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



Pete Albrecht
Sam'l Bassett
David Beers
M. Covington
Loren Heiny
Harry Helms
Jim Mischel
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

 

July 31, 2005: 36 Years Without Being Rash

Carol and I met 36 years ago today (see the story of our meeting in my July 31, 2004 entry) and we had planned a low-key private celebration with one another: a nice dinner at home, and the long-awaited ceremonial opening of the bottle of Dornfelder Rotwein 1994 we have had in the wine rack for ten years now. We were in Chicago last year and could not open the bottle as planned, so this year will do. Who's in a hurry? Part of the trick of creating a lasting relationship is patience; patience with one another's peculiarities (of which I had many in 1969) and patience with the way the relationship evolves. Carol and I worked at getting along. If we seemed to have made it look "easy," that's only because we're relatively private people and don't air our difficulties in public.

Those difficulties are now mostly gone. The mills of patient love (like those of God) grind slowly, but they grind exceedingly small.

Well. However small love grinds things, they don't always work out as planned. Late Thursday afternoon, right before supper, Carol and I were weeding the boulder terrace to the north of the house. I pulled a thistle out by the roots, stumbled, and fell over backwards into a patch of tall weeds at the bottom of our little hill. I banged up my right arm a little on the bounder but was otherwise unhurt. We went in, I dabbed some Bactine on the scrape spot on my arm, and popped a couple of Aleves.

Friday morning I noticed a red patch on my arm. It grew during the day, as another formed on my left wrist, and a few more on top of my head. By bedtime my arm looked like a war zone, with another rising on the small of my back. My arm had started to itch viciously, as did the other locations during the course of the night. Today the itch got bad enough so that we canceled our romantic dinner at home, and went to visit friends. The weeds had looked innocuous enough, but just my luck to do a full bodyflop into a patch of poison ivy camoflaged by tall grass.

So it goes. I suspect I'm going to have to go get myself looked at tomorrow, as new spots are showing up on my face and legs and everywhere else. Our romantic anniversary evening can wait until next July, as can the 1994 Dornfelder. We've been together 36 years. Another year won't hurt—even if it itches in spots.

July 30, 2005: It's a Planet! It's a Comet! It's Even More Planets!

Oi veh. It's raining new planets in here. Yesterday morning I first heard about object 2003 EL61, a Trans-Neptuanian Object in our solar system's Kupier Belt: the dustbin beyond the major planets where debris from the creation of the Solar System gathers. It's big—perhaps bigger than Pluto—and may have a large moon. Pete Albrecht crunched some numbers and charted its path in the sky for the upcoming year. Because of its distance—51 AU, 20% farther out than Pluto—it doesn't move very quickly. Pete's planning on photographing it with his 12" Meade, but due to the object's faintness at 17th magnitude, it will take some practice, and might have to wait until EL61 can be seen high in the sky in full darkness, which may not be until early winter.

But heh—a few hours later, yet another discovery was announced, of yet another planet-sized body out "where God lost his shoes," as country people say. Object 2003 UB313 is even larger, almost certainly larger than Pluto, and much farther out: 97 AU, which is almost three times Pluto's distance from the Sun, and at a 44° angle from the plane of the Solar System, where planets aren't supposed to be.

Note that these were announcements, not discoveries. Both new worlds (people are getting into fistfights over whether to call them planets) were discovered in 2003, and their discoverers were holding out for more data to help pin down their orbits, albedos, and so on. There is yet a third large body discovered earlier this year, 2005 FY9, about which little has been published.

Peculiar circumstances forced announcement of 2003 UB313: People were reading the logs of the discovery team's telescope on the Web, and had begun to wonder what they were looking at, since no other notable object is at that precise location. So other people began looking, and eventually, the extremely faint object was photographed. Pertinent stories are here, here, here, and here. If it all sounds confusing, you're right. I haven't entirely sorted out the research politics yet, and we may not get the full story for awhile.

Still, it's legal to speculate about the objects themselves. What I think we have here are immense comets, so big that the inner planets' gravitational fields have not perturbed them onto Sun-grazing paths. They are a matrix of gravel and boulders embedded in frozen gas rather than rockballs like Mars or gas giants like Jupiter. The rocks may have gravitationally migrated to the objects' centers, and one wonders if the gases have frozen out in layers by density, as Larry Niven speculated in his 1966 novel, World of Ptaavs. We may wonder for awhile; just getting there would probably take 100 years using current technology. In the meantime, does anybody else have any new planets to announce? The line forms here. No pushing. You'll all get your chance.

July 29, 2005: Embedded Database Engines and DLL Hell

(Read yesterday's entry if you haven't already.) Several people have written to suggest database engines for Aardblog, and a few asked a very good question: Why not just install MySQL on the client, and thus use MySQL for both the client database and the server that mirrors it?

Answer: I want the database engine embedded in the Aardblog application. I'll use a separately installed engine if I really have to, but I have a peculiarly intense animus against cutting an application into several independent blocks of code. To do so is the road to hell...DLL hell, specifically.

I have a strong bias toward single-block applications. Outside of standard OS calls, I want everything that the app does to exist within one monolithic .EXE file. The reason: Otherwise, there is no way to guarantee that the code library environment under which an application was compiled and tested will be identical to the code library environment under which the application is run.

Ok. Example time. Suppose you create the DogMatic kennel manager to use the MyDataBox database engine as a client-side database. DogMatic V2.0 was developed using MyDataBox V3.5. As long as DogMatic 2.0 and MyDataBox 3.5 are both installed, everything's cozy and works.

Alas, the CatHouse cattery manager was written to use MyDataBox V3.55. If Roscoe's Puppies & Kittens installs CatHouse after DogMatic, MyDataBox will overwrite its 3.5 release with its 3.55 release. Supposedly that's OK, because there are no API changes between 3.5 and 3.55. MyDataBox does not require separate installation of 3.5 and 3.55.

No API changes. Sure. However, My DataBox V3.55 changed certain under-the-covers memