Jeff's Home Page
Jeff's Resume/Bio
Photo Gallery
Contra Home
About Contra
Contra on LiveJournal
Diary Topic Index
Diary Photo Index

Link to an Entry



Sam'l Bassett
David Beers
M. Covington Loren Heiny
Harry Helms
Jim Mishel
Bill Roper
David Stafford
Scott Smith


2006

2005

2004

December 2004
November 2004
October 2004
September 2004
August 2004
July 2004
June 2004
May 2004
April 2004
March 2004
February 2004
January 2004

2003

 

January 31, 2005: Review of The Great Influenza, Part 1

I usually start reading a book explicitly intending to learn something, and one of the finest pleasures in reading for pleasure is learning more than I bargained for. So it happened this time, with a truly excellent book by John M. Barry, The Great Influenza. This is going to take me a few days, so bear with me.

To begin: I've always been mystified by World War I, a war that has never made sense to me. (I think that Europe was simply ready for another war. War was considered noble then, especially by people who didn't have to fight.) What mystifies me even more is how little I've seen and read in my life about the Great Influenza Pandemic of 1918, which marched with the forces of WWI, infected a third of humanity, and killed five percent of it.

Five percent of all humanity. 100,000,000 people. Now that would make an impression, right?

Wrong. I never heard of it in any American or world history course I ever took, in high school or college. I don't think I ever heard of it at all until I started reading history aggressively in my 40s. Basically, folks, we almost lost this one. If the virus had held out a little longer, most of what we call society might have collapsed, and entire races—especially aboriginal peoples—might have vanished entirely from the Earth, and humanity reduced by half. We got lucky. I wonder if we will always be that lucky.

The Great Influenza was long called Spanish Influenza, but while it hammered Spain, it began right here, in a rural county in Kansas, in the spring of 1918. From there it marched with American troops to Europe, and eventually to every corner of the world, killing mostly the young and the vigorous, and killing them in what was often as little as twelve or fourteen hours after symptoms appeared. Nothing like it had ever been seen before. Barry describes, with terrifying vividness, life during the roughly eight weeks that the plague did its worst work. It reminded me of how Barbara Tuchman described the Black Plague's evisceration of Europe in the 1300s: No room in hospitals, corpses hauled through the streets of Philadelphia stacked like cordwood in horse-drawn carts to be buried in anonymous pits, corpses stuffed in closets or in crawlspaces because there was no place and no one to bury them, young children starving to death in shut-tight apartments because both parents had died before either could summon help—not that help was likely to come. In the US, 675,000 people died, and among First World nations we did about the best, probably because most Americans had had flu at some point in their lives and had developed some immunity to that class of viruses. In Alaska and Labrador, where flu was a rarity, no one had any immunity, and death rates in some larger villages was 60%, and numerous smaller villages of 100 people or so were simply wiped out.

How could something like this possibly have been treated so lightly by contemporary writers? Why doesn't every schoolchild know about The Great Influenza? Barry has some suggested explanations, and they're ugly. More tomorrow.

January 30, 2005: Amazon's Search Inside the Book

As I'm sure almost everyone has noticed by now, many if not most books listed for sale by Amazon have a feature called "search inside the book." You can literally search the entire book, and scan two pages forward or back from a search hit. I've used this feature several times, especially looking for odd names in the Tolkien canon, or asking myself questions like, What happened to Alatar and Pallando? I used it to find the original context of the mysterious neologism "e nagua," from I Never Promised You a Rose Garden. Amazon created the feature to try and emulate the "flip through the book" experience at B&M bookstores, where people read isolated passages to size up the book, but can't just stand there and read the whole thing.

It's a good and useful idea, and for research at least, it works well—at least when you know what books contain the keywords you're looking for. Searching for keywords across titles is made more difficult by the fact that you have to scroll to the end of the title/author search results list, and click on the "Click here to see additional results" link. For "e nagua" this was easy. For "Brunel" you have to scroll past 264 citations to get to the search results for searching inside titles. One wonders if this was deliberate or not. Certainly, being able to jump to the end of the title/author results, or just jump to the "Search Inside" results, would be a useful enhancement to Amazon's system.

For sales, well, it's less clear. Interestingly, most of the books I've searched so far were books I already have and had read, and was looking for interesting passages that I had forgotten to mark. I can't think of any book that I've bought because I could read scattered pages inside it. Keith and I have limited Paraglyph's participation in the program to only a few titles, waiting to see if we can detect any effect on sales. We haven't seen any, one way or another. It's hard to do any serious testing, without the ability to establish a "control group" for comparison. Each book is its own group, and sales of one book say only a little about sales of another book. Paraglyph doesn't have enough books in print to weigh one or two titles against the rest and get meaningful results.

However, Amazon as a whole does. In this interview, Jeff Bezos says that books participating in the Search Inside program have increased their sales by 9 percent over sales of non-participating titles:

You launched Search Inside the Book about a year ago. What have been the main effects?

If you went into a physical bookstore, and all the books were shrink-wrapped shut, would you sell more that way? Probably not. But for the first eight years, that's what Amazon.com was. Now there are hundreds of thousands of books [that can be full-text searched]. Sales of those books are up 9 percent relative to others. We wondered about things like cookbooks and reference titles - would people just take the snippet they need and not buy the book? In fact, by letting people search inside, sales of these types of books have gone up more than average.

9 percent isn't riches, but it's certainly an indicator that more people are using the system than abusing it. I marvel at the cleverness of these guys, and I often wonder what they're going to come up with next.

January 29, 2005: Networking Odd lots

Just finished the wired networking chapter this morning. It was some serious work, but fun—I learned a lot of new things about UTP cabling, and a couple of things about the shortcomings of various versions of Windows. So it's probably time to dump a few of them as odd lots:

  • Only Windows XP can do IP over 1394. Windows ME, maybe, if the wind's from the south and your lucky penny's in your pocket. Win2K, nyet.
  • The latest releases of LapLink Filemover don't work on Windows 95. Now hang on here...what version of Windows is likely to be the one most in need of serial port file transfer?
  • This one's important, and may rank as the coolest and least suspected method of frying your PC: Running an Ethernet cable (copper, not fiber) between two houses, each with its own electrical service and ground. For various reasons, the ground potential of every house, even those built side by side, is a little different. Think of the line from the ground rod up into the house as a resistor between earth ground and the electrical system ground; basically, the breaker box. The ground potential is the voltage drop across that resistor. If you toss a cable to the neighbors and connect both ends to a PC, it's very likely that a DC current will flow through one or more of the conductors, and NICs are not designed to handle that sort of current. They'll fry. Maybe not with smoke and flame, but they will die and not come back. So don't run cable. Use WiFi.
  • I found a site called Port Forward, which is devoted to helping people configure consumer-class Ethernet routers. There is an immense amount of information on this site, for every router I've ever heard of and a couple hundred more. If you don't remember the default IP for a given brand of router, it's here. If you're clueless about port forwarding and firewall configuring, it's here. Definitely worth bookmarking.
  • Not network-related, but interesting nonetheless: The Antec Sonata case has something I've never seen before: A standard disk-drive power connector on the back panel of the case, accessible even when the case is shut up tight. I'm not sure when this might be useful, but it's one of those things that when you need it, you probably need it bad.

