My previous visit to Korea had been a six-hour transit stay just before the Olympics and I had been used as training material for a
half-dozen different security teams getting their procedures ready
for the onslaught of tourists and athletes. My current visit was considerably less high-strung, although there was still an awfully large
number of security personnel in evidence. It was a traffic violation
control day and each intersection had one or two cars pulled over
for traffic violations and the expressway had police vans every few
hundred meters, waiting for potential perpetrators to pass.
Wednesday morning, I took a cab out to the
Korea Advanced
Institute of Science and Technology (KAIST) to meet
Kilnam Chon, a
professor in the computer science department. KAIST is the center
of one of Korea's R&D networks, the System Development Network
(SDN) which, for some reason, is also known as Hana. SDN has
hubs at the two major KAIST locations in Seoul and at the Daeduk
Science Park.
Fifteen universities and research organizations connect to the
two hubs at speeds ranging from 9.6 to 56 kbps. A 56 kbps line to
the University of Hawaii links this TCP/IP network into the Internet. SDN supports OSI applications such as FTAM and X.400,
but usage was declining rather than increasing. Five years ago, in
the middle of a large OSI push, 20 percent of traffic was X.400, but
the number faded to insignificance as users switched over to SMTP-based mail handlers.
There are two other TCP/IP networks in Korea. While SDN
supports itself by fees from members, the other two are government
supported. The Korea Research Environment Open Network
(KREOnet) is sponsored by the Ministry of Science and Technology
and was originally established to provide an access path to a Cray.
KREOnet was linked to SDN in two locations, providing a fairly
seamless Internet.
KREOnet also maintained a 56 kbps link to CERFnet in San Diego, but it appeared that most Intemet traffic moved across the
PACCOM link to Hawaii. Though the Hawaii link was saturated at
times, for some reason the CERFnet link had been relegated to a role
as a backup.
The third network was the Korea Research Network (KREN),
sponsored by the Ministry of Education. The network linked eleven
BITNET sites in Seoul to Japan via a 9,600 bps line. KREN also
linked nine universities together with 9,600 bps TCP/IP lines and
provided a path into the rest of the Korean Internet.
After a morning briefing on the networks, Professor Chon invited me to choose between hamburgers and "lousy Korean food"
in the KAIST dining room. Lousy was an overstatement, but it certainly would not be described as spectacular.
What made the lunch palatable was the conversation. Professor
Chon was the chair of the JTCl Korean Committee, the key committee for OSI work, and was also active in the CCITT standards arena.
He had heard that the Bruno project had been killed and grilled me
for the reasons.
Professor Chon had been planning on replicating Bruno for Korea and we discussed whether Korea really needed the ITU's permission. As an individual, one could argue that I had in fact
needed their permission, assuming you take the copyright assertion
at face value.
A country, on the other hand, was in a different situation. After
all, the ITU had come into being when countries like Korea had
signed the ITU treaty. A key aspect of international law is that any
right not explicitly delegated is retained by sovereign states, and
one had to wonder if the ITU could prohibit Korea from distributing
standard documents to its citizens.
One could even read the ITU treaty as encouraging this behavior. The treaty required that all members take all possible steps to
achieve the broadest dissemination of ITU recommendations.
Though online distribution of standards was not specifically mentioned, it was certainly not explicitly prohibited.
I explained my concept of a "standards haven," where a country
would ensure that standards were available online to its citizens.
Not for anonymous FTP, mind you, since that would muddy the
issue of sovereignty. Once the standards were online in one country,
though, it would be a simple matter of cutting tapes for other countries.
"Let's do it!" Professor Chon said. I cautioned that there might
be political fallout, with the ITU possibly objecting to the Foreign
Ministry, a common bureaucratic home for the official delegate to
the ITU. In Korea, though, the ITU representative happened to be
in the Ministry of Telecommunications. To my delight, the Ministry
had already gone on record as actively supporting online distribution of standards and had passed the requirement for such a fileserver down to Korea Telecom, the PTT.
Lunch started to taste better and I finished my
kim chi. Korea
was looking more and more like it might become the world's first
standards haven.
The conversation then shifted to other standards documents,
particularly OSI. There were two types of OSI documents that interested Professor Chon: working group documents and the final products, the International Standards.
Working group documents constituted enormous piles of study
papers, submissions, drafts, technical corrigenda, and other documents in such profusion that simply moving paper around to participants had become the key bottleneck in the standards process.
ISO had formed a working group on working group procedures,
but the working group was still bogged down in formulating the
procedures under which it would operate. The best that they had
come up with had been to use floppy disks to exchange ASCII-based files.
