Log in

28 January 2016 @ 08:52 pm
I introduced Newton's Method to Amber tonight, since she's been very interested in the square root of two recently, and Newton's Method is a pretty good way to approximate it.  Also, I thought she might be ready for the process.

Her favorite part was where I tried bad initial guesses, and a guess of zero gave an unhelpful infinite correction.  As Amber (loudly) says, "No dividing by zero!  We don't like dividing by zero!"

We'll see how much she remembers tomorrow, and if she can generalize to estimating other square roots.
15 December 2015 @ 09:46 pm
Today I got mistaken for a senior citizen, and also got carded.

Either age is really hard, or people aren't paying attention.

Given that I got carded for buying soda, I'm going for the not-paying-attention explanation.
The Burglary, by Betty Medsger

(A gift from my parents)  I'm not entirely sure what to make of this book.  It covers a very important event in US history, more or less from primary sources.  The key event is the burglary of the Media FBI office, which revealed just how much illegal activity was going on in J. Edgar Hoover's FBI.  That said, the book has some weaknesses.  The activist-burglars were never caught (and the statute of limitations is long-since expired, making them more willing to talk), but the narrative reports many, many times on how they worried about being arrested, hurting the flow of storytelling.  The book is also utterly convinced of its political positions, spilling over into discussions of topics that aren't necessarirly connected to the main thread of the book.  For example, the discussion of nuclear weapons (in chapter 23) says, "Countless thousands of people in both cities were permanently disabled, as were future generateions of children whose parents carried genes damanged by radiation."  This is an odd point to make:  There are plenty of reasons to object to nuclear weapons (I'd start with "They tend to kill lots of civilians."), but passing on mutated genes appears to not be a large effect:  It's been looked for and not found at a statistically significant level, as people tend to either die from the effects before producing further children, or be largely unaffected at a genetic level.  The section on cryptography (chapter 27), seems off; it uses cryptographic terminology oddly, and manages to misspell Bruce Schneier's name as "Schneider".  Essentially, I'm left wondering how careful the rest of the book is, given how many things are mistaken, or lead me to ask "How could the auther know that?".

But that doesn't really diminish the important part of the book:  The core lesson is that secret govenment agencies tend to engage in activities that are illegal, counterproductive, and unknown to the people who they should be reporting to (congress and/or the public).  Citizens need to be aware of this and vigilant.

Conditionally recommended:  It covers an important pieces of history, and I'm not sure there's a better book on it.

The Martian, by Andy Weir

(A gift from Bill Mason)  This is an example of that tiny and difficult genre, the science fiction book that doesn't break any laws of physics.  The premise is fairly simple:  Astronaut Mark Watney is presumed dead in an emergency and left behind in a Mars mission.  He then has to figure out how to survive long enough to be rescued, when the environment and supplies at hand aren't well-matched to that task.  It's a gripping story, and one that I'm looking forward to seeing in movie form as well.  One side effect of the effort at realism is that there are lots of details that might be wrong:  I remember thinking that the author probably had the density of liquid oxygen wrong.  That's more a testament to how many details are in the book, and how many he got right, though.

Highly recommended.
10 September 2014 @ 09:31 pm
It's finally time to retire my old laptop, the one that we've been using as a home server for a while. The screen had stopped working, but that didn't matter so much for server use. Last night, however, something else failed and it won't do networking anymore.

So ... I got about 15 years out of it, which makes it a relic of a different era.

(Admittedly, a different computer now thinks that it's Cricketsong, but I know the difference.)
I'm trying to sort out how the various pieces fit together: Early in the crisis, there were a lot of regulatory actions that, taken as a whole amounted to "giving money to banks". For example, the Federal Reserve started paying interest on reserves, which is a bank-only bonus; similarly, loaning banks money so that they can buy treasury bonds is a small step away from paying banks a commission on monetizing debt.

Now we've reached the matching phase of "taking money from banks", again as regulatory actions. The net effect of the two taken together is to allocate "new" money to places (like "borrower relief") that would look dubious if done as a direct line-item in the federal budget. The timing of the two phases tends to keep the banks just-barely-solvent, which would be better accomplished in some other way. It's not at all clear to me where the deadweight loss from this falls, nor even a good way to estimate how much productivity is lost.

The other troubling part is how much this looks like arbitrary exercise of regulatory power. Most of the lending covered by the settlement wasn't at Bank of America: It was primarily Countrywide, and to a lesser extent Merrill Lynch that made the problematic loans. Bank of America bought both of them to prevent messy bankruptcies (Regulators usually prefer to do that as opposed to a messy bankruptcy as in Lehman Brothers). In a real sense, those purchases were done as a favor to the banking regulators; but had Bank of America's management known that they'd come with fines of 3 years' profits, they'd likely have balked. Then again, not acting according to regulators' wishes has it's costs too, as Standard & Poor's is demonstrating.

