malvasia bianca

Over the last couple of years, I’ve started to appreciate retrospectives in a way that I hadn’t before: I’d felt for a while that they made sense intellectually, but I was never actually good at them. And, while that hasn’t particularly changed (fortunately, other people on teams I’ve been on are better at them!), I have now seen several instances where I’ve felt like they’ve had concrete benefits and where I could see the seeds of a rhythm developing.

So that’s great! And I’ve also learned something about retrospectives in the process. I’d always thought of the main benefit of retrospectives being that it helps you improve your process. But there’s another benefit beyond that: the fact that you’re doing retrospectives means that you have a process.

I mean, you can do retrospectives without writing down every aspect of your process. But the fact that you’re talking about some aspect of what your team is doing means that, in that area, you have a process. Which is good, irrespective of the benefits of tuning the process! Or at least it can be good: a bad process is probably worse than no process, but neither of those options is a place where I personally would want to be.

Thinking about it a bit more, there’s another benefit lurking in the assumptions that are prerequisites for a retrospective: the fact that you have a team at all, and that the team members’ thoughts are all worth listening to.

The interesting thing about Plants vs. Zombies 2 isn’t the gameplay: that’s largely the same as in its predecessor, and remains solid. Instead, it’s the business model: the game is free to play, and makes its money off of in-app purchases. Which is something that the corner of the games community that I’m part of is strongly against; I don’t agree with that antipathy at all as a general statement, but of course the implementation matters.

So: how does Plants vs. Zombies 2 do in that regard? I felt a little uneasy at the start: there are a lot of gates, and there isn’t a “pay a fixed amount of money to unlock everything” option. But you could unlock almost all of those gates by playing, so I figured I’d play and see how it went.

And then I finished the first era, and hit a gate to proceed to the next era; also, there were some side paths to unlock new plants and new levels, and I hadn’t unlocked most of those. The cost to proceed to the next level was actually pretty reasonable (five bucks, I’d probably already gotten five bucks of enjoyment out of the game at that point), but the game also gave you an alternate way to unlock that gate: the levels you’d already finished got repopulated with three new challenges for each of them (e.g. accomplish a task with limited plants or limited sun or without planting in certain areas), and if you finished enough of those challenges, you’d unlock the next era.

So I figured I’d give that a shot; and it turned out to be super-fun! The challenges were well designed, I really enjoyed playing through them, and I played through more than were necessary to unlock the gate in question. And, as a bonus, while I was playing through those challenges, I collected enough keys to be able to unlock all of the side paths without paying any money. (Though here I have to give a caveat: item drop percentages are the sort of things that companies change over time, so other people’s experiences might be different.)

The result was an experience that I really enjoyed, and that I didn’t pay any money for; the challenges actually made the game better than the original. (Here’s a post on Kotaku that reports a similar experience.) And it’s not like you have to go through every challenge, either: if there are a few challenges that are too tough or too annoying for you, you can skip them.

After that, I was pretty optimistic. And I had a similar experience with the challenges on the second era: I enjoyed going through them to unlock the gate at the end. (There isn’t a gate on the third era.) What was different, though, was the side paths: the number of keys that the game dropped went way down. I only unlocked a couple of side paths on my own on the second era, and I haven’t unlocked any yet on the third era, though I’m sure if I went through more of the challenges then I’d unlock at least one of them.

So, the result is that, without paying money, I’ve gotten all of the narrative out of the game, I’ve gone through something like 50 challenge levels, and I’ve got many more challenge levels awaiting me if I feel like it. And I’ve unlocked about half of the extra plants and powerups that you can unlock in game; I haven’t unlocked any of the plants and powerups that you have to pay money for. (The pay-only plants are all from the first game; the powerups are not necessary to succeed.) I didn’t count the time I spent playing it, but my guess is something like 10 hours, and they were good hours.

If I think of it as a free game: that’s a very good deal. If I wanted to pay to unlock the remaining side paths, it would cost me $12 (and my guess is that, if I’d gone through more challenges on the third era, I’d save $2 off of that); I’ve already gotten more than $10 out of gameplay out of the game, so that also feels fair to me. I haven’t actually paid that money yet, mostly because I felt like I’d played the game enough for now, but I can actually imagine returning to the game at some point in the future and deciding that I’ll spend money on that. (And the game is set up to allow them to release extra eras; if they do that and charge for them, I’ll be happy to pay.)

If I wanted to unlock the purchase-only plants, it would cost me $20; if I wanted to unlock the purchase-only upgrades, it would cost me $14; and if I wanted both, I could save a buck or two off of the combined price by buying bundles. That is a higher price than I personally would be willing to pay; the price of all of that, however, still puts it at less than the price of a new console video game, and the value of the total experience is absolutely there. That feels to me like an explicit price segmentation strategy, and one that I support: people who really love the series and will put in dozens and dozens of hours will spend the extra money, and it will be worth it for them; other people won’t, and those people who won’t will still get a fine experience.

In general, that adds up to an experience that I basically support: there’s a bit more asking for money than I’d like, but I can ignore that. I wish the item drop percentage hadn’t dropped off quite as precipitously, but whatever.

Having said that, there is a big caveat: if you look at the top in-app purchase list, the most popular items aren’t any of the ones I’ve mentioned above: they are instead coin packs. To explain that, let me go back a bit: I said above that the gameplay is largely the same as its predecessor; that’s true, but there are a few additions, and one of them is a temporary special power that lets you destroy zombies en masse.

I almost never used that special power: it felt wrong to me, actively going against the core Plants vs. Zombies experience. But it’s there; and, if you want to use it, you’ll need coins. You’ll earn some coins through gameplay, so if you only need to use the special power sporadically (e.g. to get through a challenge that you’ve played a few times and can’t quite survive the final wave on), you’ll be swimming in coins. But if you want to use that special power regularly, you’ll need to spend money; judging from the in-app purchase list, a decent amount of people are doing that.

And that’s not good: I do not support pay-to-win. Now, Plants vs. Zombies 2 is handling this much better than many games do: some games give you difficulty spikes where you’re forced to pay to proceed, giving something that looks like a skill gate that is actually a pay gate. (I’ve heard that Candy Crush Saga is a prominent example of this, though I can’t say from experience.) And that’s much worse than a pay gate, because it attacks the notion of expertise at the game. That doesn’t happen in Plants vs. Zombies 2. (Or at least didn’t when I played: these are all numbers in a spreadsheet that can be updated silently behind the scenes at any time!)

But still: the game is designed in a way to ask players who are less good at the game to pay more money. And I don’t like that.

Having said that: people can make their choices, and it is not my position to tell other people whether they should prefer to bang their head against a tough level, to pay money, or to give up. (And designing a game where people never bang their heads against tough levels is also problematic in its own way.) Note also that this is problem isn’t unique to in-app purchases: many many games ask you to grind to win (with more grinding required the worse you are), and from my point of view, that is no better morally than asking to pay to win. (Time is in many ways much more tightly constrained, much more precious than money!) Also, EA is avoiding some pitfalls that I find more morally objectionable: there’s no multiplayer, limiting the damage of pay-to-win (there’s a reason why I don’t play Magic: The Gathering!), and the game is relatively straightforward about the tradeoffs that it’s asking users to make. (The aspect of Sorority Life that I felt actively ashamed of was its “crates”: those are designed to make money by taking advantage of human brains’s bad instincts when it comes to probability.)

All in all: a good game, and an interesting case study. Ends up in a place that I’m pretty happy with as a player; but I imagine that future PopCap games will be worse in this regard.

In the past, my habit has been to only write my main post about a game when I stop playing it. That makes sense for the vast majority of games that I play; but for games that I play for month after month, that I would play in perpetuity if a sequel wasn’t released, that policy perhaps makes a bit less sense? Which brings us to Rocksmith.

But: what to say about the game? The problem with living with the game for so long is that nothing feels fresh any more: every thought that I’ve had about the game I’ve had dozens of times. And, of course, I have written about Rocksmith before, in fact at some length.

So, to be brief: it’s a great game. It’s a great way to learn guitar, it’s a great way to experience music. It’s a first iteration, a proof of concept, and that shows: there are usability problems and outright bugs, and I’m very glad that it sounds like the developers are focused on usability issues for Rocksmith 2014. But the first game very much proves the concept, and the usability problems are in no sense a significant barrier to enjoying and getting a great deal out of the game; there’s no point in talking about them further here.

What really fascinates me about the game, though, is its approach to learning. So, even though I’ve already discussed that at length once, I’m going to take another swing at it.

Because there’s so much going on here, and I don’t pretend to have it sorted out. I’m a much better guitar player than I was when I started playing Rocksmith, no question about it; but it is similarly certain that, in the grand scheme of things, I’m still not a very good guitar player. Which is an entirely reasonable state to be in a year later; how has Rocksmith helped, how has Rocksmith hurt, how have my habits helped, how have my habits hurt?

And, stepping back a bit, is my goal really to become a better guitar player? No, or certainly not exclusively: I’m playing Rocksmith because I enjoy it, and I have not committed to focusing on guitar excellence as a goal. All things being equal, I would of course prefer to be a better guitar player than a worse guitar player, and it is fortunately the case that, in this instance, doing what I enjoy is not completely incompatible with improving my guitar skills, but those two also aren’t completely aligned either with each other or with playing Rocksmith.

