When I go to a high point in this city and look down, I see countless flat roofs that could easily host solar panels. Even with all the fog this city gets, that would make a significant impact on our use of non-renewable energy. But it is not to be. Homeowners tend not to like the upfront expense, they tend not to know about SolarCity, and a bunch of the homes are rented. Absent some regulation, they aren't going to install renewable energy.

I think the neatest time to add renewable energy to a building is during construction. Absent that regulation, unless the owner makes it a priority, then the architects are not going to add it to the plan. For example, my work place recently commissioned and moved into a new building. It has an unobstructed, south-facing, 2-story-high, 10-foot-wide window that we have to cover up on the inside to maintain the climate. My immediate thought was: Solar energy. But I had no authority; the people in charge just put a poorly designed curtain on it. It just doesn't occur to them that we could put renewables in this building.

Actually, in the current political climate, I think renewable energy gets negative publicity from these deployments. Conservatives under the thrall of Koch money see renewables as an admission of AGW, and reject it. No! That reason is stupid! And regardless of AGW, renewables will help us survive the depletion of the oil reserves! The Koch-funded people claim that there is no depletion. I live in a state of extreme pessimism.

That makes no sense. The diesel buses can handle some pretty steep hills. On the other hand, many trolley lines are on pretty flat areas. Especially the 14, which barely goes 500 feet up or down over a 10 mile distance.

As a regular rider, trolleys are aggravating because they're slow. If they go too fast, the wires pop off, and the driver then walks to reconnect them. Some drivers respond by driving very slowly, and the other trolleys accumulate behind them, because it's impractical for trolleys to pass each other. This is especially annoying when multiple lines share a significant segment of wires, as the 14 and 49 lines do.

I guess the trolleys are sort of nice because they have zero emissions and they're much quieter at climbing hills than the diesel buses.

When I was in 7th grade, a teacher asked me to prepare a paper on the class's computer. I found that the CRT could handle a higher resolution without problem, so I switched to that. A couple hours later, I was in trouble for "breaking" the computer. That was my introduction to how laws like the CFAA happen.

The problem is that AP classes are, pretty uniformly, badly constructed. Half of the education in AP math and science courses is How to Use the TI-83 Calculator. Half of AP Computer Science is How to Program in Java. The College Board is single-handedly blocking progress in the education of technology in math and science.

I don't know about the rest of the AP classes. I also think the College Board's role in college admissions, via SAT and AP, is fragile and counterproductive.

The init system is a very poor example of Unix common interfaces. As beelsebob and oursland point out, different Unix systems use different init systems. The Linux alternatives, upstart and systemd, were actually inspired by the clear advantages displayed by MacOS X's launchd.

And even in Linux, with SysVinit, there are different interfaces. When you want a script to run at boot, do you use update-rc.d, like Debian? Do you use rc-update like Gentoo's OpenRC? Or chkconfig like Red Hat? Or insserv like SuSE? And where do you find important details like the hostname and network configuration?

I don't find systemd to be a pleasing design, and I especially don't share their love of long command names with lots of consonants, but I think their work is very important.

It's still frustrating for the residents here.

I, for one, know that the narrative is far more complicated than just VC-funded rich dudes conspiring with greedy landowners to drive up rents. Also, I am well aware of the laws of supply and demand. The supply does not match the demand at all. It's much worse this time than last time, the dot-com bubble of the 1990's.

I'm even aware of a little-discussed wildcard: China. The financial system there is corrupt, and the people have no safe way to invest for retirement. The burgeoning middle class is desperate for options. You might remember how they drove the price of Bitcoin to over $1000 before the regulators caught on and outlawed Bitcoin exchanges. Well, another option is real estate. The poorer people invest in Chinese cities, fueling an unsustainable construction boom over there. The richer people invest overseas. Whenever a single-family dwelling goes on sale in San Francisco, it's immediately snapped up by a Chinese investor with cash. No need for a mortgage.

The effect is that I don't know any ordinary young people who can afford to live in San Francisco except with their parents. When people do move out, they move to South San Francisco or Oakland and commute, or further. A few people manage to win the lottery of Section 8 Housing or other subsidized housing. These ordinary people include the professionals who teach your children and staff your restaurants. High living costs are natural, but they are not sustainable and you shouldn't think of them as desirable. For one thing, making all your workers commute is bad for the environment.