The whole thing seems messy and suboptimal.
I'm trying to figure out the motivation behind the SEC's experiment:


(Via a reference in Kid Dynamite's blog post)

In short, they're picking a few thinly-traded stocks, and requiring that prices be a multiple of 5 cents, instead of the now-traditional 1-cent (usually). The first-order effect of this is pretty clear: It transfers money from investors to market-makers. In the past the SEC has worked in the opposite direction, both with decimalization and going after dealers for "colluding to keep bid-ask spreads wise". In particular, NASDAQ had a phase where lots of bid and ask prices were even; the SEC viewed this as an attempt to mistreat investors.

It's possible that this is an attempt to favor traditional market-makers over the high-frequency traders: With prices changing by 5 cents at a time, there's less opportunity for a high-frequency trader to pick up a trade. (Or viewed the other way, for a high-frequency trader to give an investor a better deal by a penny or two.) They may be thinking about the smaller bid and ask sizes that naturally go with a smaller spread, but if that's their objective, it seems like there would be better ways of accomplishing that.

For example, instead of the current Regulation NMS (National Market System -- it details how the many exchanges interact, and those rules do much to enable high-frequency traders), the SEC could experiment with slower trades: Have exchanges match trades every second. As an individual investor, I'm rarely in that much of a hurry, and I'd rather get a deal that's better by a dollar than a deal that's 300 milliseconds faster.
Tags: ,
08 March 2014 @ 11:22 am
One of the interesting bits of the crisis in Ukraine is that in 1994, Ukraine gave up its nuclear weapons in return for a security guarantee. Russia, the US and the UK supposedly guarantee Ukraine's integrity; the value of this guarantee looks dubious at present.

The sets an unfortunate precedent for any future leaders trying to decide whether or not to maintain a nuclear arsenal. Nuclear powers, essentially, don't suffer territorial incursions, and non-nuclear powers do...
01 December 2013 @ 10:18 pm
I find it ridiculous that the FDA objects to 23andMe's test kits: Getting your DNA tested really can't hurt you. The FDA's theory is apparently that someone might find out that they have a genetic risk for something, and do something dangerous as a result. Still, most of the medically dangerous things that you might do as a result (unnecessary surgery?) require a doctor's assistance, and that's really the proper place to inject sanity as required.

Apparently there's a whitehouse.gov petition to ask the FDA to reconsider their ban:
06 November 2013 @ 12:41 pm
I got two telephone polls yesterday, which turned out to be very different. The first one seemed to be some sort of long-term Republican strategy thing, and asked for, among other things, "in your own words" descriptions of stuff that matters to me (judging by the sound, she typed in my answers).

By contrast, the second poll was following a rather inflexible script, and wasn't satisfied with my declining to say how I voted in yesterday's election. I eventually hung up on her three times; Beth told her to stop calling the fourth time.

I wonder how much my responses (or lack thereof) wind up mattering.
19 December 2012 @ 10:05 pm
Introductory Calculus For Infants, by Omi M. Inouye

I'm not really sure how to describe this book: If you expect a calculus book, it has less calculus than you might expect. If you're expecting a kid's book, it has more calculus than you'd expect.

It follows the letter x (which the other letters didn't like) after f (who is fabulous and fun) explains how as a team they can do lots of fun things: one for each letter of the alphabet, all at least plausibly calculus-related. There's a happy ending: The other letters see how x can be a good friend.

I look forward to reading it to Peter and Amber, though I expect that I will have a lot of explaining to do. (Despite the title, I don't think I would try reading this book to a very young child; an appreciation for the alphabet is necessary.)

Conditionally recommended: It's cute and funny, but would presumably not shine as well if you're not in the mood for a kid's book, or don't have fond memories of calculus class.
Tags: ,
31 July 2012 @ 09:55 pm
The DVR that Beth got along with cable service failed recently, wiping out her recording of the early parts of the Olympics. Apparently it was a HD failure, and while the tech they sent out to troubleshoot was nice enough, their process is kind of broken.

For reasons that are unclear to me, the DVR tests its network configuration before it tests the disk. It didn't like the IP range it got DHCP'd into; for some reason the cable box had detected that I had most of my computers behind a firewall, and put us somewhere other than the 192.168.0.x it was expecting.

So they replace the cable box upstairs. After that, the downstairs box was able to diagnose that it was broken, and the tech replaced that too.

The script doesn't call for connecting my computers back to the new box, so the situation when AT&T has "fixed" the problem:

Beth's recordings are gone.
All of the devices that used the cable box as a wireless access point need reconfiguration.
The firewall isn't connected to the cable box.
Even after I reconnect the firewall, I still need to reconfigure stuff, as the network is different.