But, honestly, I don’t always know to what extent Rocksmith‘s behavior is aligned with learning. Take, for example, one of my favorite aspects of the game: the fact that its default mode is to present you with setlists of songs together with a target score to reach in each of them. This is an experience that is completely foreign from anything a human teacher has ever asked me to do (though, these days, when I sit down at the piano I generally pull a random book of music off the shelf and start playing, instead of focusing on a song); it’s also something I really enjoy. If we think of the goal (or at least a goal) of the game as to present a different way to enjoy music, then, for me, being asked to play through a random collection of songs meets that goal well.

There is something that’s at least potentially missing from that goal, however: learning to play those songs well. That’s not completely absent: the game does present you with a target score to reach, and there are times when that target score stretches you just enough. But a much more common experience is that the target score is low enough that I can succeed on the first try of a given view of a song (the target score caps out at 90,000 points, which is nowhere near perfection even on a “correct note” metric for most songs); I probably improve my playing of that song a little even when I’m doing that, but not very much. And there are also times when the target score is too high (which is especially frustrated when the song in question is one that I don’t enjoy listening to and/or playing); the game gives you an out when that occurs, lowering the target score, but the process isn’t always fun or didactically helpful.

When the target score is an appropriate stretch on a song I like, though, it’s great. That happened a lot during my first few months with the game: I had a lot to learn about playing the guitar at a basic level, so I was always below the 90,000 point level, and every time I played through a song I felt like I was getting at least a little bit better. And it still happens sometimes: the game will present me with a target that is higher than I can currently perform at but not outrageously so, I’ll play through the song a couple of times to get a feel for it, and then I’ll pick a section or two to focus on in “Riff Repeater” mode. After going through those sections 15 or 20 times, I’ll be noticeably better at them, at which point I’ll return to the song and be able to play it better than I could before, as measured by both my ears and by the score.

Stepping back a bit, one question this raises is: how does the game compare to a human teacher? A good teacher is doing a few things:

• Presenting an appropriate challenge: knowing when to push, knowing when not to.
• Giving focused advice about what you are doing well and what you aren’t.
• Modeling good performance.
• Connecting on a human level: sharing your enjoyment of beauty and of success, sympathizing with difficulty.

Rocksmith does a surprisingly good job on the first, but it has trouble beyond a certain skill level. Though even there it has some advantages over human teachers: human music teachers that I’ve worked with choose songs based on the overall difficulty of the entire song, and I think Rocksmith‘s ability to not only present stripped-down subsets of the song but to strip down the subsets more in parts of the song that give you more trouble is a really useful tool. (Especially given that this is rock music, where solos can be dramatically more difficult than the rest of the piece.)

Rocksmith isn’t great about focused advice. (Though this seems like at area where the game could improve; judging from a preview video, I’m optimistic that Rocksmith 2014 will do better.) It will give you feedback both via an overall score and by telling you which specific notes it thinks you got wrong, but beyond that, you’re mostly on your own. It does give you minigames based on specific techniques, if you want to work on them; I haven’t spent too much time on them, and I think there are problems with the implementation of some of them, but it’s a potentially promising idea? Ultimately, though: the game only models so much about playing well: you can play a song in a way that barely manages to hit all of the notes, managing a sort of sloppy lifelessness, and the game will still give you full marks.

Though there too there’s an aspect that is good: if the game is ever unsure about whether to give you credit or not for playing something correctly, it always errs on the side of thinking you played correctly. This is very important, and something that Rock Band 3 got wrong (at least measured from a didactic point of view): there is nothing more frustrating than doing something correctly and being told that you didn’t. I hope the game improves in its ability to correctly say that you’ve done something wrong, but there will continue to be a gap.

And there will also be a gap in another thing that human teachers do in this vein: suggest that you work in a specific area. The game currently does this to some extent: it will not only suggest songs for you but it will also suggest technique drills for you, and will even suggest subsets of songs to focus on. What it can’t do is suggest that you think about playing the song in a different way or in a different style within the range of correct possibilities; that sort of higher-level conversation can be really valuable with human teachers.

In terms of modeling good performance, the game does one big thing right: you don’t play a song alone, you play it along with the original band’s performance of that song. There’s room for improvement there, but that’s something I really like about it. Still, it’s always fun to play with actual humans listening and reacting to each other in real time. Then again, the game’s multiplayer mode allows that as well; I’ve only done that once, but it was a lot of fun.

You’d think that connecting on a human level would be completely absent: this is a machine, after all. But it’s not: part of the game giving you challenges to overcome is setting up a celebration when you’ve overcome those challenges, and I like that. (I’m not a fan of points as extrinsic motivators, but points as feedback is entirely different, and to me points setting up goals setting up celebrations is subtly different.) Also, the game’s forgiving nature, its tolerance of mistakes and willingness to see the seeds of correct behavior, is a personality of sorts, and one that I think is very helpful.

This next bit is mostly a digression, but: it’s interesting to see Rocksmith‘s take on memorization, given my interest in that. The game does try to encourage you to learn songs: once you’ve demonstrated that you can play a song adequately, it unlocks a “Master Mode” for that song, where you don’t see the notes for the song. (And you get double the points, if that matters to you.)

Which I liked at first: partly as an acknowledgment that really learning a song is a good goal, and partly because it opened up a rather effective way to memorize a song. Because what I would end up doing is practicing a song with the notes visible, and then practicing it again in master mode; that would then reveal how many parts I felt unsure about. And, if you don’t exit at that point, the game will then replay your performance for you; but this time it shows you the notes. So I could pay attention to the parts I missed, and next time I would do better.

That’s all well and good; the problem that selecting master mode isn’t always optional. When you finish playing through a set, the game throws an encore song at you; and, once you’ve mastered a few songs, the encore song is almost always a song that you’ve mastered, and you’ll be forced into master mode. And, if you don’t actually have that song memorized, you probably flounder and won’t enjoy the experience. So there turns out to be a disincentive to unlock master mode on songs; which in turn means that, as I buy more DLC, I sometimes almost regret buying easy songs not for musical reasons but because I’ll be forced into master mode against my will. (Sometimes after playing through the song as little as two times: really, being able to sight read an easy song is a very different things from having it memorized!)

So they made some bad interface choices there: there should be a way to never inadvertently be forced to not see the notes for a song. Having said that, I’d like to see a game that takes memorization further, using spaced repetition ideas: the game is already suggesting songs to me and paying attention to how well I do, it would be pretty cool if it could combine that with techniques to help me memorize songs.

Which I doubt will happen! But I am curious what Rocksmith 2014 will do with master mode: based on the previous videos, it will do that on parts of the song instead of being an all-or-nothing thing, and it will drop out of master mode when you mess up, both of which sound like good things.

We’ll see how that turns out. The main thing that I’m worried about, is being in a situation where I am playing a section well but I don’t know all of the notes, and the game doesn’t offer to show me the notes unless I intentionally mess up. Still, no sense worrying about that in advance.

Kathy Sierra recently posted an interesting series of tweets about becoming highly skilled. They’re light on detail, but still: something to put next to how I’ve been behaving with Rocksmith. (Setting aside the obvious fact that Rocksmith isn’t designed to turn people into experts at all, it’s for much more basic learning than that.)

Take, for example, her claim that “It is nearly ALWAYS far more effective to take 1 tiny useless-on-its-own subskill at a time to high reliability vs. pile of mediocre skills”. I suppose Rocksmith‘s game modes are compatible with that, and maybe the riff repeater focus on sections?

She followed that up by saying that that’s “Counter-intuitive for most as it means postponing ability to practice the whole thing while working on seemingly disconnected bits…” So, if I’m focusing on getting good at guitar, spending so much of my time playing through whole songs is probably not a great idea.

The thing is, though, I’m not at all sure that that is my goal. Don’t get me wrong: all things being equal, I’d like to get better, and I’m reliably putting in time practicing. Having said that: I’m playing because I enjoy it! So if what I enjoy is playing songs, I should embrace that on its own terms.

But I really do like playing songs well, and I like the physical act of having my fingers develop their dexterity. Even when I’m working on a single song, I enjoy focusing on sections of it that are giving me trouble and improving at them. So probably the balance of how I’m spending time is a little off.

Anyways: it’s been a great year; I’m looking forward to spending my next year with the next iteration of the series.

Over the last few years, I’ve become more and more interested in Apple from a business point of view. Some specific questions that I’ve been wondering about are:

• How is Apple going to defend and extend its place in the current phone ecosystem as the capabilities the iPhone provides become more and more mainstream?
• What’s the next area that Apple is going to be moving into?
• Are they going to stick with having their key product releases occur within a couple of months at the end of the year?

I tend to think about the first two of those through the lens of Christensen’s disruption theory: the first is the question of when their phone will be good enough and how Apple will react to that (and possibly the question of what the next disruption will be within phones); the second is largely the question of what disruptions will happen outside of phones/tablets. And the third makes me wonder from a lean point of view: a one year cycle time already seems a little long to me (though maybe not for major hardware—that’s the cycle Toyota works on, after all!), and Apple’s demand isn’t at all balanced within that cycle time.

And this year’s iPhone launch has gotten me thinking again about all three of these questions, answering them in different ways than I would have before the launch; so it is, perhaps, time for another installment my biannual foray into uninformed Apple punditry.

Before the announcement, I assumed that the 5C was going to start attacking the prepaid market in ways their current product line doesn’t: but, of course, it didn’t. Instead, it took their previous strategy of making the last year’s model available at a $100 discount, and made that model significantly more attractive. (Both from a purchaser’s point of view—the guts may be the same as last year’s 5, but the style is new—and (presumably, though probably to a less significant degree) from the point of view of Apple’s margins.)