Some San Francisco natives work in tech. For the rest, this is not a good situation.

Bill Shockley was the originator of the Silicon Valley arrogant genius archetype. One of the co-inventors of the transistor, he convinced an electronics entrepreneur in the Los Angeles area to pay him to set up a semiconductor laboratory near his mother's home in Palo Alto, staffed with young geniuses. Then his abrasive management drove them away, leading them to found Fairchild Semiconductor, followed by Intel, AMD, and other, less important, electronics companies in the area. In the meanwhile, Shockley went into eugenics.

HP was already around, and Fred Terman of Stanford was encouraging entrepreneurship, but Shockley brought the "silicon" to Silicon Valley. And the arrogance.

The WRT54GL is a relic of an ancient time. Most importantly, it's a relic of a time without IPv4 address exhaustion, and without realistic demonstrations of DNS cache poisoning.

DD-WRT has support for 6in4 and 6to4, but not as much support for IPv6 over PPPoE or DHCP-PD or Sixxs.net AYIYA. I prefer OpenWRT, but I also prefer plain-text configuration via the command line, so I'm weird. OpenWRT officially dropped support for the WRT54GL in the last stable release, 12.09 from April 2013, and it didn't really work right in 10.03, either.

I've been generally pleased with routers based on the Atheros AR7161, but those are obsolete (only N300 and N600), and not that easy to find. Probably the most famous from that line is the Netgear WNDR3800, the target model for CeroWRT and the EFF Open Wireless Router. 680MHz MIPS24K, 16MB of flash, and 128MB of RAM are so luxurious after the 200MHz BMIPS3300, 16MB RAM, 4MB flash of the WRT54GL.

I wouldn't depend on Buffalo or DD-WRT. DD-WRT is tolerant of closed-source drivers, which leads to long-term maintenance problems. I prefer to look for OpenWRT support. Actual support, not that fake press-release support that Belkin-Linksys did with the WRT1900AC and its lame Marvell chipset. Actually, since the WZR-600DHP is discontinued, I wouldn't recommend any of Buffalo's products right now. I don't really recommend ASUS, either.

The WZR-600DHP is good because it's built around the Atheros AR7161. Atheros donated the driver and wireless firmware to the open-source community. The WZR-600DHP2 is a completely different device built around the Broadcom BCM4708. You can't get 40MHz channels or even the Ethernet driver to work on those things without closed-source drivers. Almost everything from ASUS is powered by Broadcom.

I'm cautiously optimistic about current-generation Qualcomm Atheros devices. The QCA9880-AR1A is no good, but the QCA9880-BR4A seems decently supported in OpenWRT. But I can't be sure until I have a device to play with.

You should care more about the firmware and driver source availability than about the manufacturer. It's because, no matter how strong and how fast your router is today, tomorrow your router is slow and obsolete. When (not if) problems are discovered with your device, the availability of updates depends on the ability to recompile the firmware.

I like my Buffalo WZR-600DHP. It came with DD-WRT, but more importantly, it was built on the Atheros AR7161, like the Netgear WNDR3800, Ubiquiti RouterStation, Mikrotik RB-450G, and several others, so I prefer to put OpenWRT on it. Sadly, this chip is several years old now, and doesn't support 802.11ac, and Broadcom offers cheaper N600-N750 chipsets, so there aren't a lot of AR7161 routers. Also, some of the early AR7161 routers are a little flaky, like the Netgear WNDR3700v1. My uncle had one where the 2.4GHz radio died.

Usually, I'm opposed to Cavium, Broadcom, and Marvell, and suspicious until proven otherwise of Qualcomm Atheros, MediaTek, and Realtek. Sadly, that means I can't recommend any 802.11ac routers. The most likely to work might be the ones with the Qualcomm Atheros QCA9558 and QCA9880-BR4A combination, like the Engenius ESR1750 and the TP-Link Archer C7 v2 (not v1). Since I don't have personal experience, and the documentation is so sparse, I can't recommend those without reservation. If I had to buy an 802.11ac router right now, I would buy one of those.

This is not an argument for federal education standards. This is an argument for fundamental education reforms. "Oh, I'm sorry, we can't talk about arithmetic on mixed fractions this year, because that's a 4th grade subject. This is 5th grade. We're doing geometric figures." Or whatever. What about the 5th graders who didn't really get mixed fractions last year? Many of the best mathematicians were made to feel stupid in school because they would rather think slowly than rush through all the subjects in the scheduled time.