At least I was able to put stuff back in working order eventually.
Tags: , ,
27 June 2012 @ 11:25 pm
Somehow today's Google Glass demo resonates for me. Probably some of the trappings (a Zeppelin!), and the initial uncertainty as to whether they were really planning to jump out.

Not sure what's the best reference. I (re-)found the video through TechCrunch, having seen it over live-stream. (I also linked to the video over on my Google+ account).
24 May 2012 @ 10:18 pm
Apparently, if I need to learn about swim safety, it is also vital that I watch animations about sandwiches.
10 May 2012 @ 09:21 pm
Today was take your bike to work day, which made for some nice variety: There's some interesting scenery quite close to 85 that just isn't visible on my normal commute.

Not as significant as the last two weeks though: I don't think the bicycle got nearly as much out of take-your-bike-to-work-day as the children got out of their take-your-child-to-work-day instances.
17 April 2012 @ 10:23 pm
Tax day seems like a good day to muse on the meaning of the US having the world's highest marginal corporate tax rate. Given the actual average rates paid, this means we have world-class-abuseable deductions, too. Saner tax policy would be a good thing.
21 March 2012 @ 10:44 pm
T-Mobile's web site needs me to log in to do various things (including reporting problems). That's fine; I even know my cellphone number which is to be used as an account name.

After I log in, it wants me to set a new password, and supply some information like my name and email address. The tricky part? The fields they want me to fill in are simultaneously "mandatory" (can't submit with them blank) and "disabled" (web browser won't let me type in them).
13 March 2012 @ 11:11 pm
Originally posted by beth_leonard at End Daylight Savings Time Changes
We the people petition the Obama Administration to:
End Daylight Savings Time Changes.

Click here to sign the Whitehouse.gov petition. It needs 150 signatures to become public, and 25,000 signatures by April 13th, 2012 to get an official response. You will need to create an account if you do not already have one.

If you don't like changing the clocks twice a year, please pass it on!

Today's video-conference was confusingly scheduled, because not all of the participants have matching time zone rules. It's annoying enough having to adjust my clocks, but having to keep track of the time changes in other zones (Arizona: no change. Europe: change on a different day) adds an extra level of silliness.
08 March 2012 @ 07:00 pm
For somewhat obscure reasons, my new mobile phone doesn't know what its phone number is.

When the carrier called to welcome me to the service today, I asked them what my phone number was. Apparently for privacy reasons they can't tell me.

At some point I'll just call someone with caller-ID so I can find out.
28 January 2012 @ 04:36 pm
The Futures: The Rise of the Speculator and the Origins of the World's Biggest Markets, by Emily Lambert

This is at its core, a history of the Chicago futures markets. Other cities play into it as well, of course: The Chicago markets in stock derivatives interact strongly with the New York markets. Much of the history was new to me; I had a pretty good idea what sorts of things traded on the Chicago exchanges, but not how that came to be. They grew out of various regional agricultural exchanges. There's always a mismatch between when food is produced and when people would like to consume it, and futures markets provide (through speculators) a way to match up buyers and sellers at different time. Often this works well, other times not so well: The ban on trading onion futures is the result of one such incident. Originally there were many such markets (even two in Chicago), but as transportation improves, there's a tendency for centraliation. That's because buyers and sellers tend to prefer the largest market they can reach.

The book covers much of that history, and the rather amazing cast of colorful characters involved. Some of the interactions are memorable, too: In the aftermath of the 1987 crash, one of the New York bankers apparently asked, "Who is the Chicago Mercantile Exchange, and why do they owe us a billion dollars?". The answer, of course, being that they were the host of the stock futures market, and in the spectacular 1987 crash, much money was involved in settlement. The sometimes unexpected feedback betwen markets matters: In the 1987 crash, arbitrage between the Chicago and New York markets added to the volatility, and similar effects were significant in the more recent financial crisis. There's a lot of stuff to learn about these markets, and this book is a good source on the more human side of the markets.

Highly recommended.
28 January 2012 @ 12:51 pm
I recently read two books by Emanuel Derman, both of which have a lot to do with both physics and finance. That presumably has a lot to do with his history: He was a physicist in academia before winding up on Wall Street, eventually becoming head of the quantitative risk strategies group at Goldman Sachs. He subsequently returned to academia.

My Life as a Quant: Reflections on Physics and Finance, by Emanuel Derman

This is the more (personal) history-focused of the pair, while still having a pretty good dose of both physics and finance. In particular, his description of how VaR (value at risk) models are applied and misapplied enlightening. One of the things that struck me is how often (at least in the narrative) his life was unsatisfying. The decription of how he had to (for work reasons) spend a year away from his young son was heartbreaking.

Highly Recommended.