So: Apple is still not addressing a large part of the phone demand as segmented by price. (Here’s an article written right before the announcement that gives a good example of thinking from that point of view. Though they did take one step in that direction: in China, they’re now selling three-year old phone models instead of just going back two years, bringing their floor price down to somewhere around $350.) Instead, the 5C announcement is about recentering the iPhone line: there are two variants instead of one variant plus old models, and the slightly cheaper variant is the more attractive one.

Here’s an article from Asymco about the 5C/5S split; I agree with Horace Dediu that this split signals that Apple thinks the iPhone 5 was good enough from a technology point of view, that the technology is mature enough in important ways.

This sets up an analogy with their Mac product line: we can think of the 5C as being like the lower-end Mac line (formerly MacBook, now MacBook Air), and the 5S is like the MacBook Pro. And, if I had to predict, I would suggest: Apple will switch to an iPhone price segmentation strategy that looks quite a bit more like the MacBook segmentation strategy than their current “sell old models” strategy. Right now, the two new phone models are only $100 apart (or only 20% apart, though the 5S also offers larger memory configurations that can increase that); that price difference too small to me, I expect that to grow in future years, perhaps with the 5C going down to around $500.

Given that, I don’t think that the “sell old models” strategy is going to be around for too much longer: it’s not the sort of thing that Apple does in their longer-established product lines, and I don’t think it will stick around here, either. Which brings me to the third point that I mentioned above: the product cycle timing. There was a very interesting post on Asymco last December called Does S stand for Spring?, suggesting that Apple might switch to a semi-annual release cycle; that has turned out not to be true (at least in 2013), but there was a lot in the post that made sense.

Having said that, I think it’s natural enough to accept that hardware refreshes for a single line won’t happen more often than once a year; the big problem from a level-loading point of view is having them all happen at the same time. So: why are they all in the fall? Because that’s the key buying / gift season, so you want to be selling your most attractive product then, which means it should be new.

And I think that reasoning continues to hold for the mainstream line (the 5C and its successors). It’s not clear to me, however, that it holds for the pro line (the 5S and its successors), however: that line is more target at people who want the latest and greatest from a technology point of view, and I don’t see any reason why that audience wouldn’t buy the phone whenever it is released during the year.

So: what if the S line moved to a spring launch, while the C line stayed on a fall launch? That would help level out the demand; and, from a technology point of view, it would also let Apple put a bit more room between price segments. Assuming the technology shifted in lockstep from the S down to the C (which I’m not convinced is the case), it means that the C would either be half a year or a year and a half behind the S; given that we’re entering the “good enough” phase, I think the latter would be much more likely. And, at a conversion rate of 1 year = $100, that means that we would have a $150 price spread between the two models: the C line could move down to a $500 price point (or potentially $450 if Apple wanted to shave margins), or the S line could potentially go up to $700. (Incidentally, I think the same sort of split and changed calendar schedule makes sense for the iPad as well.) And, with that wider gap, the need to sell older models diminishes.

That assumes that we stick with two lines, but I don’t expect that to be true in perpetuity. I wouldn’t be at all surprised if we saw a larger iPhone as an option in the next S launch: that feels entirely consistent with its hypothesized pro positioning. More interesting is the price gap beneath the C line: Apple could continue to fill that with older models, and I think they’ll do that for another year, but that strategy feels like it’s going to get harder and harder to fit into Apple’s positioning.

One option is for Apple to continue to not attack the cheaper market; that’s what they’ve done with the Mac, so it would be consistent enough. But it’s not what they did with the iPod; assuming they can find a way to carry it off without significantly decreasing profits in countries where non-prepaid sales mask the true product price, I think Apple will want to go cheaper.

So: when will we see a new, significantly cheaper product line? I’ll propose that we stick with the idea that the iPhone 5 is good enough, and that “significantly cheaper” means “at least $200 cheaper than the 5C”. (Both because a $200 difference will reach a much broader market and because it will justify significant product design differences.) If we put those two together, then maybe we’ll see a third iPhone product line two years from now: reusing a variant of the iPhone 5 guts yet again, coming it at around $350.

Putting this all together, here’s a picture that makes sense to me two years from now:

• A $300 line, probably released in the summer or fall.
• A $500 line, evolving from the 5C, probably released in the fall or summer.
• A $700 line, probably released in the spring.
• A larger screen variant in at least the expensive line, possibly also in the middle line.
• Similar differentiation and timing for the iPad, though probably with only two lines instead of three

That addresses the first and third of my bullet points above; what about the second one? Honestly, I have no idea; like everybody else, I expect something wearable and something TV-related to show up soonish, but who knows exactly when.

The 5C/5S launch didn’t shed any real light on that; but Tim Cook went out of his way to talk about how forward-looking the new products were. Assuming that we treat that seriously, I’ll once again follow Asymco and suggest that the M7 coprocessor is pointing in the wearable direction: its appearance in the 5S is mostly a sideshow. I have no ideas about what the new wearable product will look like, but I’d be quite surprised if the M7 or its successor didn’t play a prominent role in it.

Then there’s the A7 chip. It mostly feels like an obvious evolution—chips get faster, and even given that today’s products don’t benefit much from 64 bits, we’re only a year or two away from wanting the larger address space. But there is once product category that Apple might move into that could use a 64-bit address space right now: if a new Apple TV is going to make a serious play at attacking the traditional game console space, then more memory would be welcome. (Not my idea, though I can’t remember where I heard / read that suggestion.)

Having said that, while I do expect an Apple TV with apps at some point, I’m not convinced it will attack current game consoles quite that directly. (Though it might: right now, the Apple TV is priced like an iPhone accessory, so they may have to make it more expensive for an app platform version to make sense, and traditional games might be part of that.) Clearly Apple is going to disrupt the game console space, but they’re already doing that; I’m not sold on the idea of running games on your iPad / iPhone and streaming them to the TV (my personal experience with that has been quite unsatisfactory), but I certainly wouldn’t rule that out as the route for disruption to play out.

And it’s probably a sign that more chips are coming; I heard speculation somewhere recently (I can’t remember where) that voice processing was a natural next place for Apple to get a boost from custom silicon. After all, wearables, the TV, and car integration all need different methods from the iPhone/iPad, and an evolution of Siri is an input route that would work in all of those contexts. (Horace Dediu has made that point repeatedly on The Critical Path over the last year.)

So, if I were to move beyond the iPhone/iPad the vision of what the product line will look like two years from now, I’ll suggest:

• Some sort of wearable product, using something in the M7 family and potentially using still more custom silicon.
• Some well-thought-out solution for apps on the TV; I tend to think that a souped-up Apple TV with a 64-bit processor (and perhaps with custom silicon for voice input) is likely, but I’m honestly not sure.
• The former will be more surprising than the latter, but I would be surprised if I were wowed by either of them. Who knows, though; one point of disruption is that it’s hard to imagine in advance…

It took me a little while to warm up to Ascension: Immortal Heroes, but I eventually ended up liking it quite a bit, about as much as the second expansion. It first seemed like the new mechanic, randomly acquiring a card from the original game that was valid for one turn only, added a bit too much randomness, but as I played more that turned into one more thing to take into account into my calculations.

More importantly, the more I played, the more I liked the balance between the factions: Mechana constructs didn’t seem to dominate, Lifebound heroes did if you could chain them together right but judging the chances of that happening took some doing, and events gave enough of a boost to the other two factions to make them interesting. The original game was in large part all about banishing; that was much much less important in this iteration, but it was still present (and still occasionally a dominating strategy if the cards fell correctly), with timing the switch from banishing to profit becoming rather more important.

So: well done, I look forward to the next expansion.

Liesl and I are in the middle of a reread of the Kushiel series, and the original trilogy continues to be, in its own way, great. I thought I’d written more about it here, but it turns out that I’ve only barely touched on the series.

So I think it’s time to touch on it again: I’ve been complaining recently about heroic motifs in video games, so perhaps I’m being inconsistent in liking a fantasy trilogy this much? Admittedly, there are video games in that vein that I’m very fond of indeed; I just think as a whole this idea of saving the world in games is overdone.

And, indeed, Phèdre does save the world (for some value of “world”), once per volume, just like in most fantasy role-playing game. And there’s a lot of fighting going on. But even in that area, it’s significantly different from most RPGs.

For one thing: in most RPGs, you’re leaving a trail of blood behind you. But Phèdre is rarely the person doing the fighting: she kills somebody in one notable scene in the third book, but beyond that I can’t think offhand of an instance where she fights people physically. She’s on the receiving end of violence rather more often, but that gets critiqued in a rather more subtle way than almost any game I can think of, using her sexuality to disentangle concepts of violence, strength, weakness, and consent.

Of course, she spends a lot of time with Joscelin, and he engages in violence significantly more frequently. But even for him killing is a quite rare act, and one that he doesn’t undertake lightly; contrast this with the constant drone of unthinking death that almost all RPGs serenade you with. (And note the increased narrative weight that Joscelin’s killing gains by its rarity.)

RPGs also have an ever-present leveling up grind; but Phèdre and Joscelin are about as capable halfway through the first book of the trilogy as they’re going to be by the end of it. Admittedly, that’s the way RPGs feel as well with their auto-leveling enemies, but here even the pretense of progression is largely gone. Or at least, the pretense of progression as measured by traditional RPG attributes.

Because there is quite a bit of progression over the course of the book: progression in understanding oneself, progression in understanding others, progression in understanding their relationship, progression in understanding what love means. And that progression of understanding comes with constant, active choice, with a meditation on the nature of choice: this isn’t abstract understanding, this is understanding that is constantly tested and enriched by surrounding circumstances, that comes from looking at the actual people you’re interacting with, observing your reactions, and actively deciding where to go next from that.

