Since I was a kid, I’ve been looking for the perfect way to read in bed. The ideal position would involve no sustained muscle effort, so I could just let my eyes drift shut as I read, without the book falling shut or my hand slipping or anything.  One way is to sit up against something and hold the book on your lap, but that’s not great for falling asleep.  So I usually end up reading on my side.

The problem is, you have to hold the book to see both pages, and in either case, you’re using some muscles to hold it where it is.

This has worked for most of my life, but it’s still not that ultimate relaxation.

However, I recently got a Kindle.  I was intending to use it mainly as a mobile web browser, but I’ve surprised myself by using it to read an awful lot.  And, with apologies to all the bibliophiles out there, I find the ergonomics better than a paperback.  When snacking and reading, I can lay it flat on a table without the use of a book weight to hold it opened, and when lying in bed, I don’t have to keep moving it to read.

But it’s not perfect.  There’s no way to hold it with a finger on the ‘next page’ buttons that doesn’t require a few muscles to hold it upright:

Either I work to hold my hand off the bed, or I awkwardly curl my fingers around it.  Either way, it tips over if I relax my arm, even if it’s leaning against a pillow, and I’m startled awake by this:

I started to wonder if I could do even better.  I got out of bed one night, went to the closet, and got a steel coat hanger and some pliers.  After a few minutes of twisting, I created this:

First of all, it holds the Kindle upright …

And second, it lets me lie there motionless, and turn the pages with just a tiny twitch of my thumb:

Finally, after decades of reading in bed, I have reached that stage of perfect relaxation.

I don’t do conventions very often, but I recently went to ConBust out in Northampton, MA, while visiting some friends.  While I was there, I had a guy propose something fascinating to me.  I can’t remember the guy’s name, so if he or one of his friends sees this, post your info in the comments. (Edit: it was a dude by name of Thom Howe.)

The guy Thom had an idea for a date.  He wanted to rent a cherry picker, drive it to her door, and pick her up in it.

Then, he’d drive to the beach, and get there at just the right time to watch the sun set.

Once the sun had set, he’d activate the cherry picker, they’d be lifted up above the beach …

… and they’d watch the sun set again.

Clearly, this is an excellent idea, and any girl would be lucky to see this guy Thom at her door.  But is it plausible?  How fast and how high does the cherry picker have to go?

I tried to work out the answer for him there at the table, but there was a line of people and there wasn’t time.  But when I got home, I remembered it again, and I’ve worked out the solution.

Here’s the situation:

By the time the earth has rotated through angle theta, the cherry picker will have to have climbed to height h.

After t seconds, theta in radians is:

The height of the lift above the center of the earth is:

So the height above the surface (sea level) is:

Substituting everything so far we get this expression for the height the lift needs to reach t seconds after sunset to stay even with the sun.

Now, an actual cherry picker has a maximum lift rate (I Googled some random cherry picker specs, and 0.3 m/s is a normal enough top lift rate.)  We’ll call that rate v, so the actual height of the lift will be this:

Substituting that in and solving for v, we get this:

(That’s arcsecant, not arcsecond). This equation tells us how fast the lift has to go to get from the ground to height h in time for the sunset¹.

But we can also get the answer by just trying a few different heights.  We plug it in to Google Calculator²:

2*pi*6 meters/(day*arcsec(6 meters/(radius of earth)+1))

and find that h=6 meters gives about the right speed.  So, given a standard cherry picker, he’ll get his second sunset when they’re about six meters up, 20 seconds later.

You might notice that I’m ignoring the fact that he’s not starting at sea level — he’s a couple meters above it.  This is actually pretty significant, since the sunset line accelerates upward, and it brings down his second-sunset height quite a bit.  If he got a faster lift, or used an elevator, the correction would become less necessary.  Extra credit³ for anyone who wants to derive the expression for the height of the second sunset given the lift speed and height of first sunset. For now, I recommend he dig a hole in the sand and park the lift in it, so their eyes are about at sea level⁴.

¹ Ideally, we’d solve for h, but it’s inside the arcsec and that looks like it’s probably hard. Do one of you wizards with Maple or Mathematica wanna find the result?

² If you work in one of the physical sciences and don’t use Google Calculator for all your evaluatin’, you’re missing out.  I wish there were a command-line version so I could more easily look/scroll through my history.  I know Google Calculator is largely a frontend to the unix tool units, but it’s better than units and available everywhere.

³ Redeemable for regular credit, which is not redeemable for anything.

⁴ I suggest a day when there aren’t many waves.

Since last night I’ve been obsessively connecting to Omegle.

It’s a simple site that just connects you to a random person, anonymously, for a conversation.

