November 17, 2011

Too smart, doesn't get quite so many things done?

We care about our craft. We're totally smart and get things done. No question.

But "smart and gets things done" has to have some kind of spectrum associated with it, right? There's at least a "smart" dimension and a "gets things done" dimension.

An easy question to ask is, "Am I overthinking?" (This is especially easy to ask if you're overthinking.)

We often quibble about how to get things done better [*] in terms of practicalities, but it often feels like people who ignore the long tail of practicalities achieve greatness with the least effort.

If you had to pick one, would it be better to over-think or to over-do?

(My advice: don't think about it too much.)

Footnotes

[*]

In some asymptotic sense of better.

Collaboration and concentration

Latest in the, "People problems are hard, I'm glad I just program computers," set of thought processes. If you have thought-provoking anecdotal data points, your comments are welcome!

There's a confounding and contentious dynamic in programming between collaboration and concentration. This goes beyond just individual and team efficiency — it's also about finding a balance that lets you enjoy and grow your craft.

There are certainly times that you want to hunker down in an LCD-monitor-adorned sensory deprivation chamber and do some badass debugging, find all the corner cases in that nasty algorithm, or suss out all the invariants in a complicated system.

But, on the other hand, there are certainly times that you need to write code during a flurry of peer interaction. If you're ramping up on a new system, the thirty seconds it takes you to ask another human being a question and receive a solid, well-informed answer usually trumps the hours you'd spend reaching the "same" conclusion. With less certainty. Oh, and you actually missed the trap everybody else already knows about, so you actually figured it out wrong.

The extent of the dichotomy is clearly not limited to these examples. It's easy to think of times you've had to pepper somebody with questions as well as times you wanted to hole up in a cave to get some code done. For my systems programming experience I'd hazard a guess that it's around an 80/20 split between concentration and collaboration.

So, it's interesting to ponder: if you were building a team, how would you structure the work space or work activities to enable collaboration when it was necessary, but enable concentration in the common case?

I suppose it's doubly difficult because you also want your team to feel empowered — capable of seeking out collaborative help when they need it in order to get things done and make progress, but also empowered to cordon themselves off and have at it.

I can pretty easily find research that involves hospital workers in the early 1990s, but I have to imagine that a team of brilliant systems programmers is a different beast altogether.

Ship, or don't die!

Perhaps a corollary is, "Release early, or release toxin."

I admit I'm a bit of a Seth Godin fanboy — he's driven, omits needless words, and gets things done. His blog rarely has an unread count in my feed reader. At the same time, when you look up to someone, you can't help but expose some vulnerability.

One of his latest kick-ass entries, What did you ship in 2010?, put me in a total funk. I shipped a modest set of things this year, by which I mean that I found my list unimpressive. Maybe it was too short, maybe it wasn't innovative/creative enough, or maybe I was just being a negative Nancy.

In any case, Nancy found... er, I found the list-writing experience incredibly disappointing. [*] However, after some thought, I realized what I would probably say to Seth if we had a chat about it over coffee at Red Rock: I don't think I should really care.

Why? Because I'm probably not going to die tomorrow.

Kindergartener existentialism

A fun-size bit of existentialism is that human essence isn't fully realized until you die. When you die, your whole lifeline has played out and your effect on the world is fundamentally complete. To use a catchy existentialist marketing slogan, human existence precedes essence.

In a related vein, kindergarteners don't try to get their macaroni pictures displayed in art museums. When you're new to a scene and acquiring pre-requisite experience, there's no need to subject the rest of the world to your crap: in the common case, there's plenty more time to cultivate your essence and make your mark on the world. In fact, experimenting, throwing your crap away and moving on may be a much better use of your time than trying to ship something naïve or artless. [†]

My parents wouldn't want to put my broken patches up on their refrigerators. Even if they did, they don't have those kinds of refrigerators that magnets can stick to.

Shipping in potentia

Like most people who suffer from over-achievement syndrome, I have fever dreams of instantaneously becoming an expert in every piece of tech I touch, innovation dripping from my fingertips as I puke rainbows and such. Discovering that talent and perseverance have limits is always a cruel come-down.

Perhaps because of these delusions, I initially found it hard to grasp my most important accomplishment of this past year: getting to know various aspects of a state-of-the-art, production, multi-platform language design/implementation and the surrounding processes and tech. That's not shipping! It is, however, necessary experience to ship higher-impact (and perhaps daydream-worthy) tech down the road.

