My Dinner with Erik

This post is about a man named Erik Naggum, who died in Oslo, Norway on June 20, 2009, apparently after a long bout with chronic illness. I say "apparently" because I didn't know Erik very well, and the official accounts of his death have been light on details (which is fine -- it's not really anyone else's business).

The reason I'm writing this is because despite the fact that I didn't know him very well, we had a very close, um, association over the years (it will become clear presently why I had to struggle a bit to find the right word to use). My name, for better or worse, has become inextricably linked with his in certain circles, and so I thought it would be appropriate for me to write something on the occasion of his passing. Death often provides an occasion for reflection.

I would like to start by setting the record straight about a couple of things. Erik and I were often very publicly at loggerheads and a lot of people believe that I had a personal dislike for the man. I didn't. I met him on only one occasion -- a story which I've never told, but which I will tell here shortly -- and we got along reasonably well, to the shock of many witnesses. To be sure we had deep and profound disagreements, but beyond those disagreements I knew (and to this day know) very little about him. I have no reason to believe that he was anything other than an honorable, hard-working man, but beyond that he was always a cipher to me. I mourn his passing because now that can never change.

My relationship with Erik is unique in that, with only a few exceptions, every interaction we ever had took place on a Usenet newsgroup called comp.lang.lisp. As a result, our relationship is chronicled in unusual detail, thanks to the eternal memory of the World Wide Web in general, and Google Groups in particular. Unfortunately, when I went to consult the record to prepare this piece I found to my dismay that Google Groups seems to be coming down with Alzheimer's. I would occasionally go there for a trip down memory lane, so I know it was working fine up until a few months ago. But last week, when I learned of Erik's death, I went to search for some of his old postings and found that Google's index of several Usenet groups has apparently been corrupted. The original raw data is all still there as far as I can tell, but the index that lets you search it has big gaping holes in it. As a result, I had to do the research for this piece manually, which is why it has taken me so long to finish this, and why I may well be leaving out some pertinent details. If you think I've missed something important, you're probably right.

To set the stage for people who may be reading this who aren't already familiar with the background, I have to tell you a little bit about comp.lang.lisp, or CLL as it is known amongst its adherents. CLL is a Usenet newsgroup, one of the earliest forms of on-line communities on what would eventually become the World Wide Web. Usenet itself goes back to 1979, and CLL to 1986.

My first post (as far as I can tell) was in February of 1990. Time Berners-Lee was still ten months away from writing the world's first web browser. The general public did not yet have access to the Internet, was not generally even aware of its existence, so everyone there was a computer geek. On-line communities (they weren't thought of as such back in those days) tended to organize around technical topics of mutual interest. CLL focused (and does to this day, at least ostensibly) on a programming language called Lisp, whose fans (and I was definitely one of them) tend to be unusually passionate about it. Lisp inspires love and devotion among programming language geeks the way that Apple's products do among gadget hounds. I won't get into why that is here because that would take me far afield. Just take my word for it.

Because of this common interest, and because Lisp tends to be somewhat of a black sheep among programming languages, CLL fostered a kind of camaraderie that went beyond its nominal scope, to the point that in March of 2003 when I felt the need to vent about the upcoming and IMO ill-advised invasion of Iraq, CLL is where I went. (Reasonable people can differ about whether or not this was appropriate. The point is, this is an illustration of how *I* felt about CLL. And I was apparently not alone because my post spawned a very long thread. But I digress.)

Erik Naggum joined CLL (again, as far as I can tell with search not working properly) in August 1994. (N.B. I'm actually pretty sure this is *not* Erik's first post because I have a vague recollection of looking for it several years ago when search was working properly and coming up with a different posting that was quite a bit earlier than this. But searching manually I couldn't find it.)

BTW, by drawing attention to the fact that Erik joined CLL several years after I did I in no way mean to imply that I had some kind of seniority privilege or anything like that. I am only pointing this out because I want to tell the whole story, and part of the story is that I've been observing Erik for a long, long time.