The secretariat for JTCl was, unfortunately, ANSI. Any attempt
to provide a file server for distribution of working documents
would require the cooperation of the ANSI secretariat and it was
highly unlikely that this notorious group of Luddites would want
any part of it. ANSI had been the single most vocal critic of the
Bruno experiment, sending their objections straight to Pekka Tarjanne.
Online distribution of working documents would certainly improve the ISO Process, but I wasn't so sure that I wanted to improve
it. It would be much more interesting to get the product from that
process available to those mere mortals who didn't have the time or
money to fly to exotic locations and stay in expensive hotels. People
in developing countries, students, and others without extensive resources should at least be able to read the international standards.
ISO standards, centrally produced and all in a common format,
were a prime candidate for scanning and optical character recognition. A single format and high-quality documents means that the
OCR software can quickly learn and can produce a fairly accurate
rendition of the standards. Modern OCR software can go so far as
to save markup information such as the font size or the location,
making semi-automatic conversion into a language such as SGML
very feasible.
The only hitch was that distribution rights went from ISO to the
national standards bodies and ISO appeared to have inserted a
minimum selling price for international standards. For draft international standards, though, there was no minimum selling price. In
other words, a national standards body could give away a DIS but
not an IS.
In most cases (but not all), the draft standard was virtually identical to the final product. The definition of a draft standard is that
there are only minor technical corrections to be made. It would certainly be nice to distribute the final product (the idea of giving students an inferior product certainly grated), but this might be a
loophole that would force some change.
The goal of setting up standards havens was not to get countries
into the standards distribution business. Instead, I was much more
interested in getting ISO and ITU to see the light and begin the
process of distribution themselves. After all, it would make so
much more sense to put a series of FTP (or FTAM) servers around
the world and release all documents over the Internet.
Professor Chon set me up to visit Korea Telecom the next day to
pitch my idea, and I went back to town to take care of unfinished
business. First on the list was a new power cord. The lounge at
Narita Airport had only 2-prong straight outlets and, desperate to
get some work done, I had finally snapped the ground prong off the
cord for my laptop.
Luckily, Korea used the same outlets as the U.S., including the
same grounded outlets. Unluckily, however, getting the concierge at
the hotel to send me in the right direction was a bit of a challenge.
Arranging a visit to the North Korean border would have been a
snap, but computer components drew a blank.
I made typing motions until the concierge shrugged her shoulders and wrote the name Se-Woon Sang Ga on a piece of paper. I
handed the paper to the cabby who shot me a puzzled look,
shrugged his shoulders, and took off. Ten minutes later, he pulled
in front of a run-down 5 story concrete building.
Turned out this was where people rebuild old video games.
Room after room was filled with old motherboards, technicians
squinting at oscilloscopes and old men playing Go. The halls were
crowded with rebuilt machines and school children came flooding
in to play free games, testing the machines.
I wandered past a store that sold nothing but joy sticks, another
that had ribbon cables, several video tube stores, and, inexplicably, a
sculptor making plaster busts of the dearly departed. Finally, I
found a shop with a few PCs running Tetris. I made jabbing motions at the wall, ran my hands in a snaking motion towards a PC,
and reached for the power switch, making flicking motions.
One of the kids, realizing that I probably wouldn't just disappear, got up from his game and I repeated my charade. He nodded
and reached up to the top shelf and handed me a power cord. I
forked over my 3,000 won (U.S. $4) and went happily back to the
hotel.
The next day, I took a cab way out to the outskirts of Seoul to the
research laboratories of Korea Telecom. The Umyon Dong building,
I found out sitting in Future Hall, was the most intelligent building
in Korea.
It certainly looked smart. Everything gleamed, the halls were
empty, and there were lots of video screens all over. Ushered up to
the second floor, I met Moon-Haeng Huh and Joo-Young Song, two
senior officials of Korea Telecom. They showed me around the
building, pointing out CATV servers, the FDDI backbone, and even
fingerprint recognition equipment for entry to high security areas.
The place didn't even have light switches or temperature controls in
the rooms: you used the telephone to call the automated control
server, keying in the temperature you wanted for a room.
After some ginseng tea, we discussed the idea of a standards
server. The labs, with an IBM 3090, Pyramid superminis, VAXen,
Sun servers, and lots of other equipment, certainly had the resources. They also had the mandate. We agreed that this looked
like a likely project and I left agreeing to send in a formal proposal.
I couldn't wait to get home to get this started, but I had a few more
stops to make first.