Which sounds sterile when I write it, but to me the series comes off as one of the most moving depictions of love that I’ve seen. Not just the first blush of romance, but the love of making a life together, and making a family together. And a family that you actively choose: the third book is all about choice in the context of the relationship between parents and children, and does so in a way as to not remove agency from either side.

Plus: lots of fucking. I’m also grateful to that aspect of the trilogy, but I won’t go into details about that here…

As I mentioned last month, I was thinking of hooking up my laptop to the TV so I could play System Shock on it. That idea didn’t go away; so I ordered a mini-DVI-to-HDMI converter and a lap desk. It took me a little while to get around to setting things up—I was playing the new XBLA Magic game to get ready for a VGHVI session—but eventually I had an evening free, so I went to plug things in.

That actually took longer than expected: I hadn’t looked closely at my receiver, so I didn’t realize that I didn’t have ports available that would let me combine HDMI video with RCA audio jacks. Eventually I gave up on that, and plugged stuff straight into the TV; that worked fine. Or at least mostly fine: when I turned it on, I was missing the sides of the screen, but I’d been through that drill before, so I knew how to turn off the TV’s overscan settings, I just apparently had to do that separately on the TV’s different inputs.

So: I fired up DOSbox, and set it to full screen mode! And: the sides of the screen were cut off again, and this time there weren’t any options on the TV to help me.

I of course googled the problem; I found nobody else describing those specific symptoms. There were some general recommendations for which values in the configuration file to fiddle with; I tried various ones of those. Most of them had no effect; then I tried appending the “no, I mean it” option to one of them, and the screen looked right! So I started up System Shock; the dimensions were still good, but the framerate was glacial.

I spent an hour and a half fiddling with config files, and then I eventually gave up and played on the laptop screen. (Actually, that’s not quite right: I tried on the full screen, and there wasn’t too much cut off. So I’m actually thinking that I’ll give it a try again on the TV soon: now that I understand the game’s HUD, I should be able to play with parts of the sides and bottom cut off.) At least I feel like I got an authentic PC experience out of the evening.

Still, it was the right choice. Mainly for one reason: System Shock is a wonderful game, and is the game that I’m in the mood to play now. (And I will continue on to System Shock 2; hopefully the Good Old Games packaging will help that work better, or maybe being in a higher resolution will help something in the display chain behave better.)

But also: it’s not the only PC game that I want to play. Especially not right now, with the recent release of Gone Home; I’d also like to play Analogue and The Yawhg. Admittedly, I’d rather play those games on the 360 or the iPad, depending, and I’m probably still going to wait a while to see if any ports appear, but I may have to take what I can get.

Good times? Well, it’s far better to have a way to play a 1994 game than not to.

Edit: Thanks to Praveen’s comment below, I tried out Boxer; it’s working great in fullscreen mode. So yay, I’m actually pretty happy with the setup now.

I’ve been putting the standard “jōyō kanji” in my collection of items to memorize, adding seven a week; this week, I finally reached the end of the list. I’ve skipped the occasional week, but not very many; that suggests that it’s probably taken me about six years. Long enough, in fact, that the list was updated once during that time period, adding an additional 196 characters.

The book that I’ve used as reference for this is Hadamitzky and Spahn’s Japanese Kanji and Kana, and I highly recommend it; I actually own all three editions of it (I got it when I was flirting with learning the language back in grad school, and it’s been revised as the list has expanded), and it’s very solid. It’s a good presentation of the basics of Japanese writing (not just the kanji list, but also information about kana, punctuation, radicals, etc.), it presents all the characters cleanly (including stroke order), it gives a nice selection of readings and definitions for each character without overwhelming you with possibilities for either.

And I will say: it’s a little odd for a six-year project to come to an end! Nice to have evidence that I can stick with something like that, though. But of course the project of learning Japanese as a whole hasn’t come to an end: which raises the question of what, if anything, I should replace this with?

One possibility is: more kanji! I have another dictionary around; it looks a lot larger, but there are only 5446 characters in it. And I already have learned some number of characters that aren’t in the standard list, so I’m probably around halfway through the book already, maybe more. Another possibility would be to go back through the standard list, but this time focusing on writing: you don’t really know how a character if you can both recognize and produce it, and even if you only care about reading, if you can’t write a given character, it’s easy to mistake other similar characters for that one.

I think, though, that I am not going to pursue either of those paths. Memorizing is all well and good, but I should focus on using the language, not learning for the sake of checking off boxes. So I want to spend more time reading, and have my memorization driven by that. I’ll add vocabulary words as they come up and seem important, and I’ll probably go more out of my way to add individual characters in compounds if I’m not already familiar with them, but I should really be spending time putting as many sentences in front of my eyes as possible. (I probably should be spending time speaking, too, but for better or for worse I’ve decided not to go back to lessons for the time being.)

I’m actually trying to dial back the memorization in other ways: in particular, I flirted with learning Chinese starting three or four years back, and while I won’t say that was a mistake, it’s also not something that I want to spend more time on beyond listening to podcasts a few days a week. And I’m a lot less solid on Chinese vocabulary than Japanese vocabulary, which means that they show up rather more often in my memorization quizzes than I’d like; so I’ve started deleting almost all of the Chinese words as they show up. (I’m leaving in core Chinese vocabulary, numbers and a few other basic words/expressions, but getting rid of everything else.)

Instead, reading. I’m over halfway through Hikaru no Go; really enjoying that (enough that I sometimes read some after I come home from work, not just over weekends), and it’s at a good level for me. I also found two bilingual collections of stories that look good, I have high hopes for them. (I’ve just started the first one.) I’m not entirely sure what I’ll replace Hikaru with once I’m done with it; maybe Buddha? (I picked up a copy of the first volume on the trip, we’ll see how the reading level is.) I expect I’ll stick with books where I have an English version readily available for some time (though I have hopes that maybe I’ll be able to read Twelve Kingdoms well enough that I’ll eventually be able to make it beyond where the English translations run out, but who knows.

It’s been a fun trip.

It probably would surprise people who have interacted with me recently to hear it, but I actually spent a fair amount of time a few years back trying to get good (as an individual, as part of a team) at estimating: reading the literature (both agile and otherwise), trying it out, and refining and repeating in an attempt to get better. These days, I do not care about estimates nearly as much (or at least I don’t care for estimates nearly as much!); that’s basically because I haven’t seen spending time on estimates as a particularly direct road towards doing software development well as measured in terms that matter to me. I understand estimates in the context of, say, XP, but within XP I see them as a means to the end, and other approaches to those ends seem more productive / fundamental to me.

Still, estimates aren’t going away (and, incidentally, Ron Jeffries recently wrote a very interesting post that touched on the subject); so, it’s time for another installment in the ongoing series of “David tries to understand his reactions to some aspect of software development practices”.

I actually don’t think that I would react negatively to all discussions of estimates. To that end, I’ve come up with three aspects of any such discussion that are important to me; if the discussion of estimates doesn’t acknowledge at least two and probably three of them, I’ll probably start off with a strong instinctive negative reaction. But if it touches on all of them, I’ll be a lot happier. Those aspects are:

1. Why are we estimating? How will we deliver software differently if we do estimate than if we don’t?
2. What does the estimate mean? We can’t predict precisely when something will be done; so what’s the underlying probability model, and how are we boiling down that probability model down to one or two numbers?
3. How are we getting better at estimation? Estimation is hard, but with feedback it is possible to improve.

These are, of course, interrelated, but probably the first of them is the most important; as a mathematician, though, the second one of those bugs me the most, so I’ll start there.

So: what is your probability model? Honestly, I’ll be happy with any initial interaction that shows that you are thinking from a probabilistic point of view at all. Here is my stab at what I think probability distributions in estimating software projects are like:

• Error models are more likely to be multiplicative than additive: i.e. rather than being off by plus or minus two days, you’re more likely to be correct if you think that it’s equally likely to take either twice as long or half as long as your estimate.
• I don’t believe that it’s multiplicative either, though: things are more likely to go catastrophically bad than they are to go gloriously well.
• If your estimates are accurate to within a 2x factor more than, say, three quarters of the time, then you’re doing a good job at estimating. But for some kinds of work, even that is hard to attain: bug fixes in particular are notoriously hard to estimate.
• Software engineers who haven’t actively practiced estimating are very unlikely to get close to the most likely spot on the estimation graph: instead, we’ll default to saying the fastest time that we can imagine something being done by.

Or, to reduce my discomfort to a smaller number of bullet points: if you’re asking me for an estimate, either ask me for a range or specify what portion of the curve of answers you want to be to the left of the number I provide. I can work with either of those.

That’s the second aspect. Moving back to the first aspect, here are some possible good answers to the question of why we are estimating:

• We are planning external commitments (a press release, a contract) around work that isn’t yet completed.
• We are trying to choose between tasks that are of approximately equal business value, and we want to know which can get done faster.
• We want to have a discussion about what it means for a piece of work to get done or what the candidate pieces of work are at all, and we are using estimates as a sneaky Jedi trick to help that discussion along.

And here are some less good answers to the question of why we are estimating:

• We have a particular fascination with predicting in advance exactly what collection of work will get done one or two or three or months from now.
• We see other people doing it, so we’ll do it as well.
• We think programmers are lazy, so we want to use estimates as a tool to fight that.

I may be somewhat eccentric in putting the first of those less good answers on the less good list: that’s garden variety Scrum or XP. The thing is, though, it strikes me as a quite difficult task that won’t actually affect what you’ll do over that time period. Whereas stripping down that question in one way or another (asking only about a subset of the tasks or else relaxing the time bound) do lead to actively actionable choices: they’re basically the first and second choices on the good list.

