Re: "Mailto" Command

Jukka Korpela (jkorpela@cc.hut.fi)
Tue, 25 Mar 1997 08:30:24 +0200 (EET)


Date: Tue, 25 Mar 1997 08:30:24 +0200 (EET)
From: Jukka Korpela <jkorpela@cc.hut.fi>
To: www-html@w3.org
Subject: Re: "Mailto" Command
In-Reply-To: <3.0.32.19970324082124.00893c60@pop3.interaccess.com>
Message-ID: <Pine.OSF.3.96.970325075312.28343A-100000@alpha.hut.fi>

On Mon, 24 Mar 1997, Harold A. Driscoll wrote:

> At 10:49 24/3/97 +0200, Jukka Korpela wrote:
> >As regards to the evolution of HTML, I think there is clear need
> >for a standardized method for specifying that a form be sent
> >as a mail (RFC 822) message to a given address or list of addresses.
> 
> Such a proposal currently exists, as an Internet-Draft
> <draft-hoffman-mailto-url-00.txt> [1].

As far as I can see, that draft discusses the modification of
the definition of mailto URLs. (As an aside, I don't think it's
so great an idea to standardize the ?Subject thing, even if it
is made as part of a more general extension. People want to use
something like ?Subject in order to have an indication of where
the mail comes. But is the Subject heading the right place for
such _origin_ information?)

My note was about sending form contents as mail messages. That is,
what should happen if the ACTION in a FORM element specifies a
mailto URL? The HTML 3.2 spec is not clear about this; 
http://www.w3.org/pub/WWW/TR/REC-html32.html#form
mentions such situation but does not specifically _require_ that
it should work, and it seems to me that popular browsers do not
seem to have wide support to it; see Notes at
http://www.hut.fi/~jkorpela/HTML3.2/5.25.html

RFC 1867, section 5.6, says:
'Independent of this proposal, it would be very useful for HTML
interpreting user agents to allow a ACTION in a form to be
"mailto:" URL.'

And it seems to me that neither HTML 3.2 nor any other specification
has made this a _requirement_.

Moreover, as far as I can see, application/x-www-form-urlencoded
is the only ENCTYPE which is required to work (implicitly, by its
being the default). Wouldn't it be good to have some other encoding, too,
for which support is required and which encodes the form contents
in a manner which is reasonably legible as such (in addition to
being formally defined, to allow automatic processing)?

> C'mon, reality check, please. The syntax of email messages is both
> particularly easy to parse, as well as very well know as an open standard.

Well, easiness depends on the user; I suppose most Web page creators
who would like to use forms have no idea whatsoever about parsing
RFC 822 headers. And I was not suggesting any modification to RFC 822,
of course. Although a solution _might_ involve using special headers
(in the RFC 822 framework), the most essential thing is the syntax
of the message _body_, which should reflect the structure of the form.

The fact is that a lot of people want to put feedback forms onto their
Web pages, expecting just to read the feedback as part of their normal
incoming mail flow. (Some of them might later wish to have some automatic
processing, but that's another story.) I think that this should be clearly
supported in a standardized manner.

Yucca, http://www.hut.fi/~jkorpela/