Survey of URI implementation work (was: Re: Follow up...)

[ Martin Hamilton wrote: ]

} Leslie Daigle writes:
} |                                                             Disbanding
} | the URI group would fail to reflect the fact that there is a LOT of
} | URI work left to be done.  However, it might just come to something as
} | drastic as that if this working group is not succeeding in making any
} | progress on that work.
} How are we doing on _implementations_ of the URI proposals ? :-)

Well, I don't keep the definitive list, but here's a few
things I'm aware of (mostly our own, for the simple reason
that these are the ones for which I write the cheques... =8-0  ):

We've started shipping the first release of DIGGER, our
freely available WHOIS++ implementation, last week (watch
for the T-Shirts in Stockholm... ;-) WHOIS++ uses an
attribute/value- based data model, has support for
distributed indexing, and our code ships with a WEB front
end, maintenance tools and Oracle and mSQL database
capability. Development  of a general template-oriented
gateway capability is now underway.

There has been talk of using WHOIS++ as a testbed server
for URN->URL resolution (of course, it's not clear that
this is the best way to go but experimentation is
definitely warranted). I don't know what else, if anything,
has been done elsewhere on testbeds (I'm basically out of
the loop on this aspect) but invite interested parties to
look at the DIGGER code. It can be fetched via anonFTP at:


The DIGGER homepage is at:


This contains details on DIGGER, as well as pointers to
WHOIS++ development at other sites.

FYI, we _really_ need something like a standardized URN in
DIGGER. Patrik Falstrom (the implementation team leader on
the DIGGER implementation) tells me they've gone ahead and
started using MD5 checksums pending anything definite from
this group and I'm sure he'll be happy to explain more to
those who need to know. He can be reached at "".

As far as I know, the Silk prototype is the first working
example of an integrated URI tool. We use active objects
(URAs) to search, access and store information, we use
something which looks a lot like a URN to name results of
accesses, we have something that looks like a form of URC
within the URA (to provide information about the URA,
among other things) and we of course use URLs for access,
passing these to Netscape/Mosaic (and plan to eventually
support other browsers) as needed. Of course, we don't use
IETF standards yet for anything but URLs, since there
aren't any, but we hope/expect that it will support such
standards one day.

We asked the URI chair for time to provide a brief
overview of the state of this tool in Stockholm and how it
relates to our understanding of URIs but this request
appears to have been refused. I've also been told that
this work is not within the charter of this working
group(!). This decision, coupled with the refusal to support
a call for architecture documents for this group, has been
a cause for grave concern to myself and others here at
Bunyip and I hope/expect to post a separate message on
this in the near future.

Be that as it may, Silk has been in beta for the past
month or so, being distributed to those who've requested
it, and we've scheduled the general release for next week
before Leslie Daigle (the Silk project team leader) heads
off to Stockholm. Details on the release, when it happens,
will be available at:


As a final FYI on Bunyip URI-related activity, I'm told by
the archie support group that by next week they will have
added a URL output format to the archie telnet frontend.
I'm not sure how long this will take to go through testing
and reach the customers, but it shouldn't take all that
much time. I expect that if we don't get some form of URN
standard soon, we'll define our own and put something out
supporting that, as well.

Outside of Bunyip, Michael Mealling has been working on
URNs and URN to URL resolution and has demoed a modified
version of Mosaic which provides the user the capability
of clicking on a URN, which is then resolved and accessed
automatically (I seem to recall he added a URN primitive
to HTML, then added the code to call a server for the
resolution step into the client). This was only proof of
concept, be clearly shows the utility URNs in this
application. This was a while ago, and I don't remember
the status of this work.

Michael also sent out a message last night concerning his
mods to BIND to allow using DNS to provide URN resolution.
Some experimentation is obviously underway.

For those who don't have it in their overflowing inbox,
Michael's email address is "".

Bill Armes at CNRI did a presentation at the Danvers IETF
on the work he has done on Handles. Check out the minutes
for details.

There's a brief snapshot of the things that popped into my
head. I'd love to hear about other implementations. We'd
behappy to put up a home page pointing to the various
projects, if there's interest in this.

					- peterd

BTW, I have been in and out of the office for much of the
past two weeks, so am way behind in my mail. If you've sent
a message and are waiting for a reply, my apologies. I'm
doing my best to get to the 600 pending messages.


  ...there is reason to hope that the machines will use us kindly, for
  their existance will be in a great measure dependent on ours; they will
  rule us with a rod of iron, but they will not eat us...

                                               - Samuel Butler, 1872

Received on Friday, 7 July 1995 13:56:27 UTC