Models. Behaving. Badly. also by Emanuel Derman

This book is more focused on the nature of theories and models, though is has a sprinkling of his personal history as well. Probably the main point he tries to get across is that the constructs of physics are fundamentally different from those in finance: They both use math, but the theories of physics are very accurate. As he says, for all intents and purposes, Maxwell's laws are the electromagnetic field, not an approximation to them. Even the best models in finance, in contrast, are at best approximations, and often dangerous ones at that. He covers a wide range of topics, including excursions into philosophy and religion.

Conditionally recommended: It's in many ways similar to the other book, and My Life as a Quant is probably the better one to read first.
21 January 2012 @ 11:55 pm
Feynman Lectures on Computation, by Richard P. Feynman (Edited by Tony Hey and Robin W. Allen. Sadly, it's a posthumous book.)

In a similar spirit to the excellent The Feynman Lectures on Physics, this is a conversion of some of Feynman's lectures to book form. It's a pretty good introduction to a lot of aspects of computers, from obvious topics like breaking up tasks into computer-instruction-sized pieces, to less obvious topics: Reversible computing, entropy, and quantum computing.

I found almost all of it easy to follow: Feynman's presentations are always very clear. The exception was chapter six, "Quantum Mechanical Computers". Much of this was phrased in the language of quantum field theory (bra-ket notation, etc.), which is something I wish I understood better. But after a little review and re-reading that chapter it made sense. Essentially, he's describing a design for a computer where the individual steps are compatible with small-scale quantum mechanics, which is necessarily reversible and so on. Unlike subsequent discussions of quantum computing, his design doesn't depend on using quantum superposition or attempt to do non-classical computing. It's more a thought experiment as to just how small (and low power) a computer can be.

The last chapter on physical aspects of computers (VLSI and so on) provides a fascinating glimpse into circuit design before I got into it: Back then, it wasn't obvious that CMOS was going to be the dominant way to design chips. Still, most of the chapter maps very well to what I did when I was a chip designer.

Highly recommended.
One of the more noteworthy provisions of the proposed legislation is that it calls for blocking entire sites to deal with infringing content. This is unfortunate in terms of collateral damage: A site like LiveJournal could be blocked over content that I (as a user) have no control over, as a consequence of going after "bad guys". (A similar mechanic is active in the recent takedown of Megaupload. I'm sure there was infringing content, but there's a lot of inconvenienced non-infringing users as well.) Sites with a lot of user-generated content would be in trouble, especially if it's uneconomic to have a site administrator police everything. Wikipedia would essentially have to shut down, as a most visible example of the problem.

From a economic viewpoint, though, this is a pretty good example of the tragedy of the anti-commons. Unlike the similarly-named case with no property rights, a case of overlapping property rights can lead to under-use of a resource. SOPA would extend the property rights of intellectual-property holders to include the privilege of blocking content that is merely co-located with infinging content. Given the number of potential rightsholders to negotiate with, this blocks off a lot of useful activity. It also does so at a less-direct level: Where is the line drawn between infringing and non-infringing content? With lots of people to potentially offend, a lot of now-legal parody and commentary becomes legally risky. Yes, it's legal, but you have to argue that after they shut down your site ... so why bother in the first place? (This chilling effect is probably a significant motive behind the legislation in the first place.)

As a side issue, adding new legal processes (paths outside the judiciary, DNS blocking requirements for ISPs, etc.) deserves careful scrutiny: It is claiming, more or less, "Due process is too much work." On the contrary, due process is vitally important.

All of this is orthogonal to the question of exactly how much protection intellectual property should have; Even if it needs still more protection (I don't think it does), the proposed mechanisms are a poor way to accomplish that.
20 January 2012 @ 01:43 pm
It looks like I'll be starting work at Google shortly. This is an exciting change, and I'm really looking forward to it. (Totally rearranging our household schedule, not so much.)

I'll be in SRE, doing some mysterious combination of software and system administration.
Tags: ,
30 December 2011 @ 10:21 pm
I was doing something related to Fermat primes, and noticed that my TI-84's interpretation of 2^2^5+1 isn't 4294967297 as I'd expect (famously divisible by 641), but rather 1025: It's interpreting it as ((2^2)^5)+1. How'd they get the order of operations wrong there?
14 December 2011 @ 03:49 pm
I was recently considering how unbacked currencies can lose value: In the old days of the gold standard, the fact that a dollar was theoretically convertible into gold kept the value of a dollar more-or-less stable. (The occasional suspension of convertibility and mining improvements notwithstanding.)

With a fiat currency, there's nothing in particular to keep the value from slowly (or quickly!) inflating away. But there may also be nothing to restrain volatility. Certainly after the Bretton Woods system collapsed the currency movements have been much wilder. The question that I don't have an answer for is to what extent is the mere fact that the currencies aren't backed responsible. Certainly there are other possible reasons: Transactions in general are faster and easier than they used to be, etc.
Tags: ,
07 December 2011 @ 10:57 am
On the way back from taking the children to school, I saw a sports car with a license plate frame reading, "Paid for with seized drug money". That's probably true: The car had a tax-exempt license plate.

