Splitting apart

I managed to work in, this weekend, reading a young adult book by Veronica Roth, Divergent.

It takes place in a post-apocalyptic Chicago (and I was glad to have been in Chicago this summer, as it got me to recognize several places) where the city still (kind of) exists next to a big marsh that used to be Lake Michigan. In response to the destruction of society, the people have been split into five groups, each of which emphasizes one trait that that group believed contributed to the End. Those groups are Abnegation (selflessness), Amity (peacefulness), Candor (honesty), Dauntless (fearlessness), and Erudite (knowledge).

Everyone, on a specific date around their sixteenth birthday, has to take an aptitude test to determine which group they're most cut out for (and it seems that most people end up staying in the group in which they were born). Shortly after (I think it was a couple of days later), they choose which group they want to be in.

But some people are shown to be Divergent by the aptitude test (which is administered VR-like via a serum). Few people are quite sure what it means, but it's deadly to admit to (for reasons unclear until very late in the book).

The main character, Beatrice, is born to an Abnegation family, and is found to be Divergent. The test-giver didn't rat on her for being Divergent, and allows her to make her own choice after telling her that the test ruled out two factions.

She chooses Dauntless, and is thrown into a bit of a nightmare of initiation, as Dauntless (at least as it is "today") is very Darwinian (and, yes, cutthroat).  It takes a while to come out, but the Dauntless are the ones that protect the city from the outside, although we never really find out from what they're protecting the city.  They are the only group that has firearms, although they don't seem to have a lot of them.

So you can see where there are a lot of parallels with Hunger Games (more than I expected, when I started it); I think the main character is, also like in Hunger Games, becoming one of the driving forces of a revolution overturning society, although I don't have the second book to confirm that.  (Which reminds me, I still haven't seen the second Hunger Games movie.  Need to rectify that.)

The book definitely has some problems; learning to use a gun takes (much) more than a day or two.  And learning to fight, hand-to-hand, takes much longer yet.  Expecting a group that has never done either to be able to compete with people who've been doing it all their lives is absurd.

And let's face it, if there are only twenty-ish new people each year joining a group of ... well, it's unclear, but seems to at least be several hundred to a thousand... Well, that won't be enough.  Especially after that group is cut down to ten.

Plus, it takes quite a few support people to keep an army running.  It wouldn't take nearly as many as today's US Army (which has, IIRC over a hundred support per soldier), but it's considerably more than zero.

And that doesn't even touch on the number of Erudite.  To manufacture the things briefly mentioned takes very large numbers of people.  You need the people to do the designing, then the people to design the manufacturing, then the people to do the manufacture, then the people to assure the power for all of those stages, the training people to teach the engineers, etc.  It would take hundreds of times the number of Dauntless (although it's possible that there are far more Erudite than Dauntless, it's very unlikely).

Actually, that is one dynamic that isn't even touched on; the relative sizes of the various factions.

One that is discussed is that the government is almost entirely run by Abnegation; the other factions have only token representation.  This does make sense, in a way; you do get the least corrupt people running things.  And in a low-resource environment, that's certainly important.

Anyway, it's a very interesting book, and will probably make a pretty decent movie (it's coming in March), but I'm undecided about whether I'm going to read any more of it.  Maybe I'll get them from the library.  One thing strongly in its favor was that I raced through it in very little time, so it definitely kept my attention (much more difficult, these days, than it used to be).

Non-interchangable SLR?

I ran across this article about a new, lensless camera earlier today, and am a bit confused about its terminology.

To start with, I think what they're starting with is a pinhole camera controlled by an LCD screen. That is to say, they open a hole in the LCD lattice to be the pinhole. That would explain what they mention about infinite depth of field (a characteristic of pinhole cameras).

And I think there's something there about opening multiple pinholes at once to get a better image (an interesting idea, for sure).

But where complete confusion sets in, for me, is when they start talking about single pixels as the sensing element behind the LCD array. They must be using a weird definition of a pixel, because one pixel using the normal definition gives almost no information. Try reducing down any picture to height and width of 1 to see what I mean. I won't say there's no information there, but it's a damned small amount.

So I think what they're actually talking about is a single DSLR-like sensor, and possibly doubling that to two sensors. That would make some sense.

Other than that, I think it's a pinhole camera with an array of "pinholes". One limitation, if that's so, is that you'll be limited in exposure length, I believe, because two opposing polarizing elements don't stop all light. Don't believe me (and it's certainly counter to what I heard in physics class in high school)? Well, variable neutral density filters work by taking two one-stop polarizing elements, and rotating one of them. All of them go from two stops to eight (which is a nice range), but even at eight stops, you can still take usable pictures (heck, I've got a ten-stop filter that still allows usable pictures).

Anyway, it's an interesting idea, and I'm curious to see what they do with it.  Thinking about it, something like this would probably be perfect for a cell phone camera, where space is very constrained.  Hmm...

