travel report - June 1996

NOTE: These are my personal notes, and do NOT form any part of the official output from that meeting, nor does it form an UNINETT position.

June in Montreal

It may be symbolic for the IETF in this city that it straddled the day when Quebec would have seceded from Canada had its plebicite gone the other way. I think I will remember this IETF long for the surprising lack of animosity, the miracles of concilliation that happened, the disasters that didn't happen and the intense interest in cooperation, often coming from unexpected partners.

We have met the enemy, and he is us

Microsoft was, for the first time in history, attending the IETF in force. And I DO mean in force - of the more than 1100 preregistered attendees (another record), 37 were from Microsoft. They seemed to be everywhere in the Apps area, always bearing the message "we want to cooperate; we want to join the standards process; we are nice guys".

What's more, I believed them.

Dance with bear, expect bruises

For the first, and probably the last time, the IETF was colocated with the INET conference, the annual conference of the Internet Society.

The approx. 2000 INET attendees proved an interesting contrast to the highly casual, technically oriented nearly to obsession crowd at the IETF conference; one wag called it the "suits and sandals" conference.

Some crossover occured; many of the speakers at INET were from the IETF crowd; some of the attendees at INET stuck their heads into IETF conferences. But the mingling was smaller than expected, and did not seem worth the trouble of having a joint meeting.
(The fact that the conference hotel was a 20-min walk from the conference centre was also heartily disliked by most participants - but hotels with room for conferences with 3000-plus attendees are VERY few and far between....)

We'll get it right this time!

The IETF has always been proud of the quality of its standards; they are tried before standardization, the use of open discussion and independent implementation is supposed to keep "everyone honest", including the standards writers.

Nonetheless, some of the most important standards: Mail, News, FTP, haven't been revised for 10-15 years, and their language is archaic, unclear and out of touch with the real world as presently seen.

In all these areas, efforts are underway to rewrite the standard, and "get it right this time" - NOT adding new functionality beyond what is needed to make sure extensions can be added later, but making sure the documents are simple to find, easy to understand, and specify the protocol we are really using.

The toothache is worse than the dentist

Some things are just obvious that we need a standard for, yet we have none, because of various reasons involving conflicting requirements, people's pride, adjacency to known ratholes or other reasons.

However, the longer we wait, the more painful the lack of a standard is.

In both the Uniform Resource Names area (providing permanent identifiers for Web resources, so that they can be copied and moved) and the even more contentious Key Exchange area (making sure you know who you are trusting, and maybe why), the pain of not having an agreed way of achieving this seems to have led to a large-scale slaughtering of sacred cows, and a willingness to compromise that has surprised hardened observers of the process.

I have hopes that a similar process is underway in the E-mail security area, but at the moment, nothing much is happening there; PGP/MIME has been put on the standards track, MOSS remains on it, S/MIME is changing and not standard, and MSP (US DoD) is a very dark horse.
We'll see if the pain rises a bit in the next 6 months....

You're not in Kansas any more, Dorothy

The Net is an international phenomenon.

For proof, there was no need to go furhter than a glance at the INET delegates, who formed a far wider range of ethnic backgrounds than the still largely white-male-American crowd that forms the IETF.

Still, a lot of IETF standards have trouble with handling of internationalized material: language, character sets and other features that are important are often forgotten at the design stage.

The IAB commissioned a workshop to look into this just before the previous IETF, and the recommendations of that workshop, while not finally published, were talked about a lot here.
It would be wrong to say that the specific recommendations (basically, UTF-8 and language tagging) were embraced heartily by all; however, large swaths of the community seemed to heave a sigh of relief that there WAS a recommendation, so that they could "do something sensible and get on with the fun parts".
Of course, the devil is in the details....

What else happened?

This is the left foot of the elephant, as usual. Lots of other stuff happened, most of which I know nothing whatsoever about; you'll have to find other people to tell that story.

Some words of note:

All in all, I feel good about this conference. Things moved; the sky didn't fall; the pleasure was present and the pain bearable.
And we have almost 6 months to go before the next one....
Last modified: Mon Jul 1 10:15:49 1996