Monday, August 27, 2012

Protecting Balls

I'm not going to comment directly on the Apple v. Samsung patent lawsuit, mainly because my employer wouldn't like it (and I obviously don't speak for them), but also because the reader would just assume I'm biased.

However, I share the apprehension that many have expressed about patents. Even if you agree that Samsung copied the iPhone and should be stopped, the power of even a single well-placed patent to block an entire class of products is fearsome. If memory serves, over a decade ago, Geoworks had a patent in the WAP browser area, and it propelled the stock from something like US$30 per share all the way to US$180. This is a lot of power.

The solution that many of these patent opponents propose, however, is extreme in the other direction. They want to abolish all patents, or at least all software patents. First of all, very few if any established players are going to support that, and lobbying dollars aside, why should the government ignore some of the country's most successful companies in favor of ranting Internet advocates? Secondly, what would that achieve?

The iPhone was a hugely expensive gamble. There was a mountain of software to develop, and Scott Forstall recalls that Steve Jobs gave him permission to bring in anyone from within the company to work on the iPhone. Forstall then went around and found those he described as "true superstars of the company" and "amazing engineers", and worked them incredibly hard for years. (No, he didn't invite me.) This is no less than betting the company, because stripping every other project in the company of their superstars has an easily-predictable negative effect that I need not belabor, yet they placed that bet.

In other words, to me it's not so much whether there was prior art to pinch-to-zoom or whatever other legal details the poor judge and jury had to slog through, but how to protect the "balls" that goes into a truly innovative product that changed an industry. The point is, even if you didn't think any of the individual distinguishing features of the iPhone were worthy of patent protection, you probably should admit that the total end package redefined smartphones. However trivial you think those inventions were, the fact is that upstart Apple jumped ahead (at considerable risk to itself, let's not forget) technologically, while the rest of the industry rehashed their allegiance to carriers by shipping yet another crippled phone.

Another point that I've seen made is that Apple was handsomely rewarded for its venture, so that should've been enough incentive to innovate. I can't agree with this point of view, true as it may be, because it can be made only in hindsight. Apple couldn't know how much it would make from this phone that it hasn't even fully developed, couldn't know if people would really accept typing on an on-screen keyboard, couldn't know if other carriers could be persuaded to carry the iPhone, and couldn't know a hundred other risk factors. If you still think Apple would've done it anyway in a world without patent protection, consider what would allow another company to leapfrog Apple, without worrying that Apple could easily ape the new invention and crush it with scale? The decision to go for it must be made long before you are assured of wild riches, and a lot of big changes can really only come from big players with lots to lose.

So the question for society is, if not patents (which I'm not really defending here, because there are indeed good reasons to worry that they are too powerful), then what? Is the answer to all future innovation really going to be "Google supporting it with ad revenue" regardless of whether it ever makes its own money? I hope not, because I don't really like that world.

Saturday, July 21, 2012

Numbers Don't Lie

When I first joined Apple in mid-2005, the stock was about US$45 a share. When I left in early 2008, it had gone up to US$180 or so. The stock floundered while I was away, and when I came back in late 2009, it was at US$190 or so. Since then, it's gone to over US$600. Look it up yourself, it's just cold, hard numbers.

I should run for president.

Saturday, July 14, 2012

Lessons from Trying to Buy a House

The Internet has become an incredibly powerful tool in ways that even a practitioner doesn't necessarily realize. When we decided that it was time to start looking for a house to buy about six months ago, the wealth of information freely available on the Internet was astounding. Armed with apps and websites, I was finding many of the properties that my real estate agent was finding. She usually had some more information (such as whether it's a short sale), but I wasn't far behind at all. It was also a great resource for many industry jargon, as well as traps and pitfalls that lie in wait.

The first thing we learned was that mortgage calculators are entirely ass-backwards. The crucial number to start and end with is how much you have to pay every month for housing. From that number, based on how much downpayment you can muster, interest rates you qualify for, and required expenses such as property taxes and insurance, comes the number telling you how much house you can afford. The online calculators start with "Home Value" as a variable you can enter, which is utterly wrong. Home value is the result of the math, not the starting point.

Another thing we realized was that Americans really are bad at math. Apparently many people (including our agent) don't actually understand that paying your mortgage bi-weekly versus monthly is the same thing. In one case, you pay more quickly and therefore pay less in interest. In the other, you get to hang on to the money you borrowed for longer, at the expense of paying more interest. It's just like the difference between a 15-year mortgage and a 30-year one, but it seems to confuse people. This isn't even really math - it's barely arithmetic!

We also learned that by about 1,500-sqft, the American home-buyer apparently wants a formal living room and a formal dining room, instead of just bigger rooms or other general-purpose rooms like dens. Since Americans don't really invite mere acquaintances home for social functions, the close friends they do invite inevitably go straight into the family room and kitchen, so a good third of these first floors just sit there and never get used. Now, if this was Texas I could understand that land is cheap, but in the Bay Area I'd have expected more pragmatism. How about a music room? Asian kids all learn the violin or piano, right? (It's okay, I can say that.) How about a hobby room, where you might at least lay out a table for a jigsaw puzzle? How about a library?

Lots of houses are built in bulk as part of phased developments, where buyers get to pick one of four or five models, and the developer ensures that your house is not identical to the next ones. It boggles our mind that a development might have dozens of these homes, and yet the interior layouts remain spectacularly bad. 2,000-sqft houses would still have cabinets that aren't square, and there'd be all sorts of odd spaces that aren't good for anything. You'd think if you were designing just four houses to build 100 times, you'd make sure each one is pretty good. After all, this is the industry that came up with "measure twice, cut once", isn't it?

Typically, the master's bedrooms take up half a floor, while the kids' bedrooms are too small to fit even a proper study desk. (As long as we're pretending, shouldn't you pretend that your kids study?) The master's bath is grandiose, but there is perhaps just enough closet space for one person. Nobody seems to have any storage space - turns out Americans park their cars in the driveway and store stuff in the garage - but, boy, do they want that master spa and that formal living room.

And what the hell is up with the white tile countertops? A basic requirement for a countertop is that it's a flat work surface!


Sunday, May 20, 2012

The Designated Hitter Solution

In the American League (AL), pitchers do not have to hit the ball. They have a designated hitter (DH) to do it for them. In the National League (NL), pitchers have to (try to) hit the ball. Since the two leagues actually play each other for some games each season (known as inter-league play) and of course for the World Series, you'll see AL pitchers almost never even swing the bat because they just might hurt themselves doing it. It's not at all uncommon to see an AL pitcher with a bat lazily resting on his shoulder, striking out just watching three fastballs down the middle of the plate. Even NL pitchers are almost invariably the worst hitters on their team, normally restricted to sacrifice bunts, and often replaced when their turn to hit comes up.

Some baseball purists argue against the DH rule, yet the fact is that the game is played at a lower level because of the pitchers hitting. Teams will deliberately walk the batter in front of the pitcher's spot (especially when there are already two outs in the inning), because the pitcher is such an easy out. The dilemma for the manager is between pulling a pitcher who's doing fine otherwise because he can't hit, or watching the pitcher waste an inevitable out. How is either outcome actually better?

The lesson of the DH rule is that, yes, in theory pitchers should learn to hit the ball, but the fact is that they can't and they don't. People who are in charge of rules and policies will do well to remember this. For example, in theory everybody would buy health insurance if they could afford it. In fact, a lot of people won't buy health insurance if they can wait until they're actually sick, and their logic is admittedly pristine. Just as wanting baseball to be "pure" doesn't make it so, the DH rule and forcing the participation of people who are too stupid or too selfish to understand how insurance works are the ugly solutions to ugly problems. The purists and the puritans, on the other hand, have no solution to either problem except to complain.

