to follow up on what Earl Hood said:
References:
<199808181924(_dot_)QAA12032(_at_)ecology(_dot_)bio(_dot_)dfo(_dot_)ca>
Date: Wed, 19 Aug 1998 09:29:42 -0700
From: Earl Hood <ehood(_at_)hydra(_dot_)acs(_dot_)uci(_dot_)edu>
X-md5sum: bb3961426450a1a18ccd3d1e11772acd
X-md5sum-Origin: ex1.ncsa.uiuc.edu
Sender: owner-mhonarc(_at_)ncsa(_dot_)uiuc(_dot_)edu
Precedence: bulk
List-Name: mhonarc
On August 18, 1998 at 16:24, Bill Silvert wrote:
I'm a bit of a novice, and couldn't find the answer in the documentation,
so forgive my ignorance. I would like to convert email addresses in the
text into mailto: url's the way it is done in the headers, but it
appears that although MHonArc does it fine in the header, I couldn't
find a way to have it parse the message other than manually going in and
adding mailto: before each address. Is there an argument or resource
that can be used to force this behaviour?
.. And Earl Hood replied:
No. The problem is that it would conflict with message-id detection
(syntactically, there is no difference between message-ids and mail
addresses).
I'm confused. Here are two chunks of suspicious text:
<199808181924(_dot_)QAA12032(_at_)ecology(_dot_)bio(_dot_)dfo(_dot_)ca>
<ehood(_at_)hydra(_dot_)acs(_dot_)uci(_dot_)edu>
You can create a hyperlink to the message with ID as given in the
first one because
a) you can see from the syntax that it might be a message-ID
reference.
b) you can see from the index that there is an archived
message in the archive with that message-ID
When you encounter <ehood(_at_)hydra(_dot_)acs(_dot_)uci(_dot_)edu> in the body
of a
message it looks syntactically just like a message-ID. Ah hah,
I'll go look in the index. No such message.
No <a href="http://path.to.archive/page";> is generated. What if
these guys were to fall through to <a href="mailto:etc.>?
I don't see how generating mailto: links conflicts with linking
to referenced messages, because without loss of valid mailto:'s
you can try on the hypothesis of message-ID first. When that
fails, you can go on the hypothesis of mailbox. Reasonable sites
don't generate message-IDs that duplicate their mailboxes, so
there is no conflict. You only need to treat something in
mailbox syntax as a message-ID when you can confirm that a
message with that ID exists. You'll generate some bogus
mailto's. That should keep the spammers busy handling bounces.
But these will be obvious to the naked eye and real people won't
click on 'em. And you won't miss any prior-message hyperlinks
and you won't miss any mailto:<person with mailbox spelled out in
message>.
People who wish to be pedantic can be unambiguous in the body of
a message by using <mid:message-ID> and <news:message-ID> forms,
BTW.
No?
Al