It feels like that scene in Fight Club where the narrator sits down next to Tyler on the plane.  Two strangers meeting, laying out their personality and sizing each other up in just a few words, with no expectations, and — thanks to anonymity — no consequences.

Except in this case, a lot of the time Tyler just screams “COCKS”, punches the narrator, and jumps out of the window.

Still, it’s fun!

For the people who wanted to know the chain of events causing the laptop problems in the last post:

1. Safe search won’t disable, and login for regular Google services is broken.  I decide it’s a cookie conflict between Google services and file a bug report with Google.
2. Google says “can’t reproduce”.  I discover that clearing private data doesn’t actually work right, and after some testing, go to file a bug with Firefox.  They ask for my version number.
3. I notice my Firefox is out-of-date, and decide that it might be a quirk of this version of Firefox+Ubuntu.  I go to upgrade Firefox in Synaptic before filing the bug.  I don’t have the notifier running and haven’t updated in a while.
4. I upgrade a package in Synaptic, but it turns out I just upgraded the meta package and not the actual package (or something.)  A more seasoned Synaptic user says “why not just run regular upgrades like a normal person?  It’ll fold Firefox into it.”  On any other day, this would have been good advice.
5. I start the upgrade.  It’s churning along nicely, and I locate my cookies file and start examining it.
6. My battery monitor disappears.  This is normal enough, actually.  I try to start it back up and get a notice about a broken configuration file.  Uh-oh.
7. My nm-applet disappears.  That’s a little stranger, but neither applet is very reliable in my ion3 setup.
8. At this point I find that my cookies file is malformed in some way, and that if I move it manually (rather than using the in-browser ‘delete cookies’) the original bug disappears.  Don’t know how that happened, but it’s solved, so I cancel the bug report with Mozilla.
9. I go to edit the cookies file in emacs, and get a message that emacs can’t start.  Uh-oh.
10. The upgrade fails with a bunch of package incompatibility messages.
11. I start to feel alarmed.  I link an apt-guru friend a screenshot of a relevant part of my Synaptic window.  He says, “one of those version numbers looks wrong.  You’re running Hardy, right?”  I am.
12. I open my sources.list (using nano, since emacs is broken.)  After a bunch of spaces at the bottom, I see something bad: A Debian repository.  I’m running Ubuntu.
13. I don’t know how I added it.  Maybe I was on the wrong system and didn’t notice the prompt.  Maybe I was trying to install one specific package from Debian (via apt, for some reason) and forgot to take it out.  I honestly don’t remember.  But since I hadn’t updated in a while, it hadn’t come up.  But now I’m in trouble.
14. I remove it, update my lists, and do a grep to see how many of my packages have upgraded to Debian versions.  750-1000 or so.  I paste the results back to some friends.  One of them looks, shakes her head, and says my system is like that guy in Star Trek after the transporter accident.
15. Over the next few days, we try pinning packages back to the Hardy version and downgrading.  There are conflicts all over and lots of ninjinuity is required.  Sometimes the system won’t boot properly, claiming kernel module problems, which turns out to be because somewhere in this process my initramfs got misconfigured.  But eventually, everything is put back in working order (except, for some strange reason, gnuplot, which refuses to install the main binary file.  I compile it from source.)
16. Watching the system boot successfully, I go to pour some milk and cereal in celebration.  I shake up the milk jug a little.  The lid is loose.  It spills all over the keyboard.  The system stops booting and the cycle of horror starts again.  (And yes, if I’d gotten a Lenovo when you all suggested it, I’d have a spill-proof keyboard with drains.  That’s my plan for the next time something horrible happens to my laptop, which should be any day now.)

1) Dvorak typists: I’m sorry if the paper I relied on has some potential flaws.  If you want to share your extensive rants on the merits of various keyboard layouts, send them to me at doctorow@boingboing.net and I’ll be sure to read them over carefully.

2) You know how sometimes trying to fix one problem causes a worse problem, which in turn causes a worse one, and so forth?  This week, I was trying to turn off SafeSearch in Google Image Search (sometimes necessary to find things¹).  In trying to turn off SafeSearch, a setting on a web form, I rendered my system unbootable². (The worst part was, everyone kept saying “oh yeah — there’s a comic about that; have you read it?”)

Any guesses as to how that happened?  I have to go for now, but I’ll include the actual chain of events in my next blog post.

¹ Like the Nate Silver Playgirl centerfold.
² I’ve finally recovered it³ via a USB boot disk and some work, but it took several days and the help of a number of wonderful hackers. Thanks you, Decklin and sneakums.
³ And then promptly spilled milk on it, so the keyboard’s been through the dishwasher and is now drying⁴ in front of the fan. This is not a happy laptop.
⁴ Nested footnotes!