The thing that I remember first catching my attention was Erik's habit of not capitalizing the first words of his sentences. The second thing that I noticed about him was that he had a very no-nonsense take-no-prisoners style. And the third thing that I noticed was that he seemed very smart. But that was pretty much it. It was five years (again AFAICT) before Erik and I interacted in any way. My first participation in a thread involving Erik in any way was in February of 1999 and my first direct response to him was in March. If you take the time to follow those links you will see that there is nothing whatsoever remarkable about them, which is pretty remarkable in light of how things ultimately turned out.

Things didn't start to get tense between Erik and me until late August of 1999. This is not to say that things weren't tense. I was not the first to clash with Erik (but as fate would have it I would be the last). Erik was already getting into regular tussles like this one.

At this point I have to digress for a moment and say a word about my background. I am not doing this to be self-aggrandizing, but because the conflict between me and Erik ultimately hinged on differences of opinion, and I'm going to make some comments on those differences of opinion, and I want to provide a basis for you the reader to assess how much stock to put in my opinion. For example, one of the things that Erik wrote in the post I cite above was:

"what? a customer of a product has _invested_ in the company?! this is too fucking nuts to bother discussing."

The opinion I wish to express is that the position that Erik is dismissing is defensible. Not necessarily *correct*, but not so wrong that it can reasonably be dismissed as "too fucking nuts to bother discussing."

So let me tell you why I feel qualified to make such an assessment. One of the things I did in my career is to convince the powers-that-be at NASA to fly Lisp on a spacecraft, which was not an easy thing to do. The culture at NASA is extremely conservative when it comes to software technology, and Lisp was viewed with extreme suspicion. But I did it, and along the way I learned a lot about how managers think, and one of the things that managers think is that they *do* make an investment in a company when they buy that company's product, particularly a software product. Right or wrong, that's what they think.

I'm walking on eggshells a bit here because I don't want to start an argument with a man who is no long here here to hold up his end of it. I only want to point this out as an example of Erik's habit of taking very strong positions on matters that ultimately hinged on subjective issues, like what the word "investment" really means.

What bothered me about Erik's behavior was not that he was wrong. In fact, it was exactly the opposite. The problem was that he was usually right, and often people would try to argue with him when he was right, and often they would find themselves on the receiving end of some very harsh rhetoric. I saw two problems with this. First, it made it very hard to tell when Erik was right and when he was wrong, because taking issue with him could be so emotionally draining. And second, CLL was turning into not a very nice place to hang out.

It's important to keep in mind that I was not the only person who felt this way. A number of people had similar opinions. A few of them expressed their views openly if a bit inartfully.

And then there was this where Erik wrote (among other things):

"one of the things you will experience is that the community is very different from the Windows communities, whence it appears you come. for instance, the Common Lisp market is not marketing-driven, it is not a pyramid game that requires ever new people nor a bug-and-upgrade scam, and it is not leveraging its operational costs across a huge volume of sales. rather, it is a pretty mature market of long-term partnerships with a steady growth..."

Now, this really bothered me, not only because he was taking it upon himself to speak for the community (and I felt like I was part of the community and I never agreed to have him speak on my behalf) but also because I thought he was wrong. The Lisp market was *not* experiencing steady growth. I believed it was in fact in slow, steady decline (a belief which, sadly, has been at least partially vindicated by subsequent history). I also thought that Erik's very visible acerbic nature might be contributing to that decline even if only in a small way. And finally, I just found the man's arrogance to be incredibly annoying.

In August of 1999 I decided to try to do something about it. Having observed a lot of interactions between Erik and people who disagreed with him I knew that simply confronting him with my views would not work, so I chose a different tactic: I decided to try to engage him and model my behavior as closely as I could on his, expecting that this would lead to a confrontation. I would then reveal (privately of course) what I had done, and hopefully he would see that his style, having been on the receiving end of it, was counterproductive.

The exchange took place in this thread. My first engagement was in this message and it escalated from there, as I expected it would, at one point fulfilling Godwin's law.

I don't want to rehash the controversy over that exchange (there was a lot of it) but there are a couple of things that I think are important from a historical perspective. First, there were only four postings from me out of 183 total in the thread. I pushed only as hard as I thought I needed to to make the point. Nonetheless, by the end of our exchange, Erik was threatening to sue me, so I decided that things had gone far enough, and took the conversation to email.