January 27, 2005: Testing the CAT5E in My New House

I don't know why it took so long to do this, but I finally tested my CAT5E in-wall network cabling here at the new house for throughput. I used Ixia's QCheck utility, with endpoints on all the PCs I have here. All my machines are running either Win2K or XP, and all my network ports are 100Base-T. I'm more than a little puzzled by two results from the tests:

  • My network cabling must be really good. On the link between my two fastest machines, I'm getting throughput readings as high as 94 Mbps—on a 100Mbps network. Now, admitting that this is my first experience with CAT5E, I thought that the housekeeping burden with TCP would be higher than 6% of the bit rate. I was expecting nothing higher than about 75 Mbps. On the other hand, I don't have anything in my library that tells me what throughput to look for on a wired Ethernet system, and nothing crisp turned up on the Web. Those who have measured throughput on wired networks, answer me this: What's typical on CAT5 or CAT5E using 100Base-T?
  • Measured throughput under the TCP protocol is much higher than measured throughput under the UDP protocol; at highest, 50 Mbps vs. 90 Mbps. UDP is a simpler protocol than TCP. I'm stumped here. Any theories? Is UDP inherently half-duplex? That might explain it, but I don't know the protocol well enough to be sure.

I measured throughput through my wireless access point as a control, and it was what it always has been: About 4.5 Mbps with encryption enabled. (The client is Wireless-B. The AP is Wireless-G.) It's all the more remarkable because the house is a good size, and some of the cable runs must be close to 70 or 80 feet long, maybe more.

The new mobo comes with 1000Base-T, and the Linksys 5-port gigabit switch is down to $85. I can get gigabit PCI cards for $25, remarkably enough. The cable seems good enough. Maybe it's time to start going gigabit. Several people have told me that gigabit NICs don't swamp the PCI bus. I'm tempted.

January 26, 2005: The New PC's Parts Lineup, So Far

After a good deal of techie soul-searching and crosseyed-scanning-of-geek-sites, Pete and I have nailed all the key issues of our new PCs, and the first parts order has been submitted to NewEgg. Here's what we've decided on:

  • The Antec Sonata case. It's got a rep for quiet, which is the idea for this design.
  • The Intel D865PERLK motherboard. The K board is the top-of-the-line D865PERL mobo, with better integrated sound and gigabit Ethernet. Tom's cites it as comparatively slow (due to conservative memory timing) but very stable, and we like that.
  • The Intel Pentium 4 3.0E Prescott. (Intel p/n BX80546PG3000E.) We could have afforded to buy the 3.2 GHz Prescott, but after staring at the numbers for far too long I became unsure whether we could afford the heat.
  • The ATI Radeon 9600SE video board. I do want to play with digital video and HDTV, but the ATI All-In-Wonder (my original choice) ties me to a TV tuner at a time when that stuff is evolving much too quickly. So we chose a relatively low-performance graphics board (we're not gamers!) without a fan, to keep things cool and (more important) quiet—and when I get time, I'll buy a separate TV tuner board.
  • The Seagate Barracuda ST3200826AS 200 GB NXQ hard drive as the main boot drive. We don't have an NCQ controller on the mobo, but they're available now as PCI cards, and once the prices come down we may buy the PCI NCQ controller to see how it works.
Those are the items that basically define the system, and the ones that Pete and I want to keep in common so we can compare notes if things ever get flaky. We're still thinking about the optical drive, and there are a host of other options for front panel goodies that I'll spend another entry on in the next few days. But we're off and running, and construction should commence shortly after I finish the chapters I'm doing on my current book.

January 25, 2005: The Golden Age of Chicago Children's Television

While sniffing around for evidence the Elmer the Elephant did actually exist on Chicago TV when I was five or six (see my entry for December 31, 2004) I stumbled across a marvelous book: The Golden Age of Chicago Children's Television, by Ted Okuda and Jack Mulqueen. Hoo-boy, what a ride down Memory Lane! Whole chapters on legends like Bozo's Circus, Gigglesnort Hotel, Garfield Goose, Ray Rayner, and Kukla, Fran, and Ollie, all of which began in Chicago, and some of which went national once they hit big in the midwest.

Oh, and a whole chapter on Elmer the Elephant and his human, John Conrad, too!

This book could have been done by a writer digging through library records and nostalgia sites, but Jack Mulqueen was actually there, and knew most of these people, and labored in the kids' TV industry himself for most of his working life. In fact, probably the best parts of the book (once the amazement of seeing stills from these shows for the first time in 45 years wears off) are the descriptions of how kids' TV happened, what the economics were, how the ratings game was played, how network afilliates sparred with the networks for shows and resources, and how shows and people bounced from one station to another. Everybody he quotes says it was fun, but to an outsider it sures sounds like Ulcers Unlimited.

Not surprisingly, Jack says a lot about his own contributions: puppet show The Mulqueens (1963-1965) which I remember clearly on Channel 9 (WGN) and Kiddie A-Go-Go (1965-1970) which was a little late for me, though I watched it once because I'd seen in the TV listings that the New Colony Six was playing on the show. The concept seems surreal today, but based on his ratings Jack had clearly hit a nerve: Kiddie A-Go-Go was a soft-rock dance show (like Shindig and Hullaballoo) for pre-teens. Yup. Five days a week, Jack's beautiful wife Elaine dressed in her distinctive all-diamonds-and-ruff harlequin costume, and taught ten-year-olds to do the Boogaloo, the Monkey, and the Pony to music lip-synced by major bands and performers, including the Four Seasons, the Flamingos, the Left Banke, and Roger Miller, as well as local bands the the Cryan Shames, the New Colony Six, and the Riddles.

For awhile Kiddie-A-Go-Go beat the formidable Garfield Goose in the ratings, but the stations that ran the show (first ABC affiliate WBKB and later UHF independent WCIU) were a little twitchy about hosting little kids dancing what a lot of conservative parents clearly thought of as rowdy or even indecent dances, and the show did not get the support from the stations that it deserved. Still, the book tells a fascinating story of an independent TV producer taking a contrarian idea and running with it.

If you were a kid in Chicago and watching TV between 1950 and 1975, this is a must-have. Good writing, lots of pictures, well-indexed. There's a very thorough list of Chicago kids' shows that includes virtually ever kids' show that was ever created in Chicago, even where little hard information about them survives. 250 pp. $17.95.

January 23, 2005: Serial ATA Disk Drives with NCQ

Pete Albrecht and I continue to work out the design of our new PCs, and over the past few days we've been confronting a constant truth within our industry: You're always on the bleeding edge of one technology or another. The one that's bleeding the most right now is Serial ATA with NCQ (Native Command Queing) which is present in most of the recent Barracuda hard drives fromn Seagate. The drives have NCQ, but the mobo-based controllers don't, except on bleeding edge mobos that we don't trust. PCI-based NCQ controllers are just coming on the market, but we're afraid that they might demand too much bandwidth from the PCI bus. And (of course) the technology is too new for good data—like, how much performance boost could we expect from NCQ over stock SATA? (My guess is not as much as Seagate might wish us to believe.)

