Does It Seem Too Much Like a Vacation?

We provide a lot of freedom in the agenda; Indeed, you are encouraged to create your own agenda to optimize your experience. Once you acknowledge that everyone starts at a different place and moves at their own pace, you can't presume to design their learning experience (help, yes. Control, no).

For one thing, the "formal" conference discussions only happen in the morning. There are less-formal things going on in the afternoon, in particular free time for skiing and other snow sports. In summer conferences, I lead hikes (usually in the mornings, since storms tend to happen in the afternoons). Although these things seem non-conferency and more vacationy, on the ski lift or the hiking trails, or at informal dinners is where some of the most interesting conversations take place because that's when people relax and feel like they can speak less formally.

Afternoons are also fair game for technical workshops, and more than a few have spontaneously organized around a subject. And in the evenings we have lightning talks (we'll probably expand these to start on Monday evening next year, to allow more slots; we were overflowing this year). Although lightning talks can cover any topic -- Bill Pugh, the FindBugs guy, did fire-eating and Carl Quinn's son gave one on shaving -- they are often technical and they always seem to stimulate creative thinking, whatever the topic.

It might look like a vacation, but it seems to produce more knowledge value than sitting in lectures and exhausting yourself. Still, we've been asked to move the details about afternoon activities further down the information page, so that some managers won't feel they are over-emphasized.

Many speak of coming away from the conference inspired and re-invigorated about their profession -- how that happens seems to be a personal experience. The fact that we can make a framework for this to happen is reason enough to hold such events.

Not For Everyone

One thing I've found fascinating, after seeing open spaces work so well, is discovering when it doesn't work. I've been asked to help several eyes-forward conferences to organize an open-spaces track in parallel, and that doesn't seem to work so well. I thought it would, because my first exposure to open spaces was at the Python conference where they worked well. However, I can't recall whether they were being held during regular eyes-forward sessions. My experience is that, if you come expecting an eyes-forward conference, it will be difficult to get you to try open spaces.

I think open spaces will eventually become much more common, but in the formative years it requires some kind of self-selection process. I always try to be very clear about what it is and what the process is when I hold conferences, but the Roundup is gaining enthusiasm to the point where some people might just hear how great it is and decide to go without learning about it first (or, in the case of the Roundup, listening to the podcast), and be surprised and uncomfortable. And even though I personally feel far more inspired at open spaces (although I still speak at eyes-forward conferences, I do so by considering at a performance), there are apparently speakers who prefer lecturing to a group.

To each his own, but perhaps I should emphasize in the literature "We want you to have a good experience; please study and understand the nature of open-spaces conferences before registering."

Progressive Dinner Fail

Last year, for the first time, we had a progressive dinner among the group-rented houses, which apparently went very well (I had just broken my leg last year so I was out of it). We decided to do it again this year, but the number of people was so large that we felt we needed to control it, for example by breaking people into groups and timing their movements. Something overdesigned like that, as programmers are wont to do.

Joe Nuxoll made the argument that we shouldn't control it too much, but our solution (break people into two groups and have them go in opposite directions) was still overcontrolled -- an object lesson to rethink your rethinking; just because it's your second idea and better than the first doesn't mean it's the right solution.

In the end we decided we should have done it the obvious way, just like last year, where everyone moved through the houses in a single large group. When that house runs out of food, the people from that house join the group and move on. If a house gets too crowded, some people will probably move on.

It's another example much like premature optimization in programming. It's exceptionally hard to stop yourself from imagining that you will know where the bottlenecks will be, and it takes a lot of experience to school yourself out of that reaction. The key is to do the experiment first, then follow the clues.

Complex Event Processing (CEP)

The only technical session I attended talked about Complex Event Processing. Although I don't have so much interest in programming anymore, I am still fascinated with complex systems (I was briefly in a Ph.d. program at UCLA in "Engineering Systems" which I thought would explore this topic, but which was cancelled -- mercifully, because it was time for me to leave academia -- midway through my first term).

It should probably be called "simplified event processing;" the misnomer suggests the resulting scheme is "complex," when it is instead designed to tame complexity, and thus allow the creation of complex systems. The book is The Power of Events.

Because such complex systems cannot afford the luxury of deadlocks, race conditions, etc., they must solve the concurrency problem at a much higher level (not one where programmers must consider every possible alternative at each step and level). As a result, this work may provide an answer to the question "what are we going to do with all those cores?" Intriguing indeed.

