Re: URA as HTML

Peter Deutsch (peterd@bunyip.com)
Wed, 22 Mar 1995 13:24:41 -0500


Message-Id: <9503221824.AA18468@expresso.bunyip.com>
From: Peter Deutsch <peterd@bunyip.com>
Date: Wed, 22 Mar 1995 13:24:41 -0500
In-Reply-To: Leslie Daigle's message as of Mar 22, 13:05
To: ura-bunyip@bunyip.com, uri@bunyip.com
Subject: Re: URA as HTML

[ Leslie Daigle wrote: ]

} 
} [terry@ora.com, Wed Mar 22 10:54:39 1995, wrote:]
} 
} > A question in aid of clarification:  we would seem to have URAs 
} > already in the form of HTML documents.  Do you consider HTML docs
} > as URAs?  if not, why not?
.  .  .
} 	2. Having said that,  I would personally be unlikely to suggest HTML 
} 	   documents as an implementation mechanism for these reasons:

<reasons deleted>

I concur with Leslie's assessment. Put another way, one of
the major defining characteristics which separates URAs
from URNs/URLs is their programable nature. Until HTML is
Turing-complete it will lack the horsepower needed to
fully accomplish this task (and once HTML _is_
Turing-complete we can go over how we all feel about the
concept of "creeping featurism"...  :-)

It is our goal with URAs that in the long run we hide from
implementation details the user. Among other things, this
means that we want to be able to hide the selection of
programming language, services accessed and so on.
Provided we build them correctly, implementations will be
able to select the appropriate combination of
version/language/etc, ship URAs to appropriate proxy
servers where appropriate, etc. If the developers of HTML
make it rich enough, it will be a candidate implementation
language. I think architecturally, forcing _one_
implementation language would be a major mistake.


					- peterd


-- 
----------------------------------------------------------------------

   I hate my silly computer, I think I'm going to sell it. It never
   does what I want it to, only what I tell it...

					- anonymous
----------------------------------------------------------------------