Erik was a flamer’s flamer, back in the golden days of Usenet. He was reflexively averse to the mainstream; a proponent of SGML before descriptive markup was fashionable, he peeled away from the community when XML hit the big time, vanishing in a puff of pungent sulfurous smoke. I think he’s left an important lesson behind him.
First, I have a story. In 1993 the European SGML conference was in Rotterdam; I went for a random walk around town and at a canalside café there was Steve Pepper, an Anglo-Norwegian pillar of the markup tribe. He waved me over and I settled down gratefully for a cold beer with Steve, and a handsome guy with what I took for a US accent; that turned out to be Erik. His English was always awesome.
I was a little tense because Erik and I had crossed swords on a few occasions; in those years I still indulged in flaming from time to time for its own sake — not at Erik’s level of course, but still.
No worry; we sat and drank and talked through a couple of hours of a warm Dutch evening; it’s a golden time in my memory. And I suspect that after that Erik moderated the flames a little when I was the target.
Erik was a smart guy and an articulate guy and he cared about a lot of the same things that I do. I’m glad to have known him.
Let me end with a question: When you’re gone, would you like to be remembered as smart and articulate, or as a famous flamer?
Contributions
Comment feed for ongoing:
From: Derek K. Miller (Jun 20 2009, at 20:06)
A famously smart and articulate flamer?
I was never any good at flame wars, though.
[link]
From: Eric Meyer (Jun 21 2009, at 00:47)
Smart and articulate.
[link]
From: Steven Champeon (Jun 21 2009, at 14:53)
When I first got started with SGML back in 1993, we got access to comp.text.sgml and read it via a Usenet-to-email gateway. Erik Naggum was as Tim said, smart, articulate, independent, and a force for good. It was from people like Erik, and the other folks that made comp.text.sgml such a great forum, that I learned how good online communities run, and what can threaten them. It was his example, along with those of folks like Len Bullard, that left me with a lifetime love of the Internet and its various communications mechanisms, and probably why I have decided to spend the past few years trying to save email from spammers. Eric Meyer sometimes thanks me for being such a good example of a jack booted list mom, so he's not alone when he does it on css-d; I learned everything I know about it from Erik Naggum.
[link]
From: Eliot Kimber (Jun 21 2009, at 20:25)
I am sad to hear of Erik's death--much too soon.
I met Erik in person exactly once. I don't remember the exact date, but it had to be the early '90s, not long before XML, when we (Erik and I, Charles Goldfarb, Steve and Peter Newcomb, and many others) were working really hard on getting HyTime 2 and DSSSL together.
Charles Goldfarb always loved Erik for being an SGML True Believer and a master programmer and the steward of comp.text.sgml. He hired Erik for some project or other, as much to keep him eating as anything (if memory serves), moved him to Saratoga, CA (where Charles has his home) and installed him in a little apartment in some anonymous complex.
I was visiting Charles or in town working (this must have been during the time I worked for Passage Systems, based in Cupertino) and went to visit Erik--we had corresponded copiously on comp.text.sgml and in private, but had never met in person.
I didn't really know what to expect but was surprised to find a much younger person, and much nicer person, than I had expected. I remember bringing a box of dougnuts, which seemed to please Erik. He eagerly showed me something he was building on his computer, I don't remember what. He was happy and excited and intense.
It must have been soon after that meeting that XML happened and Erik left the community as I have no subsequent memory of any significant communication with Erik.
Like many who have commented on Erik's legacy and his effect on then, Erik definitely forced me to think before I posted and helped to cement a desire to be a positive, constructive voice in online discussion. I also valued Erik's commitment to the cause of structured markup and The SGML Way and marveled at his mind, good and bad.
--Eliot
[link]
From: Paul Davies (Jun 22 2009, at 14:15)
I never met Erik myself, despite living in the suburbs of Oslo while I was working on an SGML database publishing system.
However, I do remember reading his contributions to the comp.text.sgml newsgroup at the time. Appended to his contributions was a pithy aphorism, my favourite of which was: "If you evaluate C++, you just get C only bigger."
[link]
From: len (Jun 22 2009, at 17:02)
Erik reveled in his brilliance. It made him giddy to use it if creating or destroying. He didn't suffer pretense of knowledge. I once made a dry comment about LISP (remember when SGML was a LISP competitor?) and Erik sent me hundreds of pages of LISP documentation: in a single email.
If one would learn, Erik would tolerate.
For all the brilliance of the flames, Erik respected those who respected truth of science, math and personal integrity. If he pushed me, he enlightened me.
I met Erik in Florida with Steve Newcomb and others but Erik was as brilliant in person as on line, but better.
Erik cared. Good journeys, Erik. I hope your questions are being answered most truthfully old friend. Thanks!
[link]
From: Paul Prescod (Jun 23 2009, at 23:11)
I think that Erik was a sort of fundamentalist. He was a fundamentalist truth-seeker.
[link]
From: Paul W. Homer (Jun 25 2009, at 11:39)
I'd be happy if people just noticed that I wasn't around any more. It's far to easy to be invisible.
[link]
From: Henry S. Thompson (Jun 29 2009, at 08:53)
Erik taught us all a great deal, and I learned a tremendous amount from him during the late SGML/HyTime/DSSSL period, before he shifted his ground. His iconoclasm has been widely celebrated here already -- I will mention only one positive aspect thereof: he disliked CONCUR, and said or wrote once somewhere that the proper treatment of overlap required standoff annotation, that is, separate markup for each distinct tree-structured view of the data. I got a lot of mileage out of that comment, and hereby acknowledge that the idea was originally Erik's.
[link]