Re: Globalizing URIs

Roy Fielding (
Mon, 14 Aug 1995 19:37:44 -0400

Message-Id: <>
To: Peter Deutsch <>
Subject: Re: Globalizing URIs 
In-Reply-To: Your message of "Mon, 14 Aug 1995 17:17:37 EDT."
Date: Mon, 14 Aug 1995 19:37:44 -0400
From: Roy Fielding <>

>[ Roy Fielding wrote: ]
>.  .  .
>} I'd just like to point out that most of this discussion is ignoring
>} the fact that the transcribability of URLs is one of the main reasons
>} the WWW has been successful in building an information base.  Although
>} it may now be possible to add persistent naming to that base, it would
>} not have been possible to create such a base with non-transcribable names.
>Actually, I think this is ascribing a disproportionate
>part of the Web's success to the transcribability of its
>URLs. Yes, URLs are relatively more human friendly than
>ISBN numbers (and I agree that this is a "Good Thing
>(tm)"), but I don't think that this transcribability was
>greatest driving force behind the Web's success.

I didn't say that.  What I said is that it was a necessary
prerequisite.  Had we started with URNs (requiring a name server
to exist before you could test the links in a document, and requiring
the contents of that server to propagate to your intended readers
before they can access the resource), the WWW would never have gotten
off the ground.  Similarly, if setting up a web server required a
$50,000 machine and root access, the WWW would never have gotten
off the ground.  And, if the only mechanism for relating documents
via hypertext required a proprietary document format or a complex
editing environment, the WWW would never have gotten off the ground.

So, now the WWW is sufficiently important that all of the above
are justifiable to be implemented or even standardized.  Great!
That does not mean URLs were wrong -- they were (and are) a necessity.
It only means that we now have a new set of design requirements
that we need to fulfill, and URNs are the answer.

>Conversely, I think it's a safe bet that more than one
>person trying to access the URL ""
>will be burned by typing "...people/fielding". After all,
>upper case shouldn't matter, right? 

Why?  Upper case certainly matters for my name.  I prefer all-lowercase
URLs myself, but that is a detail left to the server.  For example,
there is nothing preventing a server from doing approximate name-matching
when it can't find the case-sensitive mapping.  The only thing that
requiring case-sensitive names does is prevent you from using existing
technology (e.g., filesystems) as a leveraging mechanism in the absence
of a name translator.

 ....Roy T. Fielding  Department of ICS, University of California, Irvine USA
                      Visiting Scholar, MIT/LCS + World-Wide Web Consortium
                      (                (