Saturday, March 17, 2012

Challenging the iPhone, Part IV

A lot has happened in the past year. Starting in the Android camp, Motorola has engineered a brilliant sale to Google, HTC has faltered, and Samsung has emerged as the big winner. It's not in Google's interests for any single company to dominate the Android manufacturers (just as it wasn't in Google's interest for Apple to dominate smartphones), so the dynamics between the two would be interesting to observe over the next year or two.

Google also released Ice Cream Sandwich to critical acclaim, but actual adoption remains low, accounting for just 1.6% of Android devices, less than a quarter of ancient Eclair version's share. Froyo is 22-months old now, yet it still runs on over a quarter of Android devices. This slow adoption is a problem for Google in matching iOS features, and makes it difficult for third party developers to make use of new features. Some developers are finding that supporting the multitudes of Android handsets costs more than the effort brings in.

The story of the year, however, should probably be the global patent war. Apple, Samsung, Motorola, and HTC appear to be the principal players, and skirmishes have been won and lost by each party, with no end in sight. An Apple-Microsoft consortium bought thousands of Nortel patents for US$4.5 billion, and shortly after patents again appear to have been the key reason Google paid US$12.5 billion for Motorola. Away from the spotlight, Microsoft has quietly hit up several Android manufacturers to collect patent taxes as well. Even Openwave contributed a patent or two to the war.

Apple had a very good year. The iPhone 4S was released to lukewarm critical response, but the customers loved it, and the Siri voice assistant became an instant icon, appearing even in late night comedy shows. Apple would sell a record 37 million iPhones in the Christmas quarter, part of the reason its market cap surpassed Exxon-Mobil and stands today at a staggering US$546 billion.

In other news, Nokia doesn't seem to be doing a great deal better shipping Windows Phone 7 handsets, RIM continues to circle the drain, and Palm has been put out to pasture. The Microsoft-Nokia alliance may inherit third place by default, but there's certainly no evidence of a third power forming right now. The consolidation I talked about last year has happened, and the mobile landscape of 2012 is iOS versus Android, maybe even Apple versus Samsung.

But RIM deserves a mention here. Falling behind in the smartphone market, RIM allowed itself to be distracted by the iPad, and developed the Playbook with newly-acquired QNX as an answer. The Playbook was a disaster (like just about every other competing tablet), shipping without even an email client, and is now selling at or below cost. Samsung is said to be interested in investing in RIM and perhaps building phones based on the upcoming BB10 software, but I don't think that will be enough. If RIM doesn't ship a killer phone in 2012 by itself, I think they're done as a force in this market.

As for Microsoft, they'll do what they always do: plug along and hope they come upon that magical version 3 that takes over the market. A very interesting factor in 2012 is whether any suspicion between Samsung and Google would overcome their mutual interests, and develop into a real falling out, making Samsung do a real switch to BB10, Windows, or its own Bada operating system. Android is too attractive for Samsung to give up easily, but if Google tries too hard to weaken Samsung's dominance, I think it's possible.


Resolution independence has been a problem for user interface developers since shortly after an abundance of graphics hardware became available. I remember experimenting in the early 90's with floating point coordinates, normalizing each coordinate to between 0.0 and 1.0. This allowed me to scale my UI to whatever screen size. Of course, that was probably not original, and was certainly not effective. Things squeezed together on lower resolutions, and left way too much white space on higher resolutions.

Later, Microsoft also allowed the user to scale the Windows UI to use a bigger system font, but it was also a failure. It would mess up the UI of all sorts of apps that never expected to be scaled.

When the technology for ~300 ppi displays became available for the iPod in 2007, we simply used it as if it was any other screen. The iPod had no third-party apps to worry about, so we simply got the benefit of sharp text and detailed graphics. But when the technology became feasible for the iPhone 4's screen size in 2010, there were already hundreds of thousands of third-party apps that weren't written with this in mind, and could not be expected to all put in the work to fix.