I think the third answer on the good list is probably the best answer of them all; I wish I were better at that. (Here’s a good post by Esther Derby on the subject.) I think the third answer on the second list is the most common subtext for questions about estimations.

One reason why the choice of question matters even within the range of good answers is how it interrelates with your answer to the probability space question. If you’re making an external commitment to getting something done, I would recommend that you aim for a point on the probability space that’s pretty far to the right of the curve, rather than putting down a more aggressive estimate and flipping a coin. And I would also recommend only treating a small subset of the work items that way: that way, you can drop other items in order to improve your chances of meeting the date. (So I certainly wouldn’t recommend talking about a commitment to a 50% estimate for an entire iteration’s worth of work, as some Scrum treatments seem to do.) Whereas if you’re choosing between items of equal business value, then you’re probably better off trying to figure out the middle of the probability curve. (And you might want to compare the variance of the business probability curve with the variance of the engineering probability curve!) If you’re going for the third answer, then the fact that different types of work have significantly different probability curves can lead to interesting discussions.

And then there’s the question of getting better. I don’t have a lot to say here, because it’s mostly pretty obvious: pick a model with some parameters missing, gather data, and then try to figure out what the actual values are for those parameters. Story points plus velocity is one traditional way; skipping prediction entirely and measuring cycle time is a way that I’ve been curious about recently; I’m sure you can come up with other ideas. The main pitfall is not having a model at all, and the second main pitfall is leaving work out of your measurements without taking that into account in your model.

Actually, the real main pitfall is not wanting to improve at all: I tend to suspect that that’s associated with my third bad reason for wanting to estimate. Though even there, talking about improvement can potentially lead in interesting discussions: if you end up leading in a direction saying that you expect programmers to give optimistic estimates and to work evenings and weekends if those estimates are wrong or don’t meet your business desires, then I would rather have that underlying assumption be overt than not, and maybe talking about estimates will help bring that out.

So, to summarize: if we have a discussion about estimates and spend time talking about why we’re estimating, what our probability model is, and how we’ll improve our ability to estimate in a way that is congruent with those starting points, I’ll try not to be a jerk about it. (Though I may probe at whether the “why we’re estimating” answer is pointing at an underlying need that we should be attacking in other ways than estimating.) If the discussion starts from a different starting point, though, my reaction will probably depend more on my underlying emotional reserves that day.

I got off on the wrong foot with The Unfinished Swan, starting from its very first storybook page: that page begins with the sentence “The King was young, arrogant, and amazingly talented”, and that set up too many associations for me. Associations with my personal life: I spent a lot of time doing math contests as a teen, and those are full of the young, the arrogant, the talented. (I certainly qualified on all three counts.) Associations with my work life: the Silicon Valley startup scene fetishizes the young, the arrogant, the talented as well. And The Unfinished Swan was a finalist in the Student Showcase of the Independent Games Festival; based on that, it’s not much of a stretch to assume that some of the key people involved in making the game are young, they’re certainly talented, and arrogance on their part would not be unjustified.

So: the game is true to my life. This is good, no? Well, not so much, or at least not unconditionally: it’s territory that I can’t accept any more on face value either on artistic grounds or or political grounds.

I’ve been coming back to the concept of “adolescent games” over and over again for the last three and a half years: games where a boy becomes a man, discovers unexpected powers within himself, and saves the world in the process. A classic theme, but a theme that is far too overdone, and that in the aggregate I’m finding politically repulsive. And, yes, a theme about boys who are young, arrogant, and amazingly talented.

The Unfinished Swan is not such a game. It doesn’t present a world full of people who need to be saved; it doesn’t involve said savior leaving a trail of corpses in his wake. And it puts front and center the loss involved in such arrogant devotion to talent: as much as anything else, the narrative aspects of the game are about the King being estranged from his wife and son, the blind spots that his monomania has given him. So it’s entirely fair to say that it’s a critique of adolescent games, that it’s moving beyond them in important ways.

I suspect that, if the game had come out four or five years ago, I would have been significantly more forgiving of it in that regard, quite possibly actively impressed. The thing is, though: the last several years have brought forth a bunch of games about daddy issues (from both sides of the father/child relationship), as game developers are getting older and writing about the regrets that come from the time they’ve spent crunching on their games instead of being with their families. So The Unfinished Swan isn’t breaking new ground within video games in that regard: it’s the latest entrant into what is now something of a tradition.

And, as entrants go into that tradition, it’s not a particularly finely drawn portrait of such a relationship. Compare the King in The Unfinished Swan to Thane in Mass Effect 2: both made parenting choices involving pursuing their talents in ways that estranged them from their sons. But Thane’s choices and pain felt real to me in a way that the King’s choices never did; real enough that I ended up falling in love with him a little. In contrast, the loss in The Unfinished Swan felt to me like the words were there without the feeling behind them, without the details to make them real: a paint-by-numbers drawing instead of a portrait of a life.

But as much as I loved Mass Effect 2, that daddy issues theme is one that I’m also growing tired of, even when it’s done well. Why not a game about the Queen? (Not just about a woman rather than a man, but about somebody who makes the choices that she did instead of the choices he did.) What about somebody who sees these potential pitfalls in advance, who tries to avoid them without going too far in the other direction and subordinating themself to their child’s needs, who sometimes stumbles and falls in that attempt but nonetheless finds that attempt, that journey incredibly rewarding? (And rewarding in ways that don’t involve treating the child or the relationship as a puzzle box to be solved.) That would be a game that I would find true to my life, and a game that would explore relatively untrodden design space.

So, like I said: I got off on the wrong foot with The Unfinished Swan. But that doesn’t mean that it’s not a good game: I was actually really impressed with it. The basic mechanic is one that I love, both for its freshness and for its metaphorical power. Just read through my paragraphs above, with their mentions of blind spots, portraits, and painting by numbers: the metaphors practically write themselves, the game’s mechanic and spare art style lends itself so very well to those discussions, almost becoming self-commentary. (And, to that end, any portrait of a relationship that this game could present well would necessarily be somewhat elliptical: this is not a game that wants to fill in all the details.)

Take, for example, the question of finding secrets in games. In the very first level, as you start splattering the walls, the question immediately comes to mind: are there secrets hiding in the white spaces between your paint splotches? And, if you’re the sort of completionist player that I am, this sets up a real tension, which you resolve by trying to fill in the blank spaces, putting paint everywhere. But, as you do that, the scenery turns from pure white to pure black: you can’t see the secrets that way, either, they’re hidden just as well in either case! So it’s a wonderfully wry commentary on my completionist behavior; commentary that actually got me to relax and accept the fact that I was going to miss some secrets.

The game’s spare art style was also reflected in its pacing. It consists of four scenes in four areas of the world; each of them is long enough for you to get a feel for that area of the world and for the associated mechanic, but each of them also ends well before you feel that you’re in any danger of overstaying your welcome. And, while these scenes are all distinct, they are drawn together by a narrative thread coming from the storybook pages and by a consistency of design.

In that regard, The Unfinished Swan reminded me a lot of Journey: almost everything in the previous paragraph applies to Journey as well, and in both cases it’s very much to the game’s credit. And, continuing the power of the visual metaphor, it also ties back to my orsay games post: that one powerful way of structuring a video game is as a series of paintings, each of which is a (view of a) scene to immerse yourself in.

So: more of that, please. And more games that take a first-person perspective and use it for something beyond raining bullets. But perhaps a view of some different lives next time?

Since my role shift at work, I’ve been thinking about teams: partly being surprised by my reactions to some interactions, partly thinking about ways in which my instincts are a bad fit for some aspects of our house style. And three questions about organizational team structure (in general, not about my current job) that I’ve been thinking about in particular are:

1. What are the groupings within the development staff that are coherent on their own? How does the organization think about individuals? About teams? About the development staff as a whole? (Let’s assume the organization is of a size where having multiple development teams makes sense.)
2. Within a development team, what are the divisions of responsibility? Does each individual own one slice of the code, is everything shared equally? I imagine almost everybody would say that they’re not at either extreme, but where on the continuum does the team lie, and what are the interesting different ways to think of that choice beyond a simple continuum?
3. How cross-functional are teams? Are all the members traditional software developers, or are there testers, product managers, project managers, interaction designers, other roles? For that matter, just within software developers, are the architectural responsibilities within the team, or are there people outside the team who have a strong say over how they should structure the software?

My current preferences are to treat the team as a quite strong grouping, to have as much shared ownership as possible among developers on a team, and to have cross-functional teams. And I’d been thinking about these as a naturally grouped set of opinions (with the second and third following from strong teams).

Now, though, I’m less sure about the natural linkage. I mean, I understand why I link them—it’s all pretty standard eXtreme Programming stuff, and XP has had a huge influence on how I think about software development. But the fact that they are linked in that one methodology doesn’t necessarily mean that they’re linked in general.

Thinking about it more, I think there is good reason that my answer to the third question might follow from my answer to the first question. If a team is a strong unit, it has to be able to both make and execute on important decisions. But you can’t really make important decisions on your own if people outside of the team are doing most of the deciding of priorities or are specifying the design (whether from a user-facing point of view or an internal architecture point of view). And, similarly, you can’t do a good job of executing those decisions unless you have the skills within the team to perform the various implementation tasks, and the judgment to steer you on a reasonable course in the small as well as the large.

