Re: Project Gutenberg's Roget's Thesaurus

Guido.van.Rossum@cwi.nl
Message-id: <9304140819.AA26213=guido@voorn.cwi.nl>
To: Bill Janssen <janssen@parc.xerox.com>
Cc: timbl@nxoc01.cern.ch, www-talk@nxoc01.cern.ch
Subject: Re: Project Gutenberg's Roget's Thesaurus 
In-reply-to: Your message of "Tue, 13 Apr 1993 15:16:03 MDT."
             <AfmniXoB0KGWMgGNIE@holmes.parc.xerox.com> 
From: Guido.van.Rossum@cwi.nl
X-Organization: CWI (Centrum voor Wiskunde en Informatica)
X-Address: P.O. Box 4079, 1009 AB  Amsterdam, The Netherlands
X-Phone: +31 20 5924127 (work), +31 20 6225521 (home), +31 20 5924199 (fax)
Date: Wed, 14 Apr 1993 10:19:01 +0200
Sender: Guido.van.Rossum@cwi.nl
>> I see a problem coming here: how does an unreplicated document (say my
>> own home page) make a reference to such a replicated document?  If I
>
>We're trying out a system at PARC that has a similar problem.  Our
>tentative solution, translated to URL terms, is that one particular
>scheme of URL is a list of URL's, any of which can be used to access the
>document.  This would allow the client to use whichever copy of the
>document it ``decides'' is closest/cheapest.  We also have different URL
>schemes that allow us to say, ``look up this UID in a name-service of
>this type'' -- indirection.

A list of URL's might work in some cases, but in general this means
that adding a new mirror site requires updating all references...

A name service sounds like a better solution -- bot how to integrate
this in HTML or HTTP?  And who's going to run the name service?  Or
could we use some existing name service?  (Perhaps DNS is flexible
enough to return a list of hosts for certain names, from which the
client could choose?  This would be ideal.  Anyone know anything about
DNS?)

--Guido van Rossum, CWI, Amsterdam <Guido.van.Rossum@cwi.nl>