This is when Apple engineers came up with a brilliant engineering solution. They pretended that the number of pixels stayed the same as the older iPhone. Because the OS was responsible for rendering text, most text in the apps (except those the app has baked into a graphic) would look better automatically. Graphics were easily updated by creating a double-resolution version with a special name, with no code changes required, but if they weren't provided the OS would just scale up the original version. This might horrify a computer graphics purist, but it worked startlingly well. Even coming on two years after the iPhone 4, as predicted many apps still haven't been updated to take advantage of the Retina display, yet they continue to work.

This is engineering at its best. Given the constraints, engineers came up with a practical solution that worked well. It's no breakthrough in computer science by any stretch, but the point is that not everything needs to be one. It reminds me of the fable of an elevator engineer who was asked to reduce waiting times, and simply installed mirrors on each floor to give people something to look at while waiting. Engineers will do well to remember this example.

Saturday, March 3, 2012

The Republican Rebuild

Joe Nocera wrote an op-ed in the New York Times to point out that Rick Santorum as the Republican presidential nominee may suffer a defeat so humiliating that the Republican center finally wrests control back from the fringe, which would actually be a good thing for the country. It's just surreal that we're talking about contraception at all in 2012.

I've long thought that the Republican Party was hopelessly contradictory, torn between the small-government types and the religious types who very much want to use the government to force others to do what they think is right. The only thing that seems to be keeping it together is the ability of people to compartmentalize contradictory thoughts - otherwise known as hypocrisy - but it already doesn't look like Ron Paul's supporters would necessarily just fall in behind Mitt Romney or Rick Santorum, and it certainly remains to be seen whether the bulk of the party can actually unite behind either Romney or Santorum.

Nocera is right that a Romney loss will be blamed on his ideological impurity, and the problem will persist. The Republicans need a wake-up call in the form of a massive Santorum defeat, one that is undeniably the American public's rejection of these fringe beliefs.

There was a moment during the enactment of the Affordable Care Act when I realized that the most right-wing Democrats were still left of the most left-wing Republicans (or at least as left as those Republicans dared to be in public). It's a dangerous sign for the nation that the two parties appear to have no ideological overlap, because that's a recipe for legislative stalemates during a critical period of recovery for the country. I believe that if we don't play our cards right soon, the US will not be a technological leader within a generation.

I do find the parliamentary system appealing, where no party would be dominant, and several parties have to compromise in order to govern. That's not going to happen, so the Republican Party needs to be torn down and rebuilt, for the country to get back on the move. Pragmatism must win. If we ever want our leaders to consider more than just two solutions to each problem we face, then at least some Republicans must learn to compromise. "My way or the highway" actually ends up standing still.

Saturday, February 4, 2012

Obama Is...

Obama is socialist and anti-rich, but also guilty of bailing out the banks and car companies.

Obama is Muslim, but also guilty of his pastor Rev. Jeremiah Wright saying "God Damn America!"

Obama is the food-stamp president, but Romney's not worried about the very poor because they have safety net programs... such as food stamps.

Obama is responsible for the bad economy and joblessness, but Bush is not responsible for not preventing 9/11.

Obama is responsible for the deficit due to spending too much, but Republicans don't contribute to the deficit by cutting taxes.

Obama is elitist, but sides too much with the poor.

Obama is connected to terrorists, but ordered the attack on Osama bin Laden anyway.

Obama is weak because he "leads from behind" in Libya, but is too radical and unwilling to compromise with Republicans.

Obama is gutting the military, even though the military budget of the past three years were US$794 billion, $848 billion, and $904 billion, respectively.

Obama is the worst president ever, even though he took office in the most severe recession since the Great Depression, with unemployment having shot up from 5% to 8% (and going) in the previous year, and the collapse of the housing bubble that was powering much of consumer spending. Oh, and two unfunded wars, unfunded tax cuts, and unfunded prescription drug benefits. But anyway, worse than the two-term guy who left him this mess.

How do you people get these straight in your heads without it exploding?