I’m not nearly as sure about the second question, though: it seems to me entirely consistent for a team to say that it has decided that the best way for it to work is to divide up the responsibilities among the team members, with that division relatively stable. (In fact, to some extent that’s exactly the situation I was in earlier this year; some of the pressure for specialization came from outside the team I was on, but a significant portion came from inside the team.) Sure, having everybody work in the same area can contribute to team strength and unity, but having people work in areas that they’re not drawn to can just as easily distance them from others; and the main justifications I see for shared ownership within Agile thought are based on decreasing cycle times and decreasing risk, not on strengthening teams.

And, as I write this post, I’m realizing that, to some extent, the second and third points are in tension. If your team is all programmers, and especially if it’s all programmers working in the same general area, then sure, you can imagine that anybody could work on a given area of code if that were helpful. But the more cross-functional your team is, the more you’re going to have people who really can’t do each other’s jobs: when I worked on Sorority Life, I wasn’t about to start drawing dresses for the game, and the artists weren’t going to write server code.

Also, if (as I’m assuming) we’re having this discussion in the context of an organization that’s larger than a single team, then this concept of shared responsibilities makes a lot less sense for the interior teams: a first-level manager is going to be part of a team with that person’s peer managers, but those managers don’t have shared responsibilities at all: they’re managing different teams, they don’t manage whatever team happens to be convenient for them to manage on a given day! If you follow that path a bit further, you could even make the case that collective ownership has unfortunate political implications: once you’re a manager, you have an area that you can call your own, but individual contributors have to be considered as part of a collective pool.

Having written that, I still think that collective ownership is a good thing, certainly for me personally but plausibly for a wide swath of software development. And if I want to put a political spin on it, maybe it’s the reverse of that last train of thought that’s more accurate: if collective ownership is appropriate for individuals but not for bosses, then get rid of bosses! I was joking over lunch with one of my coworkers, saying that the two of us should form our own splinter anarchist team with no manager; we couldn’t get away with that, but it’s definitely tempting…

Though, when I say that I’m certain that this is what I prefer, what evidence do I have of that in general? Have I, in fact, ever been in a situation where I’m on a strong, cross-functional team with collective code ownership?

I’ve experienced aspects of that situation, no question. On StreamStar, I think we had pretty strong teams. People outside the individual development teams made decisions about the direction of the project as a whole, but within those broad strokes, the development teams had quite a bit of autonomy to decide how they were going to do things. (At least after we got acquired by Sun; before then, one of the cofounders handed down software design edicts by fiat.) And, at least on my team, we were pretty strong on collective code ownership (again, after we got acquired by Sun): different people had areas where they knew more, but if there was a big project, then everybody worked together on it.

That team wasn’t cross-functional, though: we had the programming covered well, but there weren’t any product people on the team. I didn’t really get to experience a cross-functional team until I worked on Sorority Life: there, we had programmers, testers, artists, game designers, and product managers all working together, seeing each other every day. In fact, that team did a great job of hitting my answers to all three points on the list: we were cross functional, we could make almost all the important decisions about the game ourselves, and the developers could chip in wherever was helpful. (That was the only job I’ve had where I did frontend work as well as backend work, and we all ended up being able to contribute on both sides.)

And: it was awesome? Well, it kind of was, actually: I had a lot of fun working on the game, I learned a lot from the experience, I enjoyed working with people who weren’t all programmers, we made a lot of money for Playdom, and I think the decision about architecture and about priorities that we made while I was on that team are a big part of the reason why the game was still doing well for Playdom a couple of years after I left despite its age.

But: I also left the team after nine months. So that certainly raises the question of whether what I say I want matches up with how I express my preferences in practice.

Honestly, though, I don’t think there’s a mismatch there. Part of it is that I didn’t just leave the team: I left the company. Playdom got acquired by Disney, it became a noticeably less pleasant place to work, and there were a lot of signs that it was going to get worse instead of better. Also, the work on the game itself had stabilized: in the first half year after I joined, we made what were in retrospect a quite impressive number of architectural changes, but in the last three months that I was there, it was starting to look like we’d done such a good job of reworking the architecture for the game that it just wasn’t all that challenging to continue to make our players happy and to make us money. And, talking to my former Sorority Life team members a couple of years later (or simply looking at the game), that seems pretty accurate: the same basic structure is still there. I’m sure we would have found interesting ways to improve the software architecture if I’d stayed (I’m sure people made interesting changes after I left!), but it’s not crazy to think that the product had stabilized.

So, yeah: I do like coherent, cross-functional teams with porous boundaries. But I also like working on software challenges. And, if I’m not managing to find away to have it all, I’ll just have to figure out what to jiggle…

I haven’t been following the Twine scene too closely: in my experience, Twine games generally feel like hypertext, and hypertext isn’t something that I’m particularly interested in. Porpentine’s work is distinctive enough to make me take notice, but even her work isn’t entirely to my taste, though I’m leaning towards a belief that that’s more of a gap in my taste than anything else.

I just played Depression Quest, though, and: wow.

So: what’s the difference? Part is that it does feel like more of a game to me: the systems are much more overt than in most Twine works, with the three-part status reporting that the game gives you and with the feedback loop from that status to the choices that are available. I’m sure that contributed to me playing through the game three or four times, trying out different paths, seeing what the game’s model of the effects of actions on your mental health is.

In the abstract, I could imagine that that systems modeling might make this sort of game too bloodless or too prescriptive a presentation of a specific analytical point of view; but Depression Quest didn’t feel that way to me at all. A big part of that was the way that it presents you with a list of options and then crosses off some of them (crossing off a fair amount even at the beginning, more if your mental state gets worse, fewer if things start to improve); I though that was an incredibly effective communication tool, a way of saying that it’s not as simple as saying “pull yourself together and do X”, that when you’re in bad health you frequently can’t do that.

But I found the game effective for reasons other than systems modeling, too. I suspect my reaction to the game would have been a lot more muted without the music; and the interactions between people that it presented felt real to me, too.

Very glad I played it. It only takes around 10 minutes or so for a playthrough, if you’re tempted to give it a try yourself.

I recently watched a video about the System Shock and BioShock games, and I was surprised how nostalgic the System Shock footage made me. I’d been tentatively thinking that I should play through the entire BioShock series (I’ve only played through the first game in the series), but now I’m thinking I should start off with its two spiritual predecessors.

The first question that raises is: how to get them? System Shock 2 (which I’ve also never played) is available on Good Old Games, and runs on the Mac. The original one isn’t available there, but it’s not hard to find on abandonware sites, and it’s even possible that my old CD is still lying around the house somewhere; hopefully I can run it on DosBox? So that should be fine.

Or at least fine-ish. Partly because I have a mental block against playing games on computers; partly because getting old games to work is a black art that I have no experience with; but partly because I prefer not to isolate myself while playing games, which is what I was thinking I’d have to do to play those two games.

That’s not necessarily true, though: while the obvious place to play computer games on is the iMac that we have upstairs, we do have a MacBook downstairs. And, while the computer is slow and in part held together by duct tape (it’s from April 2009, and laptops do not age so gracefully), it should still be more than powerful enough to run games from the 90’s. So maybe I should play the System Shock games on the laptop?

But that raises control issues: I would be pretty surprised if the games were playable with the trackpad on my laptop. So I guess I should use a separate mouse (and potentially a separate keyboard, though skimming the manual it looks like it doesn’t want you to have a numeric keypad at least?) And then a keyboard tray for my lap with room to put a mouse / trackpad?

Though even that is isolating: I’ll be playing the game in the same room as Liesl, but I’ll still be in my own world. So I guess I should try hooking the computer up to the TV? Dealing with the fact that the laptop comes with a Mini-DVI port, which doesn’t have sound, so I’ll have a few more cables than I’d like, but whatever.

I dunno; it’s more fiddling than I’d like, both in terms of running the game at all and in terms of having it display in a manner that I would prefer. Still, maybe for System Shock it’s worth it…

I’ve been thinking about earning, spending, and saving money more than normal recently. Part of this is getting older in a young person’s industry while having my employment be more contingent than normal over the last year; part of it is the lure of riches that the startup dream dangles in front of you; part of it is planning to remodel our kitchen and being confronted with how expensive that can get. (Cabinets! People charge a lot for them!)

It’s weird to think about how much my salary has changed over the decades and how little effect that has had on my quality of life. I had a quite good stipend in grad school (heck, the fact that I got paid rather than paying, even though I didn’t have to teach at all in three of the four years of grad school, puts me way ahead of most people in most fields), but even so: my salary tripled when I moved to my postdoc. And I also had a quite good stipend for a postdoc, but even so: my salary doubled when I left academia. My salary has been climbing much more slowly since then, but those little gains have added up, it’s gone up by about half since my first industry job.

There’s been inflation in the almost-two-decades since I started grad school, of course, and the Boston metro area, while not cheap, isn’t as expensive as Silicon Valley to live in. I’ve been living with Liesl the whole time; her salary has also gone up but not as much as mine has. And there are three of us now, not two of us. Still, those are quite substantial gains even taking all of that into account.

And what quality of life improvements have I gotten out of all of those extra gains? Sometimes, honestly, it feels like I haven’t gotten any improvements. I mean, I watched my bank statements a lot closer back then than I do now, but still: I really wasn’t depriving myself. I had plenty of books to read and games to play (and significantly more time to read / play them); we went out to eat not infrequently and sometimes those were at pretty good restaurants; we went to Europe two or three times when I was in grad school. All of that is far past basic necessities, well into luxuries. And it’s not like I’ve added more luxuries on top of my life now: my list of things I enjoy that cost money basically boils down to art and food, and my ideal vacation remains spending a couple of weeks hanging out in Paris.

