Live from Austin!

A recent BBC Panorama documentary has suggested that wifi Internet might be a major health hazard. Scary quotes about chromosome damage and radiation exposure have appeared all over the Internet.

Unfortunately, the documentary’s conclusions are junk science.

Let’s start off by noting the inverse square law, a piece of basic physics which applies to electromagnetic radiation exposure. Basically, the strength of a signal varies in proportion to the distance squared.

The people who put together the documentary measured the wi-fi signal at a distance of 1m, and the cell phone tower signal at a distance of 100m. From their measurements, they concluded that the wifi signal was “three times the highest level of the mast”.

Well, no, it wasn’t. Because the cellphone signal was measured 100x further away, it was attenuated by a factor of 100×100 = 10,000×. So an accurate quote would be that the wi-fi signal was “three times the level of the cell phone mast divided by 10,000″. Not as exciting, though, is it?

You might argue that it’s reasonable to measure at different distances because people don’t tend to sit close to cell phone masts, but they do tend to sit close to wi-fi equipment. However, think for a moment about how a cell phone works. Yes, the mast transmits a signal to your phone, which is 10,000× weaker by the time it gets to you. However, you don’t just listen to your phone; hence, it must also transmit your voice back to the network. And the same physics works the other way: the signal your phone transmits is 10,000× weaker by the time it gets back to the mast.

So as you might guess, the radiation your phone emits is much, much more powerful than the radiation that reaches you from any nearby mast. That’s the radiation levels the BBC program should have been measuring and comparing with wi-fi.

While raw power is measured in watts, the relevant measurement for assessing radiation danger levels is the Specific Absorption Rate or SAR, which is measured in watts per kilogram. An adult’s body has much more bulk to dissipate the electromagnetic field, hence it’s less susceptible than (say) a lab rat’s body.

The US limit on radiation from consumer mobile phones is 1.6W/kg. (That’s lower than the European limit, so we’ll take that as our guideline.) The limit for devices like wifi is a mere 0.08W/kg. (Figures are in FCC OET Bulletin 56.)

Those are the maximums. The actual SAR ratings of common mobile phones are well documented. A value of 0.9 is fairly normal, with few phones below 0.5. So already, it’s clear that the average mobile phone actually exposing you to 0.9W/kg is likely far more dangerous than the 0.08 W/kg theoretical maximum allowed for devices like wi-fi.

Let’s look at some actual figures for wi-fi output compared to phones. I haven’t managed to find SAR ratings for wi-fi (if you have any, let me know), so we’ll have to compare power output in both cases. Peak power output from a phone is around 2W, with the average being around 250mW, according to a handy page from the Australian Radiation Protection and Nuclear Safety Agency.

Coincidentally, 250mW is the absolute maximum power output you can get from my wi-fi router (a Linksys WRT54GS). But to get that, you have to hack the firmware. The default power output for the router is around 20mW, 100× less than the phone. Now add in the fact that mobile phones are held against your head, whereas your wifi antenna is likely at least 30cm away from you at all times, and invoke the inverse square law again. The end result is that the electromagnetic radiation you get from wifi is a tiny fraction of that which you get from mobile phones.

As a UK Health Protection Agency scientist puts it in The Times, “a year sitting in a classroom near a wireless network is roughly equivalent to 20 minutes on a mobile.”

Aha, you say—what if you don’t use the mobile phone much? I’m afraid you still get irradiated. As you travel around, the signal from the nearest cell will get weaker. The phone checks signal strength every 7 seconds. If it drops off too far, the phone sends out an “I’m here!” transmission in order to locate another cell.

Given that each cell covers a square km or two, or as little as a few blocks in cities, taking a quick drive across town can involve your phone transmitting dozens of times. So wi-fi is a pretty negligible concern compared to carrying a mobile phone, let alone using one.

Wi-fi and phones aren’t the only sources of electromagnetic radiation, though. Wi-fi operates at 2.4GHz, which just happens to be the same frequency as your microwave oven. In fact, you may have noticed that your wi-fi signal strength is lower if you’re cooking something in the microwave, especially if your laptop is in the kitchen.

You might wonder why wi-fi operates at the same frequency as microwave ovens. Well, microwave ovens operate at 2.4GHz because that’s the frequency that’s best for heating up water molecules. For the same reason, it’s a bad frequency for long distance telecommunications through damp air, so it hadn’t been grabbed for any major commercial purpose. Hence, it was declared as free unlicensed spectrum for local low-power radio. This lack of regulatory hurdles led to innovation such as cordless phones, wireless video surveillance systems, and (eventually) wi-fi.

This also means that the effect of microwave exposure at wi-fi frequencies is simply heat. It’s not like nuclear radiation, it doesn’t mutate your genetic material; it simply warms up your water molecules a bit. From a scientific perspective, people are having a hard time coming up with theories to explain why localized warming of the body might cause damage. (In fact, it’s reported anecdotally that sailors on night watch on deck during WW II would stand in front of the radar in order to keep warm. They got hundreds of times the electromagnetic radiation warming you could ever get from a phone, yet they apparently didn’t suffer major damage.)