First, the results of my laptop hunt.  Thank you so much for all the suggestions — there are a lot more ultraportable options than I realized.  I narrowed it down to two wildly different possibilties:

1. The Lenovo ThinkPad X200s.  I missed the 200s on my first survey — I didn’t realize Lenovo had one under 2.5 lbs.  It’s got perfect geek-oriented engineering (tough metal hinges, spill tray, incredibly rugged, Linux-friendly, etc).  It’s just a little bigger than I was going for — it’s pretty large for its weight.
2. The Sony Vaio P.  I love the form factor — 9″ wide but only 4″ deep; you could slip it in a pocket.  And the 220 PPI screen is everything I could want.  It’s shiny and pretty and only weighs a pound and a half, and when I played with one I fell in love.  But it’s a Sony, which means shiny appearance but no openness — repairs suck, and Linux will only work haltingly.

In the end, I decided neither was worth the cost just yet while I could squeeze a little more life out of my old laptop.  A new battery, hard drive, and some antibiotics, and it could last another year.  But if I had to pick now, I’d go with the X200s.  The ability to toss it down a mountainside and have Ubuntu boot when it lands trumps gadgety sexiness.  But only barely.

Kindle 2:

On a contrary note, I have a Kindle 2.  I’ve been really happy with it so far, other than the PDF support being poorly documented (it claims to handle PDFs natively in some of the literature, but you can’t actually just plop them down on the drive — it requires passing them through Amazon or converting them yourself).  But it handles html, text, and some other formats fine when just dropped on it via USB — it’s a lot more open than the Kindle 1.

I’m surprised at the talk of the cost being too high.  For me, the comparison is to a laptop with a cellular broadband internet card — $1440 for a standard two-year contract.  The Kindle 2 doesn’t have a full web browser, but if you’re favoring text-heavy websites (news, blogs, mail, wikis), it’s perfectly sufficient.  Plus, it’s a nice screen and has many-day battery life.  All in all I think it’s a more-than-reasonable price for something that lets me read reddit on the street corner so as to better shout at sheeple about government conspiracies.

The xkcd sysadmin, davean, notoriously scornful of any new technology, took a look at the Kindle 2 after it arrived this afternoon.  He spent several minutes playing with it, discovered he could use it for email in a pinch, confirmed that an ssh terminal could be hacked together using the browser and javascript, and bought one for himself within the hour.

Bottom line, I think it’s a really neat device that fills a niche that nothing else really does.  And seeing the prices, I’ll probably even buy some ebooks and magazines.  But the web access is the real reason I got it.  And if the advertised free access to Wikipedia and other text-y websites is curtailed, (as the Terms and Conditions suggest could happen), the Kindle’s battery life means that I can camp out drunk on the Amazon lawn yelling at Jeff Bezos for quite a while.

I will be in Ireland this Saturday, at SkyCon at the University of Limerick. I’m guessing they found me through my LimerickDB project, which may not be what they think it is. Since they rejected my original idea for a talk, which was based on John Cage’s 4’33″ I will be talking about doing internet comics for a living, and related subjects such as frozen pizzas and MarioKart.

My apologies to everyone who’s extended a speaking/convention invitation who I didn’t get back to — the answer is usually ‘no, thank you’, but I’m too meek to say so (which is probably part of why my email piles up so badly). I love traveling and talking to people all day, but it’s so exhausting that I only do a couple of them a year.

Courtesy LispClub.com:

Sue and Bob take turns rolling a 6-sided die. Once either person rolls a 6, the game is over. Sue rolls first. If she doesn’t roll a 6, Bob rolls the die; if he doesn’t roll a 6, Sue rolls again. They continue taking turns until one of them rolls a 6.

Bob rolls a 6 before Sue.

What is the probability Bob rolled the 6 on his second turn?

The answer is not 5/36.

I love puzzles which are simple to state but have a fiendishly tricky or counterintuitive answer. I just threw up a page on the xkcd IRC wiki to hold some of the better ones I’ve found over the years. I’ll be adding more over the next few weeks as I remember or find good ones. Feel free to add some of your own!

Edit: Buttons and then Daniel Barkalow got the correct answer first.  Here it is, rot13‘d.  Check your answer against this before posting smugly or people (I) will tease you:  gjb friragl svir bire gjryir avargl fvk, be nobhg gjragl-bar cbvag bar creprag.

(Warning: boring nerd rage ahead)

I need help finding a new laptop.