Pete and I are probably going to use the Seagate NCQ drives with the standard SATA controllers on the mobo, and if good data convinces us later on that we could pick up 20% or more with NCQ, we'll try the PCI controller. I suspect we won't have to. And as an aside, why don't they build disk drives that allow you to add memory to the drive cache? That would seem to me to be the easiest way to increase drive performance these days. Somehow an 8 MB cache on a 300 GB drive seems, um, stingy. (I'm a big believer in cache.)

Another odd note before I get back to work here: Online retailers catering to the build-your-own-PC crowd are no longer selling P4 CPUs slower than 3GHz. So now 3 GHz is seen by the geek crowd as an entry level processor.

More as the project proceeds.

January 22, 2005: Why Not Fiber in the Walls?

People have asked me from time to time why I didn't just have fiber optic cable put in the walls of the new house here, instead of Category 5E. I did think about it, and think hard, and I did a certain amount of shopping. Fiber optic cable suitable for a gigabit Ethernet network isn't hideously expensive (about $3/foot last time I looked) and the adapters are coming down in price. PCI cards for gigabit Ethernet over fiber are now about $250, and while that's almost 10X what a 100Base-T PCI card goes for, I might be willing to pop for it, but for a couple of issues:

  • I don't have any hard data, but I have a suspicion (based on what I know of the PCI bus and TCP/IP) that a gigabit Ethernet network would chew up a lot of processor cycles, and hog most of the bandwidth available across a typical mobo's PCI bus.
  • I don't have any serious need for the kind of bandwidth fiber would provide. Until we're throwing whole DVD-quality movies around the house, 100Base-T is more than adequate. Backing up data over the network actually doesn't take a lot of bandwidth, once you have all your static data (MP3s, archived stuff) backed up on DVDs and in the safe deposit box. After that it's all very tiny incremental saves. I create what I consider to be a fair amount of material every month as a writer, but when looked at objectively, my entire monthly output could be thrown across a 100Base-T link in less than a second.
  • This, however, is the kicker: I don't know anyone who's ever installed the stuff, and don't have any expertise myself. I found a superb installer to put my CAT 5E cables in the walls here, and got several recommendations and endorsements from his previous customers. He's never done fiber (I asked) and doesn't know anybody who's done it. Let's just say I don't want to pay several grand for fiber to be installed badly, and I don't know how to be sure that a hireling installer would do a good job.

It might be different in another ten years, or maybe even five. By then, gigabit Ethernet over fiber will be routine, and there will be people with the necessary expertise willing to work on residential installations. Until then, I don't think I'd chance it. Say you have a finished fiber optic network, test it, and find it to be a botch job, how do you replace it? Once the drywall is up, stringing new cables in the walls is miserable work, especially when (as with both CAT 5E and fiber) you have to be very careful how you take your turns inside the walls.

I'm completely happy with the network here, and it does everything I need. I think I made the correct decision.

January 21, 2005: Children of LapLink

As part of the chapter I'm currently working on, I'm looking for all the various ways to create a peer-to-peer IP-based connection between two PCs. Windows supports such connections via Ethernet crossover and FireWire right now, and I've been looking for an IP driver that will create the same sort of connection (albeit slower) over standard PC serial ports. I know about SLIP, but I've never seen anything that implements it (or its replacement, PPP) at close range over bare serial ports without intermediation via modem. I recognize that over serial port speeds, an IP-based link may not make sense because of IP's overhead, but if it exists I'd like to fool with it.

USB is an interesting issue too: I already have a $30 USB A-to-A cable package that comes with an older version of PC-Linq, but that's more like Kermit or LapLink in that it's just a file transfer package, not true IP networking. (IP could work over USB with a cable like that—the bandwidth is there—but I've not seen anything in terms of drivers.) In a very similar vein is Windows Direct Cable Connection (DCC), which exists in all versions since Win95, but I've had a great deal of trouble getting it to work properly except between two XP machines—and show me an XP machine that doesn't have an Ethernet port!

During my research on connection technologies, I also stumbled across TOSLINK, which is fiber optic cable for audio signals. Audio! Egad. Hey, for a complete exercise in cognitive dissonance, build yourself a two-tube stereo amplifier from a pair of 6T9s and then feed it signal from an old early-60s Heathkit FM tuner through a TOSLINK cable. Absolute spectral purity on your 60 cycle hum!

January 20, 2005: Hilary's Nemesis in 2008

Not much time to write here today, but having done so well in my psychic predictions for 2004, I'll go even further out on a limb and predict that our next president will almost certainly be a woman. This is good news for America, of course, but the race I foresee is one that will rip the guts out of the Democratic Party and leave them lying in the sun to rot: Hilary Clinton vs....Condoleezza Rice.


January 19, 2005: Poor Piet and His Mis-Googled Grooks

The Net threw me a curve this morning, in the wake of a number of emails (the first was from Bruce Baker) asking me where that "old poem" I quoted yesterday ("someone to love," etc.) actually came from. I recall it being one of Piet Heim's epigrammatic poems that he called "grooks", and went looking for more information on Piet Heim.

Problem is, the guy I was looking for is Piet Hein. However, enough people had mispelled his name on the Web to convince me that my spelling was correct. There were 32 citations of "Piet Heim" on Google, including one that (correctly) attributed the invention of the Soma cube to Piet Heim. A few mentioned his grooks, and 32 citations of an obscure Danish poet were enough not to arouse my suspicions.

Then I looked up "grooks" on Amazon, and, whoops! It's Piet Hein. Suddenly, 32 citations on Google became 114,000—and I realized that Hein was not an obscure Danish poet, but a famous polymath and member of the Danish Resistance, who lived from 1905-1996. A short biography is here. I borrowed a couple of his little books from a friend 25 years or so ago, and enjoyed them, and that's where I recall reading the epigram in question. Of course, it may simply be that something in my memory links Piet Hein with epigrams, so anytime I read something anonymous but epigrammatic, I think of him.

In truth, I have yet to find anyone on the Web attributing the "someone to love..." poem to Hein. I saw a couple of pages claiming it was from Elvis Presley (!!) and another ascribing it to Kenny Rogers. Most people quoted it as an element of ancient wisdom without any explicit source. If any of you can provide me with an attribution, I would much appreciate it.

In the meantime, I gotta go out and learn a little bit more about Piet Hein, and corner a few of his books again. While scouring the Net looking for the "old poem" I cited, I ran across this one of Hein's grooks:

The human spirit sublimates the impulses it thwarts;
a healthy sex life mitigates the lust for other sports.

My kind of guy!

January 18, 2005: Why I Don't Like Sports

I had an odd insight today, and I'm little surprised it took me 52 years to figure it out—not that it's a matter of great importance. I have never liked sports, not even as a kid. This drove my old man half nuts, because he was a great baseball fan and tried to teach me all the usual father-son sports things, and I wanted no part of it. And I'm not even talking about violent sports, which make me ill. I mean, even boring old baseball leaves a bad taste in my mouth, whether it's played in a stadium by millionaires, or in an empty lot by grubby-faced kids.