This seems wrong to me on a couple of levels. To start with, I'm uneasy about civil-forfeiture in general. It is a necessary tool in contexts like the historical pirate ship cases: The owner of the pirate ship can't be brought to trial, but taking the ship to stop the piracy is justifiable.

So historically, civil forfeiture applied in cases where a trial was impossible. That has been lost, and it is now used in cases where bringing a trial is inconvenient. It's a lot easier to take a pile of cash from an unsympathetic individual (and place the burden of proof on them that it's legally gained) than it is to prove wrongdoing.

But the sports car bothers me in another way: Is this really a good use of police funds? Money is fungible: If seized assets are being used to fund law enforcement, they should be used more or less like any other funding (ignoring for the moment perverse incentives). If you can't justify spending taxpayer money on something, spending seized money on it doesn't make sense either.

Obviously I don't have the whole story: I don't know whose sports car it was, or why they had it. But I still find it unsettling.
Tags: , , ,
01 December 2011 @ 01:47 pm
The Proud Tower, by Barbara W. Tuchman

This is an excellent history book, much like her other output. (I'm fond of The Guns of August, for example.) As the subtitle ("A Portrait of the World Before The War 1890-1914") implies, it's about the now-distant pre-WWI world. There's a lot of selection involved, as she points out in the foreword. The focus is limited to western Europe and English-speaking North America, but even so she says she could have written such a book three times without repeating.

The result still feels like a broad view of a world that is in many ways quite different (Aristocracy doesn't matter much anymore), but in many ways all-too-familiar: The discussion of the anti-war movement and negotiations seems very much like the climate-change negotiations of more recent history.

I found almost all of the book compelling reading, but I did get bogged down in the section about musical composers. Probably more significantly, after reading the book I have a much better idea of what the Anarchists were all about (to pick a single example out of several). The time I spent reading was well-spent.

Highly recommended
21 November 2011 @ 04:02 pm
The Meaning Of It All, by Richard P. Feynman

This is essentially a transcript of three talks that Feynman gave, on the topic of science, scientific reasoning, and unscientific reasoning.

As expected, he's clear and insightful, discussing how science works, the limits of how it can be applied, and when other approaches are appropriate. Like much of his work intended for a general audience, he avoids scientific jargon. This can be slightly frustrating: He has a wonderful discussion on hypotheses, testing hypotheses, and the problem (temptation?) of using the same data that inspired a hypothesis as a test. Having recently reviewed the appropriate parts of probability, I know exactly what math he's talking about; but the general reader won't. (They may not need it either, of course.)

Tags: ,
17 November 2011 @ 04:20 pm
Recently the stand for one of our monitors broke, and even more recently another monitor had the light for the screen fail.

Fortunately, they're the same model, so I just swapped pieces, and now have only one broken monitor instead of two.

(I realize that the monitor mount is standard, and the fact that the two monitors were nominally identical isn't actually necessary.)
President Obama's announcement that he'd make the final decision on the Keystone XL pipeline has had a presumably unintended effect. Getting approval on an alternate pipeline path would take at least another year. The Canadians want to sell their oil; if the US path to selling the oil is tied up by politics, then they'll actively seek to sell the oil to Asian markets instead (via a pipeline to Canada's west coast.)

The odd thing is that this consequence is worse than many of the other possibilities. If Obama had kept quiet, the pipeline would presumably gone forwards, with the result of a lot of construction (20,000 jobs, apparently), and improving our energy security. If he'd announced support, we'd have the same result, but he'd be able to claim some credit. If he'd announced that he'd kill the project, he'd gain favor with green constituents, (while of course giving up the economic benefits). But the in-between choice effectively kills the project, while annoying the Canadian government, and signaling governmental paralysis. There had to have been a better way to handle this.
12 November 2011 @ 04:05 pm
Rule 34, by Charles Stross

This is a sequel to Halting State, and a pretty good follow-on. It has many of the same properties as the first novel: It's written in second-person (a nod to the presentation for a lot of computer games), follows multiple characters, and involves a lot of near-future police work. Unlike Halting State, the theme isn't computer games, but rather spammers and the sorts of products they push. Other parts have changed somewhat as well: The 2nd-person writing somehow feels less intrusive, and the more subtle aspects of the story matter more: There's a lot going on that doesn't show up directly in the text of the story (reminiscent of his excellent Palimpsest), so the ending requires a fair amount of thought to understand.

Conditionally recommended. If you want more like Halting State, it's a worthy continuation; but not a really good place to start (especially given the subject matter).
06 November 2011 @ 09:17 pm
We went out with a bunch of the childrens' friends (and associated parents) to see Puss in Boots.