So, that raises the question: how are we managing to spend so much more money? I can’t realistically return to my grad school spending levels, sock away 90% of my salary, and only work one year out of ten; but I don’t have a good argument for why we can’t sock away, say, half of our salaries. (Or a good reason why redoing our kitchen is something that I should be happy to spend somewhere around two to three years of my grad school salary on; I love books and games and music, but I kind of doubt I’ve spent that much money on those pursuits across all of the last decade.) I’m pretty sure that there actually is a reason in the short term why we can’t save half our salaries; but maybe I should do the legwork to figure out what that reason is. And then, once we understand the situation better, Liesl, Miranda, and I can figure out ways to put caps on spending that will save us money without making us feel deprived.

And, as I said above: I’m getting older in a young person’s industry. So there’s also a worry at the back of my mind that I might not have a choice about needing to stop spending money. I don’t think that’s coming particularly soon, and I also hope that I’ll be able to slide into a part of the industry that’s less youth-focused if that becomes necessary, but you never know: there are a lot of people who lost a job in the recession and then got shut out of the industry.

Fortunately, it’s not like we’ve been spending all of our money: we’ve been maxing out our 401(k) contributions for years and saving a smallish but noticeable amount of money beyond that. And, while the 401(k) contributions don’t look like much on any individual year, they do add up; I surveyed the accounts a few months ago (which took a while, I should really consolidate my 401(k)s and 403(b)s from my various former employers at some point!), and I was surprised how much they added up to. Not nearly enough for me to be confident about retirement yet (and that in itself is an interesting problem: how should I model whether a give amount of money is enough to retire on?), but while I’m solidly into middle age by now, I’m not particularly close to a traditional retirement age, so that’s fine.

But finding ways to increase our savings rate would be good. (And then there’s college costs presumably showing up starting four years from now!) I did at least spend some time over the last month finding some investment options that I’m a little more confident in than what we had been doing, and we set up some automatic withdrawals into those investments that will eat up a decent portion of my last raise. So that’s a start; but there’s definitely room for improvement.

I don’t want to get obsessed with this; and I certainly don’t want to chase higher salaries in the hope that that will turn into later improvements in my life from increased savings. But I would like to increase my mindfulness in this area, instead of unthinkingly spending the money that comes in.

Like many other people, I switched to Feedly as my Google Reader replacement. The import worked well; as part of that process, in the “Uncategorized” category, I found dozens of blogs that hadn’t published for years.

I removed most of those subscriptions, but some of them I couldn’t bring myself to get rid of:

And I’m not sure that making access to the archives of those (in some cases deleted) blogs harder is entirely consistent with a vision of “Organizing the World’s Information”.

I wish I’d written about Proteus soon after playing it: it’s an amazing game, different in significant ways from almost every other game I’ve played, and while I don’t think I’d managed to come up with too many coherent thoughts articulating what I had to say, I suspect I had more at the time than I have now. We did talk about it as our VGHVI Symposium game this May, but it looks like that conversation is never going to be published as a podcast, so I can’t go back to listen to the discussion there to see what collective thoughts the game spawned, either.

Ah well; perhaps that feeling of loss is appropriate for the game’s wistful nature and presentation of irreversible change. At any rate, two fragments that stuck with me:

• The first time that the game really felt different to me was when I pressed the space bar and, rather than having that cause my character to jump, it caused my character to sit down. So: a game about pausing and drinking in your experiences; I spent a lot of time sitting down in the game.
• In my first playthrough, I was a bit rushed (we had to go to a concert), so I was trying to make it through the last season. But I couldn’t figure out what needed to happen to end the game, so I was running frantically around the island under the assumption that I just couldn’t see the spinning lights because of the cloud cover. And then I started floating away, my worries turned into wonder in one of the most transcendent sequences I’ve ever encountered in a game.

So: not much to say. Maybe, for a game like this, it’s more appropriate for me to just give pictures without commentary.

I really have fallen off of my blogging rhythm ever since getting back from vacation: the trip left me somewhat sleep-deprived, somewhat frazzled, and with a sequence of small “I finished a game” posts that I felt that I had to get out of the way. That combined to get me out of sync; I’ve still been taking notes about things that I want to write about, but I haven’t actually gotten around to writing about most of them, with the result that some of those ideas feel a little stale, and I’m not sure whether to write about them anyways to get them out of my head or to let them fester and continue to take up space with the hope that they will eventually decompose completely.

I was going to say that this post is a perfect example of that, because the events that triggered me to write about the topic took place during the first half of May, and enough has changed in that context that the details there really aren’t relevant. Except that, as I self-consciously typed the introductory paragraph, expecting that I would delete it (because, really, apologies at the start of blog posts do nothing for readers), I realized that that paragraph is actually a pretty good example of the topic in hand.

Namely: overwork and slack. Where, by “slack”, I don’t mean goofing off: I mean the idea popularized by Tom DeMarco’s book with that title of building in a bit of downtime / buffer into your work. This is, of course, hardly new to DeMarco: e.g. the lean folks are happy to present graphs showing response time going to pot as your utilization approaches 100%. Still, his book on the subject was important, and my memory is that it was rather good. (I should reread it!)

And a lack of slack was exactly where I found myself at work in early May. I’d gotten back from vacation, and immediately had to deal with a combination of digging out from two weeks of e-mail / change and dealing with end-of-quarter craziness. And the end-of-quarter craziness led directly into start-of-quarter planning overload, which unfortunately happened at the same time as my turn came up for two different periodic rotations.

The result was that I had no slack. In fact, it was worse than that: not only did I not have time to respond to all of the unusual requests coming in, I didn’t have time to do my basic normal “keep my head afloat from week to week” tasks, either. So the result was that I did a bare minimum of survival stuff and combined that with responding to a largely random subset of the unusual requests.

Which was not fun, though it didn’t suck too much, because the overload cleared out a week and a half later. What was amazing, though, was how it felt when it cleared out. All of a sudden, instead of feeling like my input queue was 125% of capacity, I felt like my input queue was 80% of capacity, and not only could I burn down my queue, I could spend more time thinking and working in a non-reactive manner. Which was great! I really felt like I was being a lot more effective than I had been when my input queue was overflowing. So I definitely want to try to hold onto that feeling, to work to maintain a buffer.

(Though to what extent I actually was more effective is subject to debate: since then, there’s been a mutual decision that my approach to planning and task assignment is a bad enough fit for my employer’s house style that I’m no longer a manager. I do think that’s mostly a fit issue, but it’s certainly true that I wasn’t working crisply enough in my preferred mode of operation to be able to push back against that fit difference.)

I’m still feeling the lack of slack at home, to some extent: a combination of recovering from vacation and having a little bit of work spill over into the rest of my life and then starting a major project. (Kitchen remodeling, whee.) And, of course, just having too much stuff on my lists. Fortunately, the way it plays out at home is rather different from the way it plays out at work: there’s not that much immovable stuff, so as long as food gets into the house and cooked, we’ll do okay. So this isn’t actively making me unhappy, it just means that I’m seeing a bit of a backlog building up in Things. Which is a useful feedback loop, I should think about what actions to take in response to it.

We went to Japan two and a half months ago, and I still haven’t gotten around to really blogging about the trip! Part of that is because I’ve been busy; most of that is because I’m not much of a travel blogger, I write about things I’ve been thinking about instead of places I’ve been to. But there is one place we went to on the trip that I wanted to blog about: the Ghibli Museum, specifically the museum’s Straw Hat Café. We saw a lot of wonderful buildings during the trip, but in its own way I think the cafe was the one that impressed me the most.

Going in, I was cautiously optimistic that I’d enjoy the cafe more than most museum cafes: as Hayao Miyazaki said in his hopes for the museum,

The cafe will be…
An important place for relaxation and enjoyment
A place that doesn’t underestimate the difficulties of running a museum cafe
A good cafe with a style all its own where running a cafe is taken seriously and done right

So the cafe was something that they were thinking about, at least, and I already had reason to believe that the museum was unusually thoughtfully designed. When we got up to the eating area, though, I was a little nonplussed: there was a fairly standard looking food stand with some tents that didn’t excite me. But after looking around a bit more, I saw the cafe, and I was a little surprised that there didn’t seem to be a line.

That was actually the first sign that the cafe was doing something unusual: observing it for a little while longer, there was a line, the line just didn’t look like the way I expected it to. There was a ring of seats near the entrance where people were waiting; one of the restaurant staff would emerge every few minutes, usher the next group in, and everybody would slide down a bit. People who were waiting were given menus to read, so they would be able to order quickly once they got inside; there were also a bunch of books lying around (mostly children’s books) for people to spend time with. So, basically, they transformed the probably inevitable wait time into a pleasant bit of relaxation; and from an organizational point of view, they used people’s time as efficiently as possible, and kept a steady trickle of people entering into the restaurant.

And after waiting for a bit, we made it into the restaurant ourselves; the first thing I was struck by was how light and how open it was. I didn’t take any pictures, unfortunately (they didn’t want you to take pictures inside the museum, and I didn’t think to ask if the cafe counted), so the only picture I have is this not very good scan from the museum catalog:

High ceilings, quite a bit of space between tables (they would have room to pack in a lot more people if they chose), wonderful light from above and from the sides. But varied ceilings, too: in that picture you can see them sloping down towards a boundary area in front of the kitchen, and the colorful tilework above the arch to the kitchen. There’s a fair amount of variation in the main seating area, too: from that picture it might look like the table sizes are relatively uniform, but in fact there are a range of tables over on the right to comfortably seat a range of parties.