Realistically, there are a number of other reasons to feel accomplished. When I left my last gig slightly under a year ago, not a single product I had written code for had shipped. (Although I'm totally rooting for one that was recently announced!) Now, every patch I write is put to the test in a development channel with millions of active daily users. I'm constantly and (relatively) shamelessly absorbing information from a team of brilliant and down-to-earth developers, my mentor Luke Wagner in particular. My scrappy throw-away side-projects keep me thinking creatively and questioning the status quo.

In all, this year was incredibly enriching.

The JS engine is more comfortable ground with each passing day. I've got the drive to give back important and innovative things. My existence precedes my essence.

I'm optimistic about the list for 2011.

Footnotes

[*]

Whatever the size of my contribution, I'm honored to say that my list for 2010 included, "Help to ship a working and efficient JägerMonkey implementation." Effin' a.

[†]

Of course, one has to be somewhat cautiously introspective — Seth also warns that continued concerns over naïvete/perfection are a natural result of a fearful mentality that he refers to as the "lizard brain".

Remove the self-selection bias from Q&A sessions

I've been in quite a few painful Q&A sessions. I think we can do better.

When somebody volunteers to ask a question, their question is not necessarily of interest to anybody in the audience other than themselves. Despite that possibility, publicly answering the question necessarily takes up the time of every person in the audience. Remember the kid in class who asked 90% of the questions — questions that nobody else ever cared about?

There's a simple way to fix this.

Step 1: At the end of your presentation, ask for a show of hands for those people interested in having a Q&A session. If very few people raise their hands, you should be worried about the quality your presentation. However, in the less gloomy scenario where a good number of people raise their hands, you have a representative sample for the audience population that might be interested in the answer to any given question. (The other people should probably leave, but suggesting that would be rude.)

Step 2: After each person volunteers to ask a question, give the audience a quick poll by saying, "Raise your hand if you're interested in the answer to this question." If very few hands go up, reassure the person that their question is a good one [*] and tell them you'd love to stick around and chat afterwards.

That's it. Each several-minute irrelevant Q&A iteration is now reduced to a few seconds and you have additional feedback as to the topics the audience is interested in — the added audience involvement can't hurt either.

Caveats

Real-time feedback software that the audience interacts with during your presentation can definitely do better. This is just a low-tech proposal to raise the bar a bit.

As my friend pointed out in discussing this, there's no simple way to determine the critical mass for answering a question — it's possible that any given question will only be interesting to a fraction of the audience. It's also possible that you, the presenter, know that the answer to a question is particularly interesting and should be answered without soliciting audience feedback. I believe in your intuition!

Footnotes

[*]

Even if it's not — otherwise, question-asking potentially becomes public humiliation, which is very undesirable.

House rental, showers, and auctions

I'm currently in the prologue of the house-rental process. Our sights are set on a four-bedroom house with an interesting property: the room resources are of differing value — not in terms of square-footage, but other, less easily scalable factors.

So how do we dole out the resources of unequal value in a manner that everybody considers fair?

(Talk about first-world problems...)

Proposed solutions

Three systems for deciding the room allocations have been proposed, and one has been rejected:

I'm personally a fan of the simplicity and total lack of competition in the random system, but if the bidding system is more likely to lead to the optimal outcome, it must be considered! (Of course, care must be taken to address possible social ramifications.)

During casual discussion we noted that the bidding system would be strategically interesting. If a person were not really interested in getting a high-value room but didn't care all that much if they did, it is clearly in their favor to drive bids on the early rooms as high as possible — this would entail a decrease in the payment rate for a later room.

I perused the Wikipedia entry on auction theory and realize that we were assuming an open/ascending bid auction, but a first-price/sealed-bid auction might be less socially stressful. [†] My main fear is that open bidding would get carried away (it's exciting, after all) and participants would come to regret their bids.

In any case, it's an interesting problem I thought I would share. Suggestions are welcome; otherwise, I'll write a follow-up entry as to how it turns out.

Footnotes

[*]

This means it has no shower. Stupid terminology, but I suppose it has a nicer ring than "toilet room".

[†]

As an aside, the game theory described on that page also seems to discuss provable equilibrium of a single auction, whereas our auctions have dependent equilibrium values.