It's sort of a mashup of a bunch of faerie tales, with substantial reinterpretation. It makes about as much sense as the original source material, but the visuals are pretty cool, and the story mostly holds together.
Conditionally Recommended: It's about what it sounds like.
I was thinking about MF Global's odd (and unfortunate) repo-to-maturity trade, and concluded that one of the developments is that government bonds are effectively money.

In the not-so-distant past, large banks loaned each other money as loans, but the current trend is to post high-grade bonds as collateral. More formally, it's a sale with an associated repurchase, which mostly works out as a loan, but is much simpler in bankruptcy.

So, in practice, financial entities (banks and so on) can get money for these bonds very easily. So easily, that they effectively are money. (Except when they suddenly aren't, as in Greek bonds.)

In such an environment, a central bank can't effectively increase the money supply by buying government bonds, because as a first approximation, they're just trading money for money. The inflationary effects now mostly happen when the money is borrowed, not when the central bankers buy the debt.
07 October 2011 @ 03:25 pm
In case anyone else is interested in a step-by-step guide to setting up an Android development environment on a Debian Linux system, here's (minus some typos that I retried), what I did:

First, system-administrator steps, as root:

Install the tools I was missing, and make them the defaults. Having the wrong version of java or keytool make for mysterious errors.
apt-get install sun-java6-jdk sun-java6-jre eclipse ant-doc
update-alternatives --config javac
update-alternatives --config java
update-alternatives --config keytool

Install the Android development stuff, from http://developer.android.com/sdk/index.html
cd /usr/local/lib
tar xvfz ~jleonard/android-sdk_r13-linux_x86.tgz

Then set the permissions to something closer to what I'm used to:
find android-sdk-linux_x86/ -nouser -print0 | xargs -0 chown root.staff
find android-sdk-linux_x86/ -print0 | xargs -0 chmod go-w

Then have it update to get the various SDK targets:
cd tools
./android list sdk
./android update sdk

Then a bunch of user steps:

Add relevant stuff to my to PATH; I appended

Make a directory for Android projects
cd src
mkdir android
cd android

I then scp'd over my keystore from my other computer; but if this were my first install, it'd be
keytool -genkey -v -keystore release.keystore -alias signkey -keyalg RSA -keysize 2048 -validity 20000

(This also involves picking passwords, and new key ID stuff)

Then, time to generate a project
android create project --target android-13 --path example --package com.slimy.example --activity example
cd example

In AndroidManifset.xml, add an SDK version tag:
 <uses-sdk android:minSdkVersion="13" />

Also, build.properties needs to know about the signing key, append

Building the project is a simple
ant release

To install and test it, connect an Android device, and enable USB debugging
adb install bin/example-release.apk

It's also quite possible to test with a software simulator, but because they're emulating a different instruction set (that's then running bytecodes), it's significantly slower than testing with real hardware.

For a debug version, doing
ant install
will build and install with a debug key.
ant uninstall
can also be useful.

Steps not involved in the process: Writing any code, paying any money. Doing an app that's more than a "Hello, World" obviously involves editing the relevant .java classes, and setting up to distribute through the Android Marketplace involves a $25 registration, which I haven't yet done.
05 October 2011 @ 08:26 pm
I've just started trying to build apps for Android, and one of the steps is to sign the package.

However, for the keys I've generated myself, I kept getting an error along the lines of:
"keystore load: Invalid keystore format"

The problem turns out to be that there are two different versions of keytool on my Debian system, one from gcj-4.4-jre-headless, and one from sun-java6-bin. Despite substantial similarities, they do not produce compatible output.

Frustrating, but solved now.
04 October 2011 @ 09:18 pm
I rebooted our home server today, after only 600-some days of uptime. It was probably due anyway, really.
Lands' End is having a sale on swimwear, and the structure of the sale is kind of interesting. Nearly everything has a big discount (65% off or so), and they're pretty clearly trying to sell off all of their inventory. (Apparently their southern-hemisphere seasonal demand isn't enough to justify year-round inventory.)

The interesting part is how they're pruning items from the display: As inventory on an item goes away, it stops showing up in searches. Along the same lines, the sizes and colors available don't list all the possibilities that (presumably) existed pre-sale. The end result (presumably) is that they'll be able to sell off nearly all of their inventory, without making a lot of user-visible changes. (It's a bit odd that a search for bikini tops yields only one result, but this is presumably nearing the end of the sale.

Not quite as big an adaptation as the demand-management computer manufacturers do (free upgrades from components that they're short of), but still pretty cool.
28 August 2011 @ 02:29 pm
I wanted an Android device, in hopes of eventually doing some software development. Since we were also preparing for a long car trip, it seemed like a good time to get an Android tablet: Tablets for games and movies are very effective at keeping the children entertained on long trips.