But let’s head back to the kitchen. Microwave ovens are allowed to leak up to 5mW/cm² at 5cm distance. A leaky oven may expose you to 0.256W/kg, at the same 5cm distance, according to measurements of leaky microwave ovens from the Australian Radiation Protection Agency. So at typical watching-lunch-rotate distance, it’s about the same level of danger as the radiation from your wifi router. So if you’re worried about wi-fi, you should be worried about your microwave oven too.

But there’s a much larger source of microwave radiation in your life. It’s called the sun. Summer sunlight at ground level can be up to 100mW/cm² of electromagnetic radiation. So standing outside on a sunny day irradiates you with 20× the radiation of a leaky microwave or wifi router, and a good chunk of it is microwave frequency.

So if you’re worried about electromagnetic radiation, perhaps the rational thing to do is what us computer scientists do—stay inside and browse the Internet via wi-fi, but never emerge blinking into the daylight…

Meanwhile, there are a growing number of people who believe that they are sensitive to low levels of microwave radiation like that found in wi-fi and sunlight. They call the phenomenon “electrosensitivity”. They claim that wi-fi and mobile phones give them headaches, make them nauseous, and so on, after just a few minutes. So, what’s the evidence?

Well, so far there have been at least 7 separate scientific trials in which allegedly electrosensitive people were asked to tell researches whether a mobile phone signal was present. In proper double-blind trials, “electrosensitive” people were unable to detect a mobile phone signal even after 50 minutes of continuous exposure. (Update: Here’s a very recent one.)

And even if they could have detected the signal—which they couldn’t—that wouldn’t have proved that the signal was responsible for their reported symptoms.

So if you believe wi-fi or mobile phones are making you ill, please do see a doctor—specifically, a psychiatrist. You may have a psychosomatic illness, you may be schizophrenic and believe you are picking up radio waves with your teeth, but you are not being made sick by people’s wireless Internet. If you truly believe you can tell when a mobile phone or wi-fi system is transmitting, sign up for a research study and prove it. You’ll be the first.

The sad thing is, though, that there are crackpots in even the highest levels of government. So expect to see more scare stories about wi-fi in the next few years.

I’m old enough to remember that back in the 80s, the scare story was about overhead power lines. They were making us nauseous, giving us cancer and leukemia, causing headaches. Funny how that risk seemed to vanish.

My previous PDA was a Palm V. 16MHz 68000, 160×160 B&W screen that could do greyscale in special modes that most software didn’t use. I didn’t particularly want to replace it, but there were a few issues I was having.

First off, the fact that it was serial based rather than USB meant it was a pain to connect to any modern computer; getting it hooked up to the Mac involved a USB to serial adaptor, special drivers, and a lot of futzing with Palm Desktop, and the end result was painfully slow. As a result, I hadn’t synced it in ages.

Secondly, I’d never liked the screen. Going from the Newton MessagePad to the Palm had been a serious downgrade, necessitated by Jobs killing the Newton. I’d been waiting around for some usable Palm devices with at least 320×480 resolution and a 10cm screen, and they finally started appearing in the last year or so.

Other than USB and a big screen, I didn’t really care too much about fancy features; just the obvious stuff—a beeper you can hear easily for alarms, either Bluetooth or wi-fi with an option for the other one, enough memory for a comprehensive GTD list, and maybe a few games and e-books, and connectivity to Mac and Linux.

I did consider the “smart phone” option (again). I came to the same conclusion as last time I considered it: it just doesn’t work. Generally speaking, I want my PDA screen to be as large as possible, and my phone to be as small as possible. Specifically, I want my PDA screen to be big enough to be usable for reviewing an outline of a hundred or so items, and my phone to be small enough to fit in the pocket of my jeans. The Treo 650 fails both tests—it’s too big for a phone and too small for a pocketbook. I’m sure it’s just the right size for some people, but not for me.

Pocket PC devices? Yeah, right. Even if I was prepared to assist Microsoft’s plans for World Domination, the Pocket PC is pretty much crippled unless you run Windows and/or Office, and I don’t run either. So Microsoft eliminated themselves from consideration.

Nokia Communicator? Tempting, but Nokia don’t seem to sell it in the USA. Or at least, I’ve never seen one, and I’m not buying one without seeing the screen first.

Psion? They seem to have given up on the consumer market, and they always price-gouged outrageously for proprietary peripherals and upgrades. No thanks.

Sony Ericsson P series? Again, I’d never actually seen one, and didn’t fancy buying sight unseen, especially not after my experiences with the Sony Ericsson t68i. Plus, you know, $700…

Zaurus? Tempting again. I kept waiting for Sharp to start selling the clamshell Zaurus machines. Unfortunately, all they sell in the US is the SL-6000. It’s thick and heavy and has a keyboard that’s too small to use, hidden in a sliding mechanism. I don’t like sliding mechanisms, they’re too unreliable. I met some guys from Sharp at a show, and told them I hated the 6000, and when would they be selling the SL-C7xx series or some other clamshell design? They said that they both wanted clamshell Zauruses too, but the Japanese mothership had decided that we were wrong, nobody in America wants clamshell machines. I wasn’t prepared to pay $800+ to Dynamism for an unwarrantied Japanese import Zaurus re-flashed with a partially translated OS, though clearly the fact that some people are rather puts a hole in Sharp’s official position. Anyway, I waited a year or so to see if Sharp would relent and sell my a PDA I wanted, then crossed them off the list when they failed to do so.

