Change in Direction on the Internet Atlas Project

[Carl Malamud <]

Dear RIPE Colleagues:

In January, 1998 I briefed the RIPE membership on a project proposal
called "the Internet Atlas."  This project's aim was to develop tools
for producing visual representations of the Internet topology for use in
navigation, resource location, and other user functions.  The nonprofit
research foundation that I chair, the Internet Multicasting Service, had
finished its last big project, the world's fair, and we managed to put
the last nail in the coffin in our ongoing struggle with the U.S. Patent
Office earlier this year by resolving the question of online trademark
and patent documents.  

My hope since mid-1996, when the fair's organizational work was winding
down, was to revive the Internet Multicasting Service and have the new
focus of the organization be this Atlas project.  As part of that
commitment to carry out the Atlas, I moved Internet Multicasting up to
the MIT Media Lab where I spent a year as a visiting scientist.  I then
moved on to Japan's Keio University in 1997 and finally made a concerted
effort throughout 1998 to make this project work in Amsterdam.

After briefing the RIPE membership in January, I set to work trying to
build a new institute in Amsterdam.  We incorporated a Dutch foundation,
rented office space, and spent a considerable amount of time trying to
get enough momentum to get the project underway.  The RIPE NCC was
extremely helpful in this regard, assisting in procuring computers,
getting health insurance, a work permit as a visiting scientist, and a
myriad other details.

The RIPE NCC was so supportive that the board pledged 100,000 ECU
towards the Atlas project.  I was hopeful that with such an initial
commitment, I could drum up support from local government, local
telecommunications companies, and others with an interest in seeing a
fairly large-scale new Internet research effort based in Amsterdam.  To
show that we were serious, I went to many of my past sponsors and they
indicated that they would be pleased to join up if the effort became
solid.  

<flame on>
Unfortunately, we just never got any local interest going.  I was never
able to get a dialogue started (indeed was not even received in many
cases) with the city, the national government, or the EC.  The concept
of an "applied research effort" with a concrete, short-term goal seemed
quite foreign to a group that I would charitably describe as
Bangemann-challenged bureaucrats.  The telcos were even worse.  My
experiences with the PTT were characterized by extremely high prices
($1600 per month for an E1 line is at least 500% higher than would
qualify for the label expensive) and indifference (the lines I did get
never did work properly).  The brave new competition, characterized by
cable companies such as A2000, all priced themselves at 90% of the PTT
tariff,  still an order of magnitude higher than what I would
characterize as non-extortionate, particularly given the low ratio of
clues to ego in these lifeless corporate bodies.
<flame off>

While I could go on at length over the overall climate for promoting
major IT initiatives, I'll save this for my next book.  The bottom line
is we crashed and burned and I was unable to get the project going.  We
never touched the RIPE NCC's 100,000 ECU, since I had pledged that it
would only be used if there was a reasonable chance that things would
take off.  I also made sure that the RIPE NCC was fully reimbursed for
expenditures on items such as health insurance, Federal Express, and
other essentials.

I've not given up on the ideas behind the Atlas, but I've decided that
the nonprofit institute is just not the right incubator for these
ideas.  While perhaps it could be done, it is clear after two years of
trying that I'm certainly not going to be the one to do it.  Rather than
give up, we've come up with a plan B, a new for-profit corporation
called Invisible Worlds, which will produce a mapping and visualization
service for the Internet.  Invisible Worlds is a spinoff from the
Internet Multicasting Service, just as universities and other research
institutes often help give birth to companies.  As many of you know, the
NLnet was formed in this way, as were several other significant groups
on the net.  Professional estimators are being used to determine the
value of the previous development of the concept by the Internet
Multicasting Service.  The for-profit corporation will pay full market
value for any such intellectual property, and the proceeds will be
placed in a long-term trust for open-source community projects.  We are
being advised by the accounting firm of Ernst & Young and by the law
firm of Baker & McKenzie to complete this transaction properly and file
it with the relevant governmental authorities.  

As a company, we're still committed to concepts such as open source and
community standards and hope that our efforts will result in something
real that you'll be able to use.  Invisible Worlds is just a startup, so
there is no guarantee that this will be any more successful than past
efforts to build the Atlas.  However, I'm heartened by the progress made
in the last month since I moved back to Silicon Valley.  Marshall Rose
has accepted a major development contract with the company as  Chief of
Protocol, Dr. David Clark has agreed to chair the technical advisory
board, and Dr. Steve Wolff has joined our board of directors.  An
award-winning design team, Becky and Larry Pranger, join me as founders
and our distinguished panel of advisors on content and design, known as
Invisible Fellows, will be chaired by composer Tod Machover.  Overall,
it is a good start and I'm hoping we'll be able to get some real results
out the door in good time.

While I did not succeed at building an institute in Amsterdam, I must
say it was an extremely rewarding sabbatical.  Watching the RIPE NCC
function at first hand gave me a new appreciation for what is without a
doubt the best registry operation in the world.  I only wish the U.S.
system was half as rational as the European system for the management of
registry functions.  In addition to the registry, the RIPE NCC is
playing a crucial role in keeping global functions such as the IANA
structured in a manner that will actually work.  

In addition to the RIPE NCC, I saw many other top-notch operations. 
NIKHEF was extremely helpful and even volunteered to assist us in
finding a facility for the institute should it get funded.  NLnet
corporation is a great ISP and NLnet the non-profit is going to make a
huge, long-lasting contribution to the Internet infrastructure through
their foundation.

Special thanks are due to Rob Blokzijl and Daniel Karrenberg, whom I'm
proud to say are both friends and colleagues.  They were extremely
supportive and gave unstintingly of their time.  If the City of
Amsterdam ever got serious about doing something real about the
information highway, they should fire their platoons of consultants (who
never seem to get beyond planning menus for program committee meetings)
and ask Rob and Daniel what to do.   RIPE is very lucky to have both of
them.

As always, I look forward to continuing to work with my colleagues
throughout Europe and I'd like to thank the membership of RIPE for
giving me the chance to get to know you better and for your initial vote
of confidence.

Carl Malamud
Invisible Worlds, Inc.
http://not.invisible.net/