Now I know why.

I have a strong emotional leaning in a particular direction, and it's a little sappy and hard to put into words. I want everything to be whole, and healed, and running like a comfortable old Chevy engine. I want everyone to have what that goofy old poem claims is all we need: Someone to love, something to do, and something to look forward to. There's always enough to share around, and if we run out of something we can always make more. I believe in civility, responsibility, generosity, and a society in which all systems only run at 75% of capacity and don't overheat. I don't believe in Hell, and I don't believe in zero-sum games. I don't want anyone to lose.

That's it right there. I don't believe that anyone should have to lose. Sports are a form of entertainment consisting of an endless series of zero-sum games. That's fun? No. That sucks. The goal in tennis is to hit the ball where the other guy can't get at it. This is dumb. Why not make the point to hit the ball precisely where the other guy can hit it back, and make it a sort of cooperative ballet among you, your partner, and the ball. I'd watch that in a second, especially if you jump up in the air, spin around twice, and still hit the ball right into the other guy's racket. Carol and I used to stand gut-deep in our Scottsdale swimming pool (which had no deep end) and send this little rubber torpedo thing back and forth, steering for one another's hands and seeing how close we could come. It took some skill and it even gave us a little exercise. (Most of the time we both had to dive for it.)

Keith and I built the largest publishing company in Arizona not by being cutthroat competitors, but by doing good work, and—egads!—cooperating with other publishers every chance we got. We split booths with our competitors, shared costs and promos and lots of other things, bought one another drinks at conferences, laughed and had a good time. We all sold books and we all made a living. The grim-faced drones at Macmillan and other monster presses didn't have any fun. All they wanted to do was win. I'm sure we drove them batshit by not playing their stupid game.

Ok, ok. I'm not so naive as to think that we can actually achieve a society where nobody loses. It's a personal ideal of mine, one to be put into play whenever possible, but also one tempered by a good grip on reality. I'll compete when I must, but it's not what's really fun in life. So sports fail for me, but dancing succeeds. I'll apologize to my father when I get up there and see him again. He didn't understand me, and I didn't understand him, but life is for figuring those things out.

Maybe we'll play catch. Finally.

January 17, 2005: The Forties Never Happened

Michael Covington, in his own Web diary, recently gave us a sage essay (in his January 1, 2005 entry) on something I touched upon some time back, albeit in a kind of half-snotty way: Our habit of dividing our history in ten-year "eras", each with its own identifiable characteristics. I'm with Michael in most respects, with one major difference: There were no Forties. There were the Thirties, there was the War, and then there were the Fifties.

I agree with Michael: What we're reaching for is some sort of handle on changes in our own culture. And it's true that these seem to come in roughly ten-year cycles. My point is that there was no culture we can look at and say "Forties!" unless it's the culture of WWII: "A" cards, rationing, victory gardens, "Is this trip really necessary?" and so on. The music popular during the War was not radically different from the Thirties. Big Band music emerged from jazz in the Thirties, and after peaking in the late 1950s, started down the Long Tail.

The Thirties began in 1929, with the collapse of the prosperity of the Twenties (helped along by economic nationalism and idiotic tariffs) and didn't end until the War forced prosperity back upon a desperate nation. Almost immediately after the War (which occupied the US for not quite four years) half a generation of young men who had delayed beginning their adult lives came back all at once, filled with the sort of resolve, discipline, and moral compass that came of fighting what might well be American history's only truly just war. Massive numbers of new houses began to be built by early 1947, and the culture of The Victorious GI coalesced around affordable detached homes. Michael's contention that TV created the Fifties is partly right, because there were regular TV broadcasts in the big cities as early as 1947. TV certainly helped create demand for consumer goods, which people could now see on the flickering gray screen, see at least well enough to plant seeds of yearning. Prior to TV, advertising on radio was mostly things you needed: Food, cars. After TV, well, advertising was about creating demand, sometimes for things that any sane person would consider ridiculous.

I don't much care for the psychology of the Fifties, but in terms of society, it seemed to work, and work well. The model continued mostly unmolested until the first children of the Victorious GIs got to an age where they began to take over the culture: 1964 is a good start, and it was in full wail by 1967, as Norman Mailer documents in The Armies of the Night. So I could make an argument that the Fifties lasted almost twenty years, from 1946 to 1964. I also think that the Eighties didn't begun until interest rates declined in 1982 or 1983, allowing the economy to get into the high gear that most of us connect with that time period. Apart from the Fifties anomaly, it does seem a little odd how things do seem to fall out into roughly ten-year blocks, just not blocks aligned on the calendar. One would have guessed 25 years, which is the time it takes a generation to come to adulthood...or maybe we're just looking for boundary points in what is really a continuum for our own use.

How would a computer select significant cultural changes? That's one piece of research I would love to see come out of AI. Would it see the same patterns we do? I hope to live long enough to find out.

January 15, 2005: How Apple (and USB) Killed FireWire

Whew. My chapter on USB and FireWire is just about finished, and let me tell you, I learned a lot—not all of which pleased me. I've known for some time that USB is a little kludgy, especially for high-speed operation, and that FireWire 400 is a superior technology at high data rates. FireWire 400 is not quite fast enough to keep up with modern hard drives, and so an external hard drive connected to a PC through a FireWire 400 cable is going to be throughput-limited by the cable.

FireWire 800, however, is another story: In terms of throughput, it leaves USB 2.0 in its dust, and falls midway between 100Base-T and 1000Base-T. It's fast enough to outrun most common hard drives and thus is a reasonable interface for external data storage. (Tom's Hardware has a very nice article on this.) It can even be used as a Datalink/Physical layer under TCP/IP to provide networking faster than 100Base-T, though from what I've read, Windows support for this is lousy.

Much of what displeased me came from this white paper (PDF), from the CEO of WiebeTech, a maker of external storage enclosures for the Mac market. The gist of the paper holds that the cost of FireWire 800 add-in ports and peripherals has been so high as to keep the number of available products down almost in the noise, and a number of Apple fumbles has kept the use of FireWire 800 gear down, even on Mac equipment. In the meantime, the cheap and reliable SATA (Serial ATA) has quietly taken over the high-speed internal drive interface market. (In other words, it's used to connect internal disk drives to motherboard controller logic.) It uses skinny little seven-conductor half-inch wide ribbon cables that are much easier to snake around a case than parallel ATA. SATA now runs at a bit rate of 1.5 Gbps, and the upcoming SATA II spec will push that to 3 Gbps. Significantly, SATA II will include connector and cable specs for external connections, and built-in support for RAIDs.

The author and I came to the same conclusion: This looks like the end of the road for FireWire, which I first heard about in 1993 and have always admired from afar. (It was exclusively a Mac technology until fairly recently.) Apple seems to have inherited the PARC curse: Build super technologies, and then prove unable to dominate any market with them. (Perhaps this only means that poor Xerox has had its revenge upon Apple for "borrowing" much of the Xerox Star's UI back in the early 80s.) Will Windows someday support TCP/IP over SATA II? I'm suspicious of the cable requirements for a 3 Gbps link (that's microwave territory, and faster than the basic RF frequency for Wi-Fi!) but if they can make it work, eek! We're gonna have some fun in the next few networked years!

