> Do you disagree with draft-ietf-html-i18n-05.txt, or with my
> interpretation of it?  I can't really tell from your comments.

Yes, I disagree with <draft-ietf-html-i18n-05.txt> and I did so
more than once on the HTML WG list.  I'd rather not discuss it again.

> I have not read all of draft-ietf-html-i18n-05.txt, so I may be
> missing something, but the quote above seems quite clear.  POSTs are
> the route draft-ietf-html-i18n-05.txt seems to be taking, and
> draft-ietf-html-i18n-05.txt is approved as a proposed standard.  

For the particular problem they were discussing, yes.  That particular
problem only occurs under ALL of the conditions I described.  What i18n
did was try to find a theoretically cleaner solution to a problem that
has already been solved in a number of ways in practice.  What they should
have done is just recommend using the existing methods of handling
charset-varying input until such time as there was a new method for GET+body.

> My proposal attempts to identify and clear away an obstacle to the
> deployment of draft-ietf-html-i18n-05.txt.  If you can convince me it
> does not, I will retract my proposal.

What obstacle?  Having the user hit the OK button is an inconvenience.
Given that the conditions under which a charset-sensitive input dialog
is needed is ridiculously low (and no, this has nothing to do with
western vs non-western -- it is only when the input text is in a different
charset than what the server would expect, the charset of the form),
I just don't consider it a worthy justification for yet-another HTTP
control field. [And if it was, the browser community would implement it
long before the standards committee, regardless of other HTTP requirements.]

I don't mind the proposal -- just be sure the justification for it
matches the actual need and doesn't invent problems that don't exist.


 ...Roy T. Fielding
    Department of Information & Computer Science    (fielding@ics.uci.edu)
    University of California, Irvine, CA 92697-3425    fax:+1(714)824-4056
    http://www.ics.uci.edu/~fielding/