Re: Mailserver URL proposal

Dirk Herr-Hoyman (hoymand@gate.net)
Tue, 26 Jul 1994 09:24:05 -0400 (EDT)

From: Dirk Herr-Hoyman <hoymand@gate.net>
Message-Id: <199407261324.JAA233582@inca.gate.net>
Subject: Re: Mailserver URL proposal
To: berg@pool.informatik.rwth-aachen.de (Stephen R. van den Berg)
Date: Tue, 26 Jul 1994 09:24:05 -0400 (EDT)
In-Reply-To: <9407251112.AA07032@tabaqui> from "Stephen R. van den Berg" at Jul 25, 94 01:12:06 pm

>
> Marc VanHeyningen <mvanheyn@cs.indiana.edu> wrote:
>
> I'm not sure how much I missed about this discussion, but, here are
> some things that I had in mind:
>
> 1. Having read the mailserver proposal I got the impression that you intend
> for the server to actually send the mail. I think this definitely is the
> wrong way to go. The client should be calling sendmail (or whatever),
> this also resolves the problem of somehow passing the sender address, since
> it will be generated by sendmail (who is actually the only one capable of
> deciding what it should look like).
>
The client (or whatever agent is using this URL) would form a mail
message and then pass it onto a mail transport agent, such as sendmail.
This is really no different than what a typical mail user agent (such as
elm or eudora) does.

> 2. The mailto: URL already allows for arbitray headers to be defined.
> What about implementing the "mailserver" URL like this:
>
> <FORM METHOD="POST" ACTION="mailto:berg@pool.informatik.rwth-aachen.de">
> <MH HIDDEN>Subject: WWW Questionaire</MH>
> <MB HIDDEN>
> This will be in the body
> of the mail
> </MB>
> <P ALIGN=CENTER><INPUT TYPE=submit>
> </FORM>
>
>
Not quite Mark. This is html+, which is specific to the 2nd generation
WWW clients. This works fine in that context. But, we are designing a
general standard here, not WWW. Peter D's comments on this matter are
exactly my own.

-- 
Dirk Herr-Hoyman                  |                      
Lake Worth, Florida, USA          |
hoymand@gate.net                  |              
+1.407.540.8309                   |