Date: Thu, 2 Dec 93 17:07:09 +0100
From: Tim Berners-Lee <timbl@www3.cern.ch>
Message-Id: <9312021607.AA00832@www3.cern.ch>
To: uri@bunyip.com
Subject: Minor clarification on url: prefix
>From: hoymand@joe.uwex.edu (Dirk Herr-Hoyman)
>I believe I heard another point
>
>For: Don't need to know EVERY URL access method.
>
>Against: Clients like Mosaic will rip off the url: prefix
> and look at the access method anyways.
>
>The only area I see this being really useful is in plaintext, like
e-mail
>messages. But, I thought a <> wrapper had been proposed for this
purpose
>anyhow (is this still in?)
It is in as suggested practice for delimiting URLs within
plain ascii text (ie not SGML or RFC822 headers or any other markup)
Voted in in Amsterdam.
> Trying to find URIs in plaintext is going to be
>an inexact science no matter what, since the syntax is already
complicated
>and it's REAL people typing these in.
>If we were to have a <> wrapper,
> then doing <URL ftp://myhost/dir/file>
>would be better, since it's SGML parsable.
It isn't SGML parsable.
<SOMETHING URL="ftp://myhost/dir/file">
is SGML parsable, as the value is quoted
and attached to a suitable attribute value.
This is quite sensible, but noone surely would
suggest that the URL=" should be considered part
of the URL.
How about suggesting that in an RFC822 header context
that a specific header be used for URLs?
Newsgroups: ietf.neverending.uri
URL: ftp://myhost/dir/file
Date: 1997 Mar 31 21:67
but leave the URL syntax unencumbered.
Tim