January 14, 2005: Still More Odd lots

These came in too late to include yesterday, but are too good to let slide:

  • Roy Harvey put me on to NASA's marvelous Earth Observatory site, with a special emphasis on its Natural Hazards section. There's another before-and-after shot here that confirms what many have been saying: That erosion of coastlines was severe in the tsunami, and the very shape of some shallow landforms has been radically altered.
  • Pete Albrecht sent me a pointer to die.net's World Sunlight map, which is a clever conflation of a computer generated map image plus actual satellite images of cloudforms. The same site can also show you a close up of the phases of the Moon.
  • As I write this, ESA's long-journeying Huygens probe has apparently landed safely on Titan and is sending back data. Sometime tomorrow, we may get the first surface shots of a moon other than our own, and the first shots of any world with an atmosphere anything like as dense as Earth's.
  • Many thanks to all who have sent me info on USB. Nuggets are scattered all over the noosphere, but nobody seems interested in pulling them all together. My one poor chapter isn't the whole story, but it makes me wonder if there isn't a good short book on USB from the modestly technical user perspective, rather than the perspective of electrical engineers designing circuit boards. Here's a nice white paper from TI (thanks to Eero Kankia) of particular interest to the Circuit Cellar crowd, who may still be piecing these things together from loose chips. As sparse as it is, however, we're veritably floating in USB information compared to FireWire. There isn't even a FireWire for Dummies book. How's the world to learn anything in that case?

January 13, 2005: More Odd lots

Too busy still for any extensive essay-ing, so let's dump a few more odd lots:

  • Alana Foster Abbott wrote to tell me that she had heard on the radio that the natives of the Maldives Islands (hard-hit by the tsunami) had anticipated the giant waves and fled to high ground without much in the line of casualities, and furthermore were refusing aid from outside—even to the extent of shooting arrows at helicopters trying to land supplies. This sure smells like urban legend to me, but snopes.com hasn't posted anything yet. Unlikely rumors like this run rampant after catastrophes, and I would possibly include the item (see my entry for December 31, 2001) about animals anticipating the waves and escaping as well. If you have any better data, I'd like to hear it.
  • Most people who wrote regarding the pair of before-and-after tsunami photos I linked to yesterday called them legitimate. Send a couple of monster waves over a small spit of sand overbuilt by humans, and a lot of the base will simply wash away. Carol and I watched some amazing home videos on the news the other night, of expensive homes somewhere in Utah collapsing into a gorge where a tranquil creek had become a raging torrent in recent rains. Never live near water—or even a gulch that might, in bad weather, hold water.
  • My high school friend Pete Albrecht actually sued a telemarketer who ignored the federal Do-Not-Call List—and won! The finance company who had called him repeatedly, pushing mortgaging refinancing, ignored his complaints, so Pete sued in small-claims court. A week before the case was set to go before a judge, the company called Pete and offered him a $2000 cash settlement. (Pete gave me permission to quote actual numbers.) His total cost on the case was $95. That company was only one of several in violation, and now that he knows it can be done, Pete intends to go after them all. Now, if only it were that simple to go after spammers...

January 12, 2005: Odd lots

I've been pretty busy the last few days, and the Odd Lots file has been filling up. Perforce:

  • In answer to the USB question I posed in my January 9, 2005 entry: It looks like each USB controller (not each port) gets a full block of USB bandwidth (the size of which depends on the USB port's version) and this block is shared by all ports in that controller's root hub. Thanks to Jim Mischel for confirming that suspicion. So it makes sense to have multiple USB controllers, rather than ports—and for maximum throughput, use only one port in each root hub you have. Bob Halloran pointed me to a nice article on Tom's Hardware that pays special attention to some gnarly problems with USB hubs and how they divide bandwidth among USB 1.1 and 2.0 devices.
  • My VP of Sales Steve Sayre sent me this link, and although it's possible that it's a hoax, I doubt it. I'm puzzled by the fact that the sea's general level looks higher in the after photo than in the before photo. Some analysis of the photos would be helpful, but sheesh, taken at face value, they are riveting. (Don't waait for me to describe them! Go see!)
  • Pete Albrecht sent me a pointer describing an undelete utility for digital photos stored on Flash cards formatted using FAT. Undelete utilities are good to have around (as battle-scarred PC geeks who've been at it since the dawn of time should know) and while I haven't yet had to use one on a Flash card, the clock's ticking.
  • Tim Goss was the first (of several) to send me a pointer to Microsoft's new anti-spyware utility, which supposedly (I haven't had time to test it yet) picks up more bad guys than either Ad-Aware or Spybot S&D. Many are now asking: What's the catch? Have we become so cynical that we can't imagine that there isn't any catch?
  • While researching anti-insomnia drugs, I ran across a useful site that allows you to search for drug interactions. It's interesting in that it includes herbal remedies and potions like valerian, rather than "real" drugs only. There's a selectable option allowing you to test whether a given drug or drugs interacts with alcohol or various foods as well.
  • Finally, I had this notion that one way to make a PC faster is simply to run older software on it. You wanna see Office fly? Run Office 97 on a 3.2 GHz Pentium 4 with plenty of RAM. I call this "underclocking." (You're not adding cycles to the PC; you're squeezing them out of the apps!) I've been doing it quite a bit since 1999 or so. (Contra is created using Dreamweaver 3.0, released in 1999. Works quite briskly on a 1.7 GHz Xeon!) Newer software isn't always better. But you knew that, right? Right?

January 10, 2005: Those Pesky Plus Signs

I've been stung by a plus sign. Again.

It's embarrassing. I knew about this problem. Months or even years ago, I rolled my eyes and laughed at yet another idiotic VHS vs. Betamax style format war: DVD-R vs. DVD+R. Still, I went ahead and installed a DVD+R drive in my Dell Xeon, and then trotted back to Office Max and bought a stack of DVD-R blanks. The "+" drive won't recognize the "-" blanks. Duhhh.

I bought the DVD+R drive because I think it's a technically superior format, but I think that, like Betamax, it will lose to DVD-R in the end. No big deal—the drives are pretty cheap, and when the time comes that no one's using DVD+R anymore, I'll scrap it and go with the majority opinion. My DVD player claims to be able to read DVD+R discs, though I haven't tested it yet. We'll see.

Something like this happened to me once before. In 1979 I had a ram-charged 1 MHz 8080 CP/M system, and when I heard of Pascal/MT+, I ordered it from a phone-order software vendor. When the product arrived, it was labeled Pascal/MT. Ever in a hurry, I opened it up and tried it, and found that several features I'd read about in the review were not present, and some of them—random file I/O chief among them—were the main reasons I wanted the product. I complained to the product's vendor, and they explained that MT+ was the new release, with all the new features. I tried to return the copy of Pascal/MT that I had purchased via mail-order, and the owner of the mail order company not only refused to take it back, he accused me of being a software pirate and hung up on me.