I don't have those emails any more, and I wouldn't reveal their contents even if I did out of respect to Erik. (I did find this message from Erik in which he references that exchange just to show that I'm not making this up.) Suffice it to say, it turned out that I had misjudged the situation quite badly. My mental model of Erik had been very wrong, and my plan had, to put it mildly, not worked.

It was in this context that Erik and I had our first and only face-to-face meeting. It took place at the 1999 Lisp User Group Meeting (LUGM99) in San Francisco. This was the conference at which Erik delivered his now legendary "Long Painful History of Time" paper, as far as I can tell the only peer-reviewed work he ever published. To make a long story short, I managed to convince him to talk to me, and we had dinner together on the first evening of the conference. No one else was present.

When I was outlining this essay I had planned to write a paragraph at this point about my physical impression of Erik, which was nothing like what I had been expecting. But I've changed my mind for a couple of reasons. First, it's not relevant. Second, Erik was a very private person. He never published details of his personal life, nor as far as I know ever put a photograph of himself on the web. So I will respect his obvious desire for privacy and say simply that in person he was soft-spoken and very polite.

We reached an understanding at that dinner. I don't remember the details of what the understanding was -- this was ten years ago and it's not like we signed a peace treaty or anything like that. But by the time dinner was over, Erik understood -- or at least acted like he understood -- what I had done and why I had done it. The only evidence I have for this was that we got along for the rest of the conference, and this was witnessed by all the other attendees. I could have sworn someone wrote up an account of this on CLL because many people remarked on it at the time, but I can't find it. (Maybe it was in an email.)

(Ironically, at that same conference I also approached Kenny Tilton, with whom I had also been having some friction on CLL. He refused to speak to me, and to this day I have never had a real conversation with Kenny.)

I thought that would be the end of it. Things were quiet for a while. Erik and I didn't interact again until December, in a very unremarkable exchange, and then not again until February.

It all started to unravel in late February of 2000, when I posted this. Detente was not breached immediately. It was a slow, steady unwinding, and it's all there if you want to read it. I'll just note this where Erik wrote:

"I also vote that somebody write "the complete idiot's guide to special variables" instead of proposing silly language changes."

The result of that suggestion was this often-cited paper which owes its existence and its title to Erik.

It took a very long time for things to really deteriorate. Things were tense but still professional for a long time. As far as I can make out, the real turning point came in May, which ultimately led to this.

Shortly thereafter I quit JPL to go work for Google. I was absent from CLL for over a year, during which time Xah Lee stepped up to the plate. If ever there were two people who deserved each other, it was Erik and Xah. Watching those two go at each other was a truly tragic comedy.

I returned to JPL and CLL in August of 2001. My first post after my long hiatus was this one, which again is remarkable only for how unremarkable it is.

Things apparently hadn't changed much in my absence. I watched with some dismay asJohn Foderaro was driven from CLL. I thought this in particular was a little over the top. The understanding we had reached in 1999 seemed to be long forgotten.

I myself didn't enter the fray again until late August. This thread is notable because for about half a dozen messages I responded to Erik by cutting-and-pasting exact quotes from earlier messages of his. I was hoping this "replay attack" (a term of art from computer security) would remind him of our conversation at LUGM99. No such luck.

In retrospect I find it fascinating how the level of tension on CLL seemed to suddenly build in the days before the World Trade Center was attacked. Maybe there was something in the air. (For the record, I didn't participate in that thread.)

Erik and I never managed to rebuild the fragile truce we had forged in San Francisco. Our last exchange was this one after which, as far as I could tell, he not only dropped out of CLL but stopped posting in English altogether. (He posted a couple of things about me in Norwegian but I don't speak Norwegian so I have no idea what he said. Frankly, I think I'm just as happy not knowing.)

So that's my Erik Naggum story. I've told it as accurately as I can. There might be mistakes. I did the best I could under the circumstances. And I think Erik did too.

So farewell, Erik Naggum. Notwithstanding our differences, I actually learned a lot from you, and not least among those lessons is that life is short. So I will always remember our meeting in October of 1999 with great fondness. Everything else is lost in the noise.