And the attention to detail carried down to the dishes and tableware, too. The curry that I ordered was served on a plate with, if I’m remembering correctly, Totoros in a ring around the border counting off numbers. (Ah yes, it was the plate on the first and third pictures of this blog post.) But different orders had different plates; and everybody got a little flag in their meal with a Ghibli illustration on it. (They do Ghibli latte art, too.) More surprisingly to me, the different orders came with different specialized cutlery: it would have worked just as well if they’d set out a knife, fork, and spoon for everybody, but they instead took the occasion to signal to people that they’re paying attention to the specifics of your order.

Good food, too: significantly better than standard museum fare, quite possible one of the two or three best museum meals I’ve ever eaten. (The quality of the food on this trip was very high indeed: I don’t know that there was any one meal that was stupendous, but several were quite good, and there was only one restaurant out of the two weeks that we were there that we regretted having gone into, which is amazing.) Yes, they’re in a museum, but they’re taking care with their food, they treat both the food and the clientele with respect.

The very first goal that Miyazaki lists for the museum is that he wants “A museum that is interesting and which relaxes the soul”: based on the cafe, they absolutely succeeded in that. And the museum as a whole is wonderful: there are delightful architectural touches everywhere you turn.

I would love to see how Christopher Alexander would analyze the cafe. Just thumbing through A Pattern Language, I see several patterns which seem relevant: Intimacy Gradient, Indoor Sunlight, The Flow through Rooms, Sequence of Sitting Spaces, Light on Two Sides of Every Room, Eating Atmosphere (in spirit, I think, though details are different for restaurant eating), Ceiling Height Variety, The Shape of Indoor Space, Windows Overlooking Life, Half-Open Wall, Interior Windows, Open Shelves, Structure Follows Social Spaces. (And that’s just for the cafe; I can’t recall another public place I’ve been to that did so well with Secret Place or Child Caves.) But it’s not just the patterns: it’s this feeling of attention to detail, of how the elements of the space work together to harmonize with each other and to harmonize with the people and activities that give those elements reason to be.

The last book that I read on the trip was Alexander’s latest, The Battle for the Life and Beauty of the Earth: A Struggle Between Two World-Systems. If I’d read it at the start instead of the end, I would have been tempted to go visit the Eishin Campus; that would have taken a fair amount of time, though, so maybe the way things turned out is just as well. That latter book turns out to be really more about social organization than architecture, anyways: I will be thinking about it for a long time, I suspect, but I’m not at all ready to write about it here yet, I think I need a few more years of observation informed by that book first.

One of the books I read over vacation was Object Thinking, by David West. I should probably reread it, I certainly don’t claim to understand it well yet or know to what extent I trust the ideas therein, but it’s the sort of melange that appeals to me.

The book doesn’t shy away from grand philosophical statements, including framing competing schools of software development as a clash between formalism and hermeneutics. It portrays formalism as, in general, the winner in western culture (and in particular in western programming); it says that a true “object thinking” style of object-oriented development is on the hermeneutics side of that split, and that most of what is presented as object-oriented development isn’t really worthy of the name.

So this isn’t just a question about the specifics of how you use a programming language: even within software development, it’s broader than that. The book takes an explicitly XP-centric approach here in its approach towards software development, seeing XP as squarely within the hermeneutics camp, and of course XP’s approach is significantly broader than a question of what object-oriented programming means to you. Actually, until reading this book, I hadn’t necessarily thought of an approach to objects as being foregrounded within XP; one strength of the book is that it makes Metaphor, the most mysterious (to me, at least) XP practice, a little more concrete, but of course Simple Design and TDD very much push you in a particular design direction, and for that matter Whole Team (including the Onsite Customer) help support certain conceptions of objects.

Having said that, the book also supports XP for pragmatic reasons that have nothing to do with philosophical schools: one of its theses is that good software is likely to be produced most efficiently by good programmers, and one of the strengths of XP is that it gives you concrete techniques to practice that may improve your skill. So even if those techniques might in the abstract not be your favorite ones, the fact that they exist gives XP a leg up over a less focused approach towards improving your skill as a programmer. (Or improving your whole team’s skill together as programmers.)

The book doesn’t propose object thinking as an approach suitable for all programming. As it says,

Traditional approaches to software—and the formalist philosophy behind them—are quite possibly the best approach if you are working close to the machine—that is, you are working with device drivers or embedded software. Specific modules in business applications are appropriately designed with more formalism than most. One example is the module that calculates the balance of my bank account. A neural network, on the other hand, might be more hermeneutic and objectlike, in part because precision and accuracy are not expected of that kind of system.

It then presents an “applicability graph” opposing the “deterministic world” to the “natural-sociocultural world” and “comprehension” to “implementation”, saying that object thinking is suitable for the “comprehension / natural-sociocultural world” quadrant while the “computer science paradigm” is suitable for the “implementation / deterministic world” quadrant. I’m not entirely sure I buy those splits, but it does make me wonder: I’ve spent a fair amount of time over the last decade working on aspects of server software that, if not deterministic, are at least not in the “natural-sociocultural world”, and while they’re not pure implementation, they’re also not about comprehending the real world.

So, by the book’s light, it makes sense for a lot of that work not to be pure object thinking. Indeed, a lot of times the software that I work on depends ultimately more on how large volumes of data move around than anything else. And if we follow the book’s grander narrative, it might also make sense that the programming worlds that I work in have fairly strong antibodies to some of the other XP practices, e.g. Whole Team or Collective Code Ownership. (Though some of the other XP practices get a lot more traction, in whole or in part.)

Still, a lot of the code I work on could be improved by better object design. And, all philosophical issues aside, the book seems like it’s potentially pointing in an interestingly productive direction? I wish the examples were more detailed, though: he makes pretty grand claims about the reduction in implementation time and number of classes that come from correct object thinking, and I didn’t get more than a hint of that from the actual code in the book.

Player of Games has always been my favorite Iain Banks novel, but I’ve never been sure how much of that is because it’s good and how much of that is because it presses my buttons. I like the Culture quite a bit as a universe, and I’m also pretty obsessed with games, so it’s only natural that I would gravitate towards a book that combines the two.

I reread the book this week, and it meant quite a bit more to me this time. The way Gurgeh is portrayed as not fitting into the Culture’s norms for gender and sexual behavior; but he fits somewhat better into traditional American norms. The way he finds aspects of Azadian culture somewhat intriguing, and slips a bit more into their point of view as he immerses himself in their language and their game.

But his point of view, his Culture background comes in as well: the game is explicit in this, in its requirement (at least in later rounds) that players register their philosophical premises. The link between your philosophical premises and your style of play is implicitly present in a lot of games that I play: most obviously in Android: Netrunner, but a philosophical approach is also there in, say, 7 Wonders or Race for the Galaxy. Heck, it’s even there in Ascension, in your choice of biasing towards combat or towards improved production, and it’s certainly lurking within go’s abstractions.

I don’t think it’s an excuse that the first examples that come to mind are board games, not video games. Video games do frequently foreground choices, but those choices generally have more of a class-based feel, in how you construct yourself and your powers. And that, to me, has less of a “personal philosophy” feel than the board game approach of choosing how to interact with systems that are global to the game. So, indeed, maybe Netrunner isn’t a good choice after all, or rather, if it is a good choice, it’s not one because of the surface choices (Corp versus Runner, factions within those two), it’s because of the web of underlying mechanics that everybody has an option of how they want to interact with it. Still, class-based mechanics are by no means divorced from philosophical premises, and some video games (e.g. the Civilization series) lend themselves just as strongly to a premise-based approach/analysis as board games do.

So: who you are. Who your people are. How well you fit within your people. And a systems approach to all of this. The Culture naturally lends itself to a systems approach, with multiple layers: the Minds are playing at a whole other level, with humans as pawns, albeit pawns with agency. You can think of that as a depressing depersonalization (and the Minds aren’t the only intelligences in the book treating people as pawns, we see lots of that in the Azad matches as well); you can think of that as a personalization of a Deming-style belief that individual performance is largely governed by the system that they work within. Which, of course, we see reflected within Azad, and all the other games that are played. (Deming also makes the point that you should understand normal variation within a system before looking at variation of individual behavior, which fits in well with a focus on games that are, in part, games of chance.)

This all comes together in that last game of Azad, where Gurgeh has to confront the fact that his play is suffused with his Culture background. Indeed, his play always has been throughout the series; generally, that’s been to his advantage, because it’s introduced a (literally) foreign element that his opponents had a hard time grappling with, but now it’s hurting him, because gameplay that is informed by habitual Culture behavior isn’t up for confronting the forceful violence of the Azadian empire.

This isn’t a sign that the Culture is weak, however; it’s a sign that the Culture has to behave differently (while still sticking to its core beliefs) when confronted with certain enemies. And so, with a shift towards the “Culture militant”, Gurgeh wins, though not without drama.

So: some clashes that are on the surface about something else are instead really clashes of philosophies, clashes of who you are at the core. And if the antagonists’ philosophies are different enough, you may have to shift the way you express your philosophy. (Hopefully not in a way in which you lose who you really are, something which the book perhaps doesn’t explore as much as it might.)

And if you do that well enough, and if you have enough skill at playing the game, you and your philosophy will end up being dominant. (And skill is very important: as the book says, if Azad had evolved before the Culture, the roles of victor and victim would quite possibly have been reversed, this isn’t manifest destiny at work.)

But, even if you win, that battle can drain you, can leave you empty. The very end of the book was the part that I found most unexpectedly personally moving: the way Gurgeh started falling into depression after his victory, how he wanted to do nothing other than sleep, sleep for years, sleep without dreams.

The book ends with a note of joy and optimism. But Gurgeh has a lot of self-reconstruction to do, too. I would like to think that, after that reconstruction, he’ll emerge with a richer life.