Ultimately, the Pascal/MT+ people took back the older copy and sent me MT+ in its place, and the good will I generated for them in the several subsequent years more than compensated for the small loss they ate by being good guys. As for the mail order vendor, they went out of business a couple of years later, after a reputation for being snarly and unreasonable got around. There's a lesson in that, and (for me) another lesson in the value of small mathematical symbols. Details count.

Let's hope it's another 25 years before I make this mistake again!

January 9, 2005: USB Root Hub Mysteries

I haven't talked about it much here yet, but I'm working on a new book collaboration with Joli Ballew, called Degunking Your PC. Unlike Degunking Windows (on which Joli and I also collaborated) this book is about PC hardware, and involves both cleaning and tuning, at a relatively high level. Nonetheless, to give good advice I need to understand the hardware at a fairly low level, and certain aspects of USB are giving me fits.

For example: On all the PCs I have here, USB ports come in stacked blocks of two, either on the front panel or the back panel, or both. Each of those blocks of two corresponds to a "root hub" and has its own controller. What I want to know is whether the two ports of a root hub are both capable of simultaneous maximum bandwidth. In other words, if it's a USB 1.1 controller capable of 12 Mbps, does each port have access to its own 12 Mbps block of USB bandwidth, or do the two ports in a root hub share that 12 Mbps? It's pretty obvious that non-root hubs (the kind you attach to a root hub via cable to split one port into several) share the bandwidth provided by the single port among all the hub ports. However, it's not obvious whether the root hub controller gives each of the ports in a root hub its own independent bandwidth block.

This matters in certain cases, especially if you have two thumb drives mounted in the two ports belonging to a single root hub, as I generally do here. (See the photo above.) If I copy files from one thumb drive to the other, the file copy operation will take longer if the two drives share a single 12 Mbps block of USB bandwidth, rather than if each drive had its own 12 Mbps channel to use to talk to the other drive. Some manual stopwatch work suggests that the two thumb drives share a single 12 Mbps bandwidth block, but I'd like to be sure. Anybody have any experience here? I find it remarkable how neither of the books I own on USB (USB Serial Bus Architecture by Don Anderson and USB Complete by Jan Axelson) explicitly addresses this question. If all ports attached to a root hub on the controller share a single block of bandwidth, then it would make sense to use one port each of two independent root hubs for simultaneous high-bandwidth tasks. Any insights?

January 8, 2005: More Cycles Are Not the Answer

Settling on a CPU and (by implication) a motherboard has been a sticking point in spec-ing out my new machine, especially since I may be asking for the impossible: A 3 GHz CPU that runs without distracting fan noise. I recall digging through PC Mobos (See my entry for December 17, 2004) some time ago, and thinking, "Gee...haven't we been hovering at 3 GHz for some time now?"

We have. Even if you don't ordinarily follow the links that I provide here in Contra, follow this one. Herb Sutter has really nailed what a lot of us have been suspecting: That the riproaring increases in CPU speed that we've seen since, well, forever, have just about stopped. We reached 3 GHz in early 2003, and here we are, in 2005, and we're still just a little bit beyond that. Bathe your CPU in liquid nitrogen and you might possibly make 4 or 5 GHz, but as a fan-cooled, mainstream, CompUSA sellable black-box design, 3.4 GHz is about it.

So when I ponder dropping back to 2.8 GHz or a little slower, I don't feel so bad now. I can keep my machine quiet at that speed—and not feel like all the noise-numb youngsters with headphones on are getting a whole lot that I'm not. In terms of quiet PCs, 3 GHz is the end of the road for awhile. Maybe forever.

(George Ewing suggested bringing a dryer-vent pipe down from the attic, connecting the fan to the attic end of the pipe, and blowing frigid Colorado winter air down the pipe to keep the PC cool without any nearby noise at all. That'll work now—but what do I do in August when it's 100 degrees in the attic?)

I have a suggestion for programmers that I think Herb Sutter was afraid to pose: Pull all that unnecessary feature-creep crap out of your software! Chasing full pips in PC Magazine's feature comparison charts has nearly been the death of desktop software. Norton Antivirus 2005 is slow because it's loaded with idiotic stuff like a partial firewall...which sounds about as useful to me as a perforated condom. It's a marketing gimmick, and has nothing to do with keeping worms out of the system. Such stuff is everywhere. I intuit that smaller, simpler, more modular software would be easier to adapt to the future imperatives of PC design.

Herb suggested everything else that might be useful. The most important is clearly to optimize code with the goal of keeping the working set (the currently swapped-in memory page belonging to a process) entirely within cache while it's running. I think that when chip designers realize it's become tough to increase the clock speed, cache size will replace clock speed as a design priority, and that's where the transistor budget of future processors will best be spent. Beyond cache, memory is the ugly stepsister in PC design. I made a pitiful Pentium III 550 run much faster (in terms of perceived application response time) by topping out RAM from 128 MB to a full gigabyte. I think there's still a fair amount of performance to be wrung out of memory access; our current RAM systems run much slower than our CPUs.

Whatever the path that chip designers choose, we need to meditate on this truth and come up with a better one: More cycles are not the answer. The day of reckoning is here. It'll be interesting to see what happens, now that we can no longer rely on higher clock rates to make crappy software look good.

January 7, 2005: What's That in My Toothpaste?

Yesterday was the Feast of the Epiphany, and "We Three Kings" has been running through my head, along with James Taylor's brilliant "Home By Another Way." Carol and I were pretending to be one with the Granola Gang out at Wild Oats earlier today, doing some shopping for a few things that you can't get at King Soopers. I get milk there that's special in a way that's worth an entry on its own, but my immediate story concerns toothpaste. While in the toothpaste aisle, I noticed a species of toothpaste that now contains...myrrh. Well, well. Recall the last and not-always-sung verse of "We Three Kings":

Myrrh is mine, its bitter perfume
Breathes a life of gathering doom.
Sorrowing, sighing, bleeding, dying,
Sealed in a stone-cold tomb.

Right. Now they're putting embalming fluid in toothpaste! I don't care if it can be used in treating gum disease, I don't want my breath to smell fresh like...a tomb. My goth days are long past. I'll rely on echinacea, thanks very much.