So, my short list of options was: Tungsten T3, Tungsten T5, Sony CLIE PEG-TH55, Tapwave Zodiac.

The Tungsten T3 has a gratuitous sliding mechanism. The slider wasn’t going to protect the screen, and I couldn’t imagine any concievable circumstance where I’d want to make the screen smaller than it already was, so what was the point? The T5 kinda illustrates the uselessness of it. I expect it was purely a matter of wanting to recycle the case of the Tungsten T and just drop in a different screen instead of having to do a major redesign. So, not really very appealing.

Ah, the Tungsten T5. Looks like a fabulous device until you read some reviews. The biggest problem is that Pa1mOne b0rked the OS on the T5 and the Treo 650, so that every single database entry is now allocated in 512 byte chunks, like on a desktop PC. So if (like me) you carry small databases with hundreds of phone numbers, to-do items and scraps of info, suddenly they bloat out to 10x the size. To me, it seems like that rather ruins the point of the thing. Palm say they are going to fix it, but the fact that they’ve given Treo 650 owners free memory cards to make up for it suggests otherwise, and they’re keeping very quiet about fixing it for the T5. The T5 has 128MB, so allowing for bloat of the kind Treo 650 users have observed, it’s like a 32MB machine—and Palm think that’s OK.

Even ignoring the memory issue, though, there are other problems. The T5 has software compatibility issues. Most software developers are scrambling with updates, but good luck getting classic Palm freeware to run on it. The connector for sync is yet another new design, so none of the existing peripherals will work. And worst of all, it has no vibration mode. Yup, if you’re in a meeting, cinema, church or whatever, you can’t have it vibrate instead of beeping for alarms. It’s the same rather anemic speaker as the T3, and it’s mounted in the center of the back of the device, so as soon as you lay it on a desk, put it in a carrying case, or even hold it in your hand, the sound is badly muffled.

So in short, the T5 fails the basic functionality requirements due to some very poor design choices by Palm.

Next to be eliminated was the CLIE. Sony decided they weren’t interested in selling in the US any more. That left the Tapwave Zodiac, and I bought one.

Things I like about the Zodiac:

• It has the biggest rechargeable battery capacity of any Palm device.
• The case is made of metal, not plastic like the T5.
• Because it was designed for gaming, it has two front-mounted speakers for loud stereo sound, and a strong vibration function.
• It has a graphics processor, leaving the CPU free to do actual CPU stuff, so performance is lightning-fast.
• The internal memory behaves as regular Palm memory, with most of the free space being used to simulate an internal memory card. This is important because Palm OS craps out once files get large, so your photos and MP3s and e-books need to go in “card memory”. With the Zodiac, you get some “card memory” built in…
• …and then you’ve got two SD expansion slots, one of which takes SDIO cards.
• A proper navigation joystick and a complete set of application buttons.
• Most color Palm software seems to run, including titles which I know don’t run on the T5.
• Real headphone jack capable of driving a pair of portable Sennheisers.
• It’s black.

Things I don’t like:

• The stylus just clips on the back. I can see that getting lost.
• The sync cable is hard to clip on; it tends to feel like it’s clipped on, only to suddenly drop off half way through a data transfer.
• Case and dock cost extra. C’mon, guys, would it kill you to bundle a cheap neoprene carrying pouch?

Basically, it’s the nicest Palm device I’ve ever seen. It’s a shame that Tapwave’s strategy is to sell it as a game console, because it’s not so hot at being one of those. As a Palm organizer, though, it easily beats the competition—at least as far as the hardware is concerned.

After several days of enthusiastic rain, the weather finally let up this afternoon. We had gone to the Spider House Café, and I’d done a good chunk of work, so when the sun came out we decided to go and see the new Zilker Park Waterfall. It seems that the Colorado river is rather beefier than usual, and is flowing over the dam in a great brown stinky torrent.

Since we were downtown and hungry, we decided to go to Z’ Tejas. We’d almost gone there once before, but had made a hasty retreat when faced with a sign saying ‘Valet Parking’. Where I’m from, valet parking means you need a personal loan to afford the appetizers.

However, it turns out that that’s not the case for Z’ Tejas. I guess you could view it as expensive by Austin standards, in that you can easily spend $30 each if you go for an appetizer and dessert. Anyhow, the food was excellent—we had enchiladas that were packed with more mushrooms than Timothy Leary and drenched in mole. Dessert was the ancho chile fudge pie, declared the second best food treat in Austin in a magazine this month. I can believe it. Who knew that adding chile to chocolate fudge would work and bring out the chocolate flavor?

Incidentally, working in small coffee houses is very feasible here. Everywhere has free wi-fi. In Austin, even Hooters has free wi-fi, I keep meaning to take a photo of the big orange banner they’ve put up to tell everyone.