Future of Programming

My Programming in the Mid-Future blog did more than just seed ideas for a discussion session. One person told me it pushed him over the edge into registering and attending the conference.

The session was fascinating not just because of the ideas that arose, but from the breadth of those ideas. It seemed to depend on how you thought about computers as to how far you thought they would progress in 25 years. I was most interested in Joe Nuxoll's view; Joe was one of the lead developers for JBuilder so he knows the low-level details, but he has moved away from programming and into user experience in recent years, where he focuses on "what does the user want to happen" rather than "how do we make the computer do something." Joe's belief, which I tend to share, is that the machines themselves will withdraw into the surroundings -- I think these surroundings will include our own bodies -- and be listening and watching to find out what we want. Joe's example was "bring me a burrito" and some people thought this meant that the burrito would be nano-fabricated, but for me it was just that, somehow, the matrix would get a burrito delivered to me (I suspect the burrito itself will be built in the conventional fashion).

Lightning Talks

One takeaway that people have reported from previous conferences has been lightning talks. This is a nice, discrete tool that numerous people have taken back to their companies and put to good use. While they are certainly more structured than open spaces, they give a bit of the flavor and may encourage people to do further experiments. For example, this year we had a scala coding dojo, which is a relatively new concept.

The Goal of an Experiment is to Create Knowledge

Jason Nerothin gave a lightning talk where he argued that at the core of value creation is knowledge creation. Dick Wall and I pondered whether the Roundup represented a company that lasts for only one week and produces quite a bit of value in that time. We cannot know exactly what knowledge is created because it's different for every person and the discovery of the knowledge often happens later. Thus it's very hard to know what you've created, even though it seems to be very valuable. Such a company would need a different view of monetization.

Occasionally an experiment might also create something else, like a salable product. But if that's your requirement for success, you'll usually fail. If you aim to create knowledge, you'll usually succeed.

Business Insights

I organized several business-oriented sessions, as did Joe Nuxoll. These were very illuminating. The most fascinating concepts were things that we persist in doing that we know are counterproductive (e.g. multitasking, extrinsic motivators, furniture-police behaviors), and things that make us happy at work (fairness, feeling of daily progress, autonomy, contributing to something larger than yourself).

The big question was too hard to answer: Why don't we stop doing the things that are destructive, and start doing the things that make us happy?

In one of the discussions I started, I mentioned the furniture police (from DeMarco & Lister's Peopleware) and was saddened to hear from another person in the discussion that this very thing was happening right now at Hewlett Packard, a once-great company where people would leave to find something better and then come back because it was such a special place to work. What we see now is the aftermath of Carly Fiorina's destructive tenure (more MBA mayhem). Fiorina received 20 million when she was removed as CEO, because she, like all such CEOs, have magical powers that will make the company super, so ridiculous compensation is necessary.

It doesn't matter that, again and again, saving money on floor space by cramming cubicles together causes people to leave, which costs far more than the pittance saved. Cubicle costs can be measured, and thus "managed" in the view of business-school theory, so when things get tight that's what you do.

The story that business schools tell is too compelling; it doesn't matter whether the effects are shown to be negative as long as we keep repeating the story and believing in magic. The fascinating thing about spells is how, once incorporated into the target audience, they constantly re-invoke themselves and cause the behavior to repeat. Not unlike a computer virus.

I also held a very practical session discussing the logistics of holding a series of conferences for creating exceptional businesses, which I plan to start this summer. Participants were very helpful in giving guidance. (Perhaps the initial theme should be "Breaking the Spell").

Twitter

I finally used Twitter and it turned out to be somewhat useful, but only in the limited sense of reading and posting messages about Roundup activities. Actually, this was not really that much of an improvement over the Google group for the Roundup, except that a number of people were following Twitter on their mobile devices so maybe the word would get out faster. It's not clear to me whether these same people might have also been following the Google group the same way, in which case it would be redundant.

I noticed that some people were able to meet up at the Denver airport using Twitter. I could see setting up an account to send special announcements as text messages to mobile phones; however, the normal "status" messages would have to be filtered out of this or it would get too noisy and annoying.

Dick Wall told me he found Twitter valuable to do a kind of crowdsourced research when he's trying to figure something out. And Twitter seems to do what it's supposed to do in a very simple and obvious way. But I found no compelling reason to keep using it, so it will go back on the shelf until the next conference.