January 6, 2005: PC Decisions That Have Been Settled

Continuing yesterday's thread, concerning the custom PC I've been designing for some time. The big decision, as I described yesterday, was what I consider Priority One in the system, and that's quiet running. Everything follows from that. I'm still comparing CPUs, but I'm leaning toward a 3.0 GHz Pentium 4. Faster chips are available, but may not run cool enough without aggressive (read here: deafening) cooling technology. I need to do more research here, and that's actually what's holding me up at the moment. Many other decisions have been made:

  • Networking. No contest: It'll be the on-mobo 100Base-T port. I have CAT 5E in the walls here, and I'm happy with the speed with which I can throw files between the work box here in my office and the server downstairs. Why not gigabit Ethernet? I may play with it someday, but few people have meditated on the fact that a software-based TCP/IP stack handling a gigabit physical layer probably takes a gigahertz of dedicated CPU throughput, and that's more than I feel like spending, simply to be able to bring an old document up here in a smaller fraction of a second. Gigabit networking would speed backups a little, but those don't require that I sit here supervising. If I play with gigabit networking at all, it'll be among my three PCs downstairs, all of which are within a few feet of each other. Cabling at gigabit speeds is tricky and touchy, and while it's something I really need to learn about, that's a project for another...year.
  • Video. I'm probably going to choose one of the ATI All-In-Wonder boards, which contain an on-board TV tuner and support software. Horrors! TV! Hey, my thinking runs this way: I have little interest in competitive 3-D gaming, but quite a bit of interest in video, not so much to watch cable TV on my monitor as to edit camcorder files and make DVDs. On the other hand, I have a live cable TV jack right behind the desk here, and it would be useful to bring up The Weather Channel in the corner of my screen for a quick look at the local radar. (Yes, yes, I'll cop to a certain amount of gadget fever here.) The All-In-Wonder 9000 Pro is not a big heat generator, which matters as well. Its 3-D performance is more than adequate for anything I might want to do in graphics.
  • Sound. Most recent motherboards have sound support built in, so I probably won't bother with a separate PCI card. I play MP3s up here on occasion (though not often, and usually when I'm just sorting papers or other junk) and use Skype. That's it. The Intel D865PERL motherboard I'm looking at (try out that cool Java zoom-and-rotate feature at the link!) contains decent on-board sound support, and I don't want to add another heat-generating card to the box if I can avoid it.
  • Memory. That depends heavily on the motherboard. I want to be able to plug 4 GB of RAM into this thing, even if not on Day One, and the Intel D865PERL has sockets enough for that. I'm not trying to squeeze the last drop of performance out of the system, so I won't quibble about what kind of RAM it takes. I'll buy the best RAM the mobo supports and not fret further.
That's all I have room for this morning. More as time (and room) allow.

January 5, 2005: The PC Prime Directive: Silence!

I'm about to begin gathering parts for my new work system here, and it might be time to talk a little about why I chose what I did. There is an incredible richness of choice available these days in building a custom PC—universes ahead of what there was when I last did this. (My last custom system was a Pentium 90; do the math.) The big driver in custom PC tech these days is 3-D gaming, in which you need the biggest, baddest CPU there is, plus a graphics coprocessor hot enough to need its own fan. Not me: My Prime Directive in the new PC design is quiet.

I bought a Dell Dimension Pentium III 550 back in early 1999, and it was my main system until early 2002. It had a fan, but the fan was so faint that you could barely hear it. In 2002 I bought another Dell, a big swaggering Xeon 1.7 GHz, and the damned thing sounds like an idling jet engine. I thought I would get used to it, but I didn't, and I find myself tolerating background noise as I write and program less and less as I grow older. (Part of that may be working at home—I haven't worked anywhere near a noisy cube farm since March 2002, when Coriolis imploded.) The Xeon is a fine box and will become a useful lab machine (especially with a new and bigger hard drive) but it's going to live downstairs in my workshop, where silence isn't a requirement to get things done.

There are specific custom PCs designed to be fanless and thus absolutely silent, but they're not especially fast, and they're also heavy users of custom components that would prevent inevitable upgrades four or five years down the road. I want a PC that I can tweak and change for years to come, and I'll balance that power against a reasonable amount of fan noise.

After much research and several recommendations from friends, I chose the Antec Sonata case. It's about as quiet as my old Dell PIII 550, assuming you don't stick a gonzo graphics card in it. To use the Sonata without adding another fan to the case (the space is there and the fans are available) I may have to limit myself to 3 GHz or below, and that's OK—there's the ancient and venerable dodge of loading the box up with RAM to make it seem faster than it is.

After all, what I do here in my office isn't cycle-intensive: I write, I draw (with Visio), I program (with Delphi 6), and I do email and Web. I'm not a gamer, unless you count games like Snood, which can run on a Pocket PC. I use another performance dodge: I use 2000-era software. Office, Visio, and Delphi are mature technologies. I have yet to see anything to make me want to upgrade past their 2000 versions. Not going to the limits in terms of CPU speed will not crimp my style. The only thing I really need cycles in abundance for is security (firewall and virus.) There will be plenty left over for Word 2000.

Some of the decisions are nonetheless difficult to make and haven't been firmed up yet. Can I go with a Prescott core? Those bleed a lot of heat, and will require a CPU cooler with a fan. There are CPU coolers that control their fans according to case interior temperature, and if you don't pack the box with a lot of stuff, the fan goes so slowly as to generate no noise at all. (So far I've chosen the Zalman Flower 6000, which does this trick.) The point I'm making is that I didn't start by choosing a CPU or a level of performance. I started by choosing near-silence, and working from there.

Other issues are minor, and much easier to spec. More later.

January 4, 2005: Backdoor Prohibition

Late last year, Colorado was the 49th state to change its drunk-driving laws to a 0.08 alcohol blood level, from the longstanding 0.10 level. It was done under Federal mandate: States had to move down to 0.08 or risk losing highway funding. In other states that complied years ago, there was an unintended consequence: Conscientious diners at restaurants chose not to order wine by the bottle at all, because they feared finishing the bottle at the table would raise their blood alcohol to illegal levels. Restauranteurs saw their revenues plunge, and pressure began for "recorking" laws, which allowed diners to take unfinished bottles of wine home in their vehicles without running afoul of universal open container laws.

Carol and I have done this only once since the law was changed, and overall it's a good thing, as I only allow myself one glass with dinner when driving, and thus rarely buy wine at restaurants by the bottle. All the crackpots (like MADD, which, alas, has become a crackpot group in my reckoning) were screaming that rednecks would be swigging from open bottles of wine while driving. As best I can tell, this hasn't happened even once, and it wouldn't matter anyway, since the bottle has to be recorked with the cork pushed in beyond "finger grip" level, and then sealed in a bag by the restaurant and left untouched outside the reach of the front seat. (We buried our bottle in the little bungee cord compartment over the right rear wheel, not for concealment but as proof that would couldn't get at it.) Swigging rednecks would be in every bit as much violation of open container laws as they were before the change. But some groups (like MADD or the ACLU) are now basically ideological in nature—and often virulently partisan—and don't want to hear anything that challenges their ideologies.

I object to the lowering of blood alcohol levels to 0.08 for a fundamental reason: It's lazy law enforcement. A blood level of 0.08 doesn't guarantee drunk driving (large people can handle more alcohol without impairment than slight people) and it's way too easy for prosecutors to get a conviction without any actual eyewitness accusation of unsafe driving. This is back-door prohibition, and the back door is getting wider: MADD clearly wants legislation banning driving with any measurable blood alcohol at all. And back-door prohibition is one facet of a much larger question of how much freedom we have to surrender to prevent wrongdoing ahead of the fact. It's really no different from the issues of "driving while Black" or "flying while Arab." We could reduce the threat of air terrorism by simply forbidding any Arabic or Arabic-looking person from flying, but where's the justice in that? As much as I hate tobacco (which killed my father horribly) I would object strenuously to making tobacco illegal. I don't gamble, but I think gambling should be legal, where it can be watched, rather than back in the alley where Stag-O-Lee famously and fatally caught Billy cheating. Many new subdivisions now forbid CB and amateur radio in their covenants even when used completely indoors (including all antennas) as a preventive measure against "interference."