It also just occurred to me... Maybe what they were referring to, in using the word pixel, was a specific spot (any specific spot) on the LCD "screen"ing the sensor.  If so, the article was not well phrased.  It's also a bad term to use, because pixel is short for "picture element", and the screen blocking the sensor does not have any picture elements in it.

Pro-portions

I've been reading comments on one of the articles I linked before, and I'm seeing some fairly common thoughts.

One is dealing with the size (understandable, as users really don't care all that much about the size of desktop computers), asking whether a unit merely 1/2 the size of the old Mac Pro (vs 1/8th) might have been better.

And the simple answer is that it wouldn't be. It would add expansion capability in some form, true. But your options are to go over, under, or next to the existing "core". None of those options work well. Going over means you're going to have some components getting a huge amount of heat thrown on them. Going under means you have accessibility (for repair/replacement) issues (and you're moving the most valuable parts into a less-stable position). And going "behind" means you break the thermal core, because it stops being symmetrical. Suddenly, some of that heat is escaping towards the side, instead of up.

I think the whole core is understanding that the heat sink drove the entire design, and that's why it's so forward-thinking, because heat requirements will have a hard time going down, while there are a few ways to improve the airflow with this basic design.

Plus, thunderbolt will cover just about any eventuality that you might need (assuming the graphics cards can be upgraded, which is, it's true, not a given).

Pro Power

I've been meaning to write about Apple's new Mac Pro since it was announced at WWDC. I'm not sure why I haven't.

I've been thinking about getting a Mac Pro for a couple of years now, mostly because of memory issues (my iMac tops out at 16GB, and while the upgrade to 16GB was substantial, I still find myself wishing for more). A lot of that is driven by keeping lots of things open at once, for sure, but especially by photo editing.

So, anyway, because of that, I've been checking the tea leaves for what to expect, and the new model still caught me off-guard.

Here are two reviews worth perusing (especially for all the pictures):
Ars Technica
Apple Insider

Having a rack in my house, I wouldn't've minded the thing being rack-mountable. Not needed, but would have been welcome. So a bit of a loss there.

Also only has four RAM slots. This is actually a bigger problem, as it seems likely to top out (for the next year or two, anyway) at 64GB. Not shabby, of course, but was hoping it would be expandable to more. We'll see what happens here. One thing I definitely wouldn't want to see is an inability to add larger DIMMs as they become available (assuming no additional pins needed, of course; if they are, then of course that wouldn't be possible); I've been screwed by that a couple of times.

The last potential problem I see is that it wouldn't be happy being stuck under a desk. Not only because the striking design would be wasted for lack of visibility, but also because it'll be pumping out a lot of hot air, and you need that air to get far away.

Beyond that, though, I think it's brilliant. It was, of course, entirely predictable, doing away with the optical drive, although I didn't think it'd be gone quite this soon. And I like the forward-thinking nature of doing all the expansion via thunderbolt (although there are definitely cost issues there). But I think the cooling design will work for a very long time, and they even claim that it'll share the cooling when one or more of the main processors is working at less than full load. I'm not sure about the latter part, but it looks very promising anyway.

This is definitely a forward-thinking machine, and I expect that other companies will put out copies pretty soon. Although I'm thinking Apple will have huge advantages of scale with it (especially if they come out with cheaper versions than the probably-$10k model shown at WWDC); it should be able to go from a fairly small processor up to the heavy workhorses in the model shown.

The one thing I worry about is if they'll have a model in my price range; I don't need two stupidly-powerful GPUs for my use and wouldn't pay for them. But, this architecture, if done right, could still put something into the sweet spot for purchase for me.

And the one bit of complete speculation. They made such a big deal about being able to drive three 4k displays with it. Will they announce an affordable 4k monitor upon release? That would REALLY get me stoked; I'd love to get one of those for photo editing.  In fact, that would be a bigger deal for me than anything they could do within the computer.  There are even intimations (screen backgrounds within 10.9 Mavericks) of a double-resolution 27" display; that would be even better.

Will the next big (still) camera have a mirror?

I was interested to read Trey Ratcliff's essay about the future of cameras.

While I certainly agree with him in principle, I'm not so sure about timing. That is, it's certainly coming, for many good and valid reasons, but he seems to be saying that it's right around the corner, and I think he's a bit early, there.

I'm fairly sure that the next generation (iteration, if you prefer) of top-of-the-line cameras will be SLRs, and I wouldn't be terribly surprised if the generation after that is, as well. But I wouldn't be at all surprised if the generation after that is a mirrorless camera. I would expect it to be full frame (so, bigger lenses and bodies than what we're currently seeing), with a much fuller selection of lenses than any currently available mirrorless system.

I'm kind of looking forward to it, actually (even though I have a new SLR on order). One good thing is that lenses should be quite a bit cheaper. One nice bit about never having a person looking directly through the lens, is that even the viewfinder image (whether delivered to the back of the camera or to an eye-hole through the back of the camera) can be digitally corrected for various lens imperfections. That should lead to fewer tradeoffs needing to be made during engineering, and, therefore, cheaper lenses. We'll see, but I think that's one factor.