After a small amount of research, I concluded that there are three good choices for tablets at present: The iPad 2, the Samsung Galaxy Tab, and the Asus Eee Pad Transformer. Since I already have an iPad (albeit 1st generation), and I wanted an Android device, I looked at the differences between the Galaxy and the Transformer. The Transformer is cheaper in the keyboardless form, but its expansion keyboard is significantly more expensive than the Galaxy's keyboard (keyboard+tablet is about the same). After typing on each of them a little bit, I decided that I liked the Transformer's keyboard better. Neither keyboard is really what I want in a keyboard, due to the severe size constraints. If a keyboard has to match a 10" tablet, it's going to be cramped. The Transformer's keyboard does have some other neat features, though: It has two USB ports, a SD card slot, and best of all, extra battery. It attaches/detaches fairly easily, and closes with the tablet as a combination sleep signal/case. Essentially, I picked the Transformer over the Galaxy because I liked how the optional keyboard worked better.

Another feature I'm happy with is the microSD port on the tablet itself. With my iPad, I can't fit all the movies I'd want on it, which is tough because the children can't reliably predict what they'll want to watch. With the Transformer, I can put movies on a microSD card (or cards), and the 16 GB of built-in storage (I didn't go for the 32 GB model) is plenty.

The Android environment has taken a little getting used to, mostly because my tablet experience has been on the iPad. The design philosophy is just different: It has a visible filesystem, and apps are often structured as a stack of screens. The app store model is significantly different as well: Apple is pretty picky about what goes into the app store, tending to enforce a minimum quality level and exclude some kinds of content (is there a Scheme interpreter there?). It also delays patches, which can be annoying. The Android market, in constrast, seems to be much more of a free-for-all. I'll have a better feel for that after a little more exposure.

The system does seem a little bit quirky. For example, the various apps deal differently with the keyboard. Some of them lock in landscape orientation when the keyboard's attached, but one (that I've found so far) remains locked in portrait mode, which is kind of strange. Still, there's a lot to like; I'm not sure whether I'll wind up using my iPad or my Transformer more.

Beth and I took the children to see Legoland and California Adventure earlier this week. It seemed like a good time: Many schools (but not ours) have started, so the parks shouldn't be too crowded. Also, it was Beth's birthday.

Legoland was fun, having pretty much what you'd expect. There's something of a paradox, though: Fundamentally, the most fun Lego-related activity is to build stuff, but we can do that at home. If Legoland was closer, I'd probably get a membership and take the kids occasionally (they did have a good time!), but I don't think we'll be making the trip very often.

California Adventure was sort of strange: It's very much a Disney theme park, only without all the familiar attractions. They were fun, but I found myself wishing we could do some of the more familiar ones. Our family in particular kept running into the height requirements: The big roller-coaster had a 48" minimum, which disappointed Peter. He was (eventually) OK with riding the other coaster, though (There are only 2 in the park). Amber at 40", plus-or-minus, was disappointed several times. She enjoyed the flying ride enough to want to retry, only to be disappointed when they ruled her not-tall-enough the second time around. Similarly, there are some 42"-minimum areas in the wilderness-themed play area, which would have otherwise been ideal for Amber. (I'm mystified why the horizontal climbing wall has that high of a requirement.)

At the end, we watched the "World of Color" show, which was spectacular. It's by far the most impressive thing I've seen done with fountains, even without the other elements. It's strongly reminiscent of Fantasmic, but with more space, and a more suitable park layout. There wasn't particularly much story to it, but we were watching it for the effects anyway.

I'm not sure how often we'll go back: Park-hopper tickets are noticeably more expensive, and if I have to pick one park, it'll be Disneyland. If somehow we did a multi-day visit, park-hoppers would make sense, but the children don't really have the stamina for even a full single day. Amber fell asleep for the World of Color part, and Peter fell asleep while riding on my shoulders on the way back.

Overall, fun, and good to see with friends, but I don't think we'll do the trip the same way again.
27 August 2011 @ 02:45 pm
At Legoland, I picked up three more of the Lego games: Banana Balance, Pirate Plank, and Minotaurus. (There was a 30% off special on games, and I kind of wanted them anyway.)
We have a few of the other Lego games, and they've been popular with the children, especially the Heroica sets. Most of the games are fairly simple (at least by my standards), as appropriate for the target age ranges. Still, the games do encourage tinkering with the rules, which could add some depth.

In that regard, Pirate Plank is a welcome surprise: The key mechanic is fairly simple: Either add another player's color to the die, or advance the pirates for each square already on the die. Unlike most of the Lego games, the optimal strategy isn't immediately easy to solve for. There's a tension between moving the other pirates early on, and adding markers for later play.