Freedom entails a certain amount of risk, and I fear that we're laying the foundations of a legal culture that forbids activities that can be pursued legally without damage to others, "just to be safe." This worries me considerably more than terrorism, which is now the universal excuse for prior restraint. There's plenty of room here for discussion (perhaps gamblers and wine taker-homers should be licensed based on their past records) but the sad part is that I don't see the discussion going on anywhere. We're clearly a lot farther down that very slippery slope than we think.

January 3, 2005: Virtual Tahiti in East Germany

Twenty or thirty years ago, there was a lot of misbegotten enthusiasm for "arcologies": Monstrous city-sized apartment blocks with occasional trees in pots somewhere to keep them from looking so much like prisons. Niven and Pournelle did a terrible novel about the concept, the name of which I don't recall, as when I threw it at the wall it went right through and landed somewhere in Kazakhstan.

Several people sent me an item that was cited on Slashdot, about a new wrinkle on the arcology idea: An enormous zeppelin hangar outside of Berlin that's been converted into a sort of virtual Tahiti, with jungles, sand beaches, lagoons, waves, restaurants, and casino-style musical review shows. Heated in colder months and open 24/7/365, it's a slice of summer set in the dreary nowheres of what used to be East Germany, and while the Germans are taking to it rather slowly, it strikes me as the best possible use to make of an abandoned zeppelin hangar. One article on the...um..whatever category it falls in (which the Germans unimaginitively call Tropical Islands) cited an older chap who wished it was a nudist resort. Nudism is apparently a Real Big Thing in Germany (it's almost extinct in the US) and one has to wonder if market forces will push Tropical Islands in that direction eventually. Pets are not allowed, but Tropical Islands does have free Wi-Fi. No hotel rooms inside the arcology just yet, but they're planning on renting igloos (that's really the word they used!) for sleeping on the large sand beach.

The zeppelin hanger came about as part of an aborted project aimed at hauling cargo around Europe in enormous helium (whew!) filled dirigibles. Every so often somebody suggests this, and money is sometimes spent on it. It's a terrible idea, but clearly, fooling around with the leftovers has been lots of fun. Where else can you find an enclosure big enough to build an arcology in for ten cents on the dollar?

January 2, 2005: Review of Building the Perfect PC

Perhaps the best single book on building your own PC from loose parts came to me about a month ago: Building the Perfect PC by Robert Bruce Thompson and Barbara Fritchman Thompson. I came to it fairly late in the design process of my own new custom PC, but it was uncanny how closely the book tracked my own needs.

The book's approach is superb: After a few chapters on the basics of building PCs and buying parts, the authors describe five different PCs: A mainstream system, a SOHO server, a "kick-ass LAN party PC", a home theater PC, and a small form factor PC. The issues for parts choice are carefully explained, and there are step-by-step instructions on assembling each one, with some of the best photos I've ever seen in a book like this, all of them in color.

Having been up and down and around the block on what I want here on my working desk, what I came up with is a great deal like their "mainstream PC." I was targeting a 3 GHz P4, in as nearly silent a package as possible. I had already chosen the very quiet Antec Sonata case, based on strong recommendations from several people I know, and the book pointed me at the Zalman Flower HSF, a CPU cooler that runs almost silently. I learned a number of things about high-performance disk drives that I didn't already know, ditto RAIDs. I may someday want to build a media server, and I learned a lot reading the section on the home theater PC, though I had to tuck it all in long term memory for future reference. (The next PC I build will almost certainly be a high-performance lab machine, perhaps a multi-CPU box. Such a machine will not be anything like quiet, so it'll have to be exiled downstairs.)

The book is probably most valuable for people who don't want to spend the time studying every PC component technology to the extent that they could confidently spec their own custom system. It would be quite easy to just make a "Chinese copy" (as we used to say) of one of the designs presented here, and the mainstream PC and small form factor PC look very effective.

The book is beautifully written, clearly laid out, and probably the most useful of the small pile of PC hardware books I've been accumulating and poring over this past year. Very highly recommended.

January 1, 2005: How Well Did I Call the Events of 2004?

Happy New Year! It wasn't until the Google era that I ever bothered to research what "auld lang syne" actually meant. (Some questions aren't really worth asking until it becomes really easy to ask them...) It's an old Scots expression (Robert Burns wrote the song, after all!) meaning "days long past"—though that was clear from context.

Anyway. Busy day ahead of me here, so let's get down to my predictions for 2004, which can be found in my January 1, 2004 entry. Don't bother clicking; I'll reproduce them here, for laughs if nothing else:

  • Jeff and Carol will move into their new house sometime during the month of February. (I had to make sure I'd get at least one right!) MISS: We moved in on Saint Patty's Day. Moral: Never assume victory. There are no sure things.
  • Assuming Howard Dean wins the Democratic nomination, George Bush will re-take the presidency handily this fall. Tougher call against Lieberman, especially with Nader sidelined. HIT: Sometimes a psychic can be right for all the wrong reasons.
  • The SCO lawsuit will collapse, and though no legal points on the GPL are likely to be decided, I think no other company will ever be so bold again as to try something this legally insane. MUDDY MISS: The lawsuit is still—barely—alive and flopping around, like a fish on a beach after a tsunami. Should die soon. We hope. (Word to the wise: Don't run down to the beach to try and grab that fish!)
  • As we continue to hand power in Iraq back to the Iraqis, the place will quiet down and Bush will claim victory. Troops will remain, though casualties will plummet. MISS. Sigh. I guess I'm the King of Wishful Thinking.
  • Darknets will become the issue in the online arena, as file trading basically vanishes into heavily encrypted and authenticated private networks. BIG HIT: See this article on Wired.
  • Wireless-G will basically sweep Wireless-B and Wireless-A into the dustbin of history. Whether or not consumers truly understand what "megabits per second" are, they've been well-trained for decades that more is better. HIT. Everybody I know has already gone to Wireless-G, except those who know that Wireless-B is faster than any known broadband connection.
  • Paid public hotspots will go into decline, and free public hotspots as promotional value-added for restaurants and bookstores will rise sharply. See Panera Bread as an early major example. SOFT HIT: Free hotspots are booming, but in hotels—to reel in business travelers—and not bookstores.
  • Pope John Paul II will still not die. HIT: The old guy is amazing. He wants to break the record for longest reining Pope, but to do that he'll have to hold out until 2013.
  • The Episcopal Church will not split. HIT: The malcontents can't take the buildings with them, and don't want to lose their pensions. Sooner or later their congregations will confront the whining doofs, socks in hand.
  • Kite aerial photography will become The Next Big Thing. That, or maybe curling. UNDEFINED: 2004 handed us no Next Big Thing. Sorry, curling fans.
So let's tote it all up: Six hits, three misses, one pass. Not bad for a liberal arts major, huh? Of course, it helps to choose your predictions. As for this year's preductions, get real. My fondest wish is that nothing "interesting" will happen 2005. We all need a rest. Good luck, have fun, and don't forget to sleep once in a while!