For the last couple years, I’ve have a Fujitsu P1610 (which replaced my stolen P1510). It should last me for another year or two, but I’m pretty hard on my computers. This one has started developing problems typical of an overused, aging machine — your standard hardware failures, screen damage, random crashes, unexpected hangings, paradoxical package syndrome, retrograde memory leakage, Hamitic keyboard, fan tearing, Indian screen burn-in, dysfunctional Haskell, LCD torsion, tab key thrombosis, trackpad sclerosis, dry RAM syndrome, scroll-wheel arousal, systemic rainbow failure, peristeronic disk fragmentation, hyperactive tilde key, heat-sink apricity, kernel duplicity, space bar plasticity, and cross-site chlamydia.

So I know that in the next year or two, I’ll be in the market for a new laptop. Here’s my problem: as far as I can tell, nobody makes the laptop I want. I really only have two criteria:

1. Weight under 2.5 lbs
2. Screen at least 1280 pixels wide

This isn’t an impossible combination. My current laptop fits it. The Fujitsu Q2010 fit it. The Fujitsu U820 fits it but is actually too small and light (the size of a CD case, with a 270 dpi screen!). The problem doesn’t seem to be technology. And, given the recent popularity of netbooks, it’s probably not lack of demand for small laptops, either. Everyone just seems to assume that if you want small, you also want a cheap and crappy display. Is there nothing between the netbooks and the Lenovo X series (which are a bit bulkier than I want)? Other than buying a laptop that is several years old (hunting down a Q2010) or has a touch screen that I don’t really want (the newer model of my current one, the P1630), do I have any options?

(Side note: I find it odd that there are so few laptops with high-DPI screens (>175)? Everything above 150 on that Wikipedia list is either a Fujitsu or a smartphone. Aren’t we supposed to be moving out of the era of pixels by now?)

I went sledding in Danehy Park in Cambridge recently, with my brother and some friends (including Mike, one of the other Boston-area people with a ball pit).  The snow was packed, icy, and awfully slick, and we were wondering just how fast the sleds were going at the bottom.

When you slide down something with no friction, your speed doesn’t depend on the path you take — just on how far you fall.  There are a number of simple equations derived from F=ma that are handy to memorize.  One of them gives the speed of an object after it’s fallen height “h” in Earth’s gravity:

In this equation, “h” is in meters and the answer is in m/s.  It’s actually 2*a*h, where a is the acceleration of gravity.  We round 9.8 m/s^2 up to 10.  (Other handy ones are that the time to fall that distance is sqrt(h/5) and the maximum range of a thrown projectile (45-degree angle) is v^2/10.)

This formula tells you that if your car nosedives off a 50-meter bridge (about double eastern US treetop height) you’ll be going about 30 m/s (interstate speed) when you hit the ground, making the crash the equivalent of hitting a concrete wall at highway speed.  It also tells you that if a (purely gravity-based) roller coaster’s highest cumulative drop, top to bottom, is 35 m (a typical large coaster), it can’t go faster than 26 m/s (which is roughly the old speed limit of US interstates).

I eyeballed the height of the hill to be about 11 meters, since I was about eye-level with the top windows of nearby three-story houses (Google Earth later verified this). So, the theoretical maximum sledding speed in Danehy park is sqrt(20*11), or about 15 m/s. In practice, because of friction, it will be lower (interesting note: the ratio of the vertical to horizontal distance the sled travels is roughly the coefficient of friction of the sled on the snow.)

Checking with our handy table, we see that 15 m/s is faster than the fastest sprinter, about the speed of a cat or rabbit (but — critically — slightly slower than a raptor), and not near highway speed.  We got the GPS from the car and did a few runs with it, recording the maximum speed each time.  It was a pretty reliable 10 m/s (11 if we pushed), which is a lot faster than running speed for everyone except Usain Bolt.

So, in every state except Wyoming, North Dakota, South Dakota, and Maine, there is a building high enough that Marty McFly could have taken the innards of his DeLorean up a freight elevator and acheived the required 88 mph (40 m/s) by jumping off the roof.  (Because of air resistance, I wouldn’t try it in NH/MT/ID/WV/AK/VT, either.) He just needs to leave a note at the bottom explaining that in 30 years they should set out a trampoline.

Edit: Sorry for the brief downtime today.  Also, to everyone posting that the GPS will only give horizontal speed and will underestimate speed on the slope: the sled reaches top speed near the bottom.  It only starts to decellerate when the grade of the slope is less than the sled’s coefficient of friction (plus a bit for air resistance), which seems to be less than 0.10 (people have trouble eyeballing slope grades, which are almost always shallower than you’d guess).  Since sin(x) ~= x for small x, this correction (1/cos(grade)) comes out to at most a percent or two in reality.  However, if the fastest part of the slope looked like the one in the second drawing, it would indeed be a big correction.