Jo Boaler has been arguing that math education should be centered around Low Floor High Ceiling Tasks. Then it matters much less when your student enters the class, because they can learn from the activity at whatever level they've mastered. Somewhere else she argues that students should work on projects over an extended period of time.

The annoying part is that educational approaches take a very long time to see if they're really effective, so it's annoying to work out what is BS and what is useful out of the things that educational reformers say.

I think this is a horrible, bad idea because we don't really know what computer science is. It's such a young discipline that many of the important pioneers are still around.

Well, the very first generation, the people who figured out how dancing machinery could represent arbitrary mathematical operations, those people died a generation ago. But many of the foundations of modern computer science, those were pretty arbitrary, and those people are either still around or recently dead: John McCarthy of LISP (1927-2011), Donald Knuth of sequential algorithms (TAOCP, b. 1938), Douglas Engelbart of the human-centered GUI (NLS, 1925-2013), Claude Shannon of Information Theory (1916-2001), Paul Baran of packet networking (1926-2011), Edsger Dijkstra of structured programming (1930-2002), John Backus and Peter Naur of programming language specification (BNF, 1924-2007 and b. 1928), and so on.

Naturally, there are disagreements about what exactly computer science should be about. Dijkstra argued it should be fundamentally mathematical, and forbade students in his intro to CS class from touching a computer or trying to "run" the algorithms that they worked through. Abelson and Sussman said it should be about program structure and interpretation, and their intro to CS class uses a language intended for clarity of teaching rather than for efficient execution. Some people think it should be about algorithms, as seen in those Code.org drag-and-drop algorithm block exercises. Clearly, most people think it should be about writing programs in whatever programming language is commercially useful, so most intro to CS classes are about Java. Yuck.

Since there is this wide variety of opinions about what computer science should be about, and especially the wide gulf between what the best do (MIT, Berkeley: SICP, in Scheme or Python) versus the worst (College Board, Community Colleges: Java), I think it's very premature to ask politicians to start mandating CS across this nation. You just know, whatever they decide, it will be wrong and slow to change. Let the field shake out another generation or two, and our grandchildren will see if the subject has matured enough by then.

As with the existing technological hassles in my life, I would use a smart watch only if it did something significantly new.

In the old days (1980s), my laptop would go weeks without a battery charge. Now, my laptop barely makes it through a day, if I'm not actually using it much during that day. But my new laptop is vastly more capable, with high-DPI IPS display and 802.11ac WiFi and the ability to run a C++ compiler many times in a single hour.

In the less old days, my phone would go a bit over a week without a battery charge. Now, my phone usually makes it through a day, but not if I'm using its GPS or its processor extensively; and it's much bigger. But my new phone has a camera that doesn't entirely suck and a lot of apps, some of them useful, and visual voicemail. Still, I wouldn't have bought it if it didn't have another compelling feature: Really cheap unlimited plans.

That's 2 devices that I have to plug in every day to keep using. A smart watch would be a third. So far, I haven't heard of any compelling features. The current crop has what? The ability to show notifications. Which my phone already does when I leave it on the table next to my mouse, and which I'm already consciously choosing to ignore when I want to maintain focus. And Samsung's watch has its trademark heart rate sensor, which works only if you're not exercising.

I can imagine some uses for a smart watch, in concept, if it could do stuff independently of the phone. A camera that you don't even have to dig out of your pocket (or purse, if you have a Samsung). A communications device that you can carry without pockets. A security/control device (if it doesn't come from Google, Apple, or Microsoft, and runs free software). The concept is interesting. It just needs good execution.

The main purpose of AP Statistics (and AP Calculus) seems to be to teach limited subsets of the functionality of the TI-89 calculator series. The programmability features of that calculator are never taught in American schools.

Not that AP Computer Science is much better. Its main purpose seems to be to teach the Serious Programming Language du Jour, currently Java. Any algorithmic learning has to happen in between the struggles with that language.

I'm not pleased with the College Board's position in American society.

I refuse to build a Connected Home without Free Software. Imagine the security nightmares of SCADA and consumer electronics, together at last.

This has to apply to the drivers and the peripheral firmware, too, because the Linux kernel has its own vulnerabilities.