Banana Balance has a physical dexterity aspect to it: If you knock the monkey off the tree, it costs you a banana, and knocking over the tree costs you all your bananas (and ends the game). I anticipate being able to balance the game with rules like "Daddy has to play one-handed".

Recommended (any of them) for the playing-games-with-young-children case, not so much for grown-up play.
Tags: ,
19 August 2011 @ 08:34 pm
Occasionally I see comments about some company hoarding cash, often in a Keynesian context of how much better the economy would be if these companies would just spend everything they could.

But I don't think that's the relevant factor. It's probably more useful to think of it in terms of raising money in the credit markets. Companies don't have particularly smooth cash flows: Some quarters are better than others, and expenses can be quite uneven if they include big-ticket items like land purchases or acquiring other companies.

Many companies can borrow easily, particularly banks and bank-like entities with liquid collateral. But places at the other end of the spectrum (tech companies, for example) can't borrow as easily: the interest rates are higher, and it's harder to borrow at all.

So they do the next best thing: They retain some of their earnings (instead of paying dividends or some such), and invest that in the short-term money markets. In doing so, they buy the finance flexibility of a higher credit rating. Instead of borrowing at the last minute, they sell liquid securities, or just don't roll over short-term loans. So, to a degree, differences in borrowing ability or willingness to move investments around get arbitraged away.

This isn't just limited to companies: China's amassing of huge foreign-currency reserves could just be a workaround for the fact that foreign lenders aren't eager to loan into a less-transparent economy with a local currency.
03 August 2011 @ 09:34 pm
Beth and I were doing some trading earlier today, and were looking at a stock we might want to own. Usually we do this sort of thing with options: Sell a put contract, so we get the stock if it goes down (and keep some money if it goes up). But it goes ex-dividend soon, which favors a buy-write, but none of the call strike prices were really that appealing.

So we just bought the stock.

(Well, we placed an executable limit order: A market order is bad during a flash-crash or other weird situation.)
02 August 2011 @ 12:16 pm
Last night Beth & I went out to see Cowboys & Aliens.

The movie is one that depends a lot on your expectations: If you go expecting a mindless action flick, it's really pretty good. But if you expect depth, you'll be disappointed. It did, as promised, have both cowboys and aliens in it.

01 July 2011 @ 10:27 am
Tiassa, by Steven Brust

Tiassa is the most recent Vlad Taltos novel, and fits the series nicely. It's a compelling read, and fills in still more details of the Drageran world. Its unexpected element is a section about Khaavren, in the style of The Phoenix Guards. I like both styles, but it's a bit disconcerting to see them together. The two views do provide a fascinatingly more complete picture of Khaavren.

Unlike some of the other recent books in the series, this one is essentially spoiler-free. There are certainly references to major spoilers, but ones that won't necessarily give away the surprises.

Highly recommended.
I recently bought a Wacom Bamboo stylus for use with my iPad, on the theory that for some uses, a stylus is a better input device than a finger.

In actual use, it's a sometimes thing: There are a number of tasks for which I prefer to use fingers (especially if multitouch is relevant), some for which either was is fine, and some (drawing) for which the stylus is definitely better.

It turns out that there are quite a few types of styluses for sale, and from reading reviews, there's a lot of personal preference involved. I picked the Wacom one because, having used some of their tablets, I'm convinced that they know how to make good styluses. Their entry, at about $30, is far from the cheapest choice, but looks to be durable and well-designed for the role. I had somehow expected the tip to feel solid, but it's clearly hollow and deforms easily. That's an advantage, as it winds up giving a nice tactile feedback about contact.

It works with the iPod touch as well, but that usage seems less compelling somehow.

Recommended, though personal preferences will matter a lot.
28 June 2011 @ 03:47 pm
I was just walking next to Amber, and she was holding my umbrella (uncharacteristically for this date, it's raining).

I'm really not used to looking down at the top of an umbrella.
31 March 2011 @ 11:00 pm
Doing xkcd in 3D is interesting enough, but the method used is very xkcd: It's the lightly-tuned output of a computer-vision system. Words fail me.
Tags: ,
25 March 2011 @ 08:26 pm
Beth and I were discussing a statistics problem the other day, and it turns out that no one had ever explained to her how to do a multiple-variable linear regression. (Possibly because it doesn't come up all that often.)

So I showed her the derivation and method for solving one, and she got it very quickly. I've grown accustomed to having to work hard explaining math, and it was break to just have someone get something more or less immediately.
18 March 2011 @ 11:40 pm
A few weeks ago I had a fever, which presented the opportunity to corrlate what I felt like doing with my temperature.

(in Celcius, as the numbers happen to work out that way)

At 37, I feel like doing normal stuff.
At 38, I don't feel like doing anything.
At 39, I feel like lying in bed (sleep optional).

Fortunately, I didn't have the opportunity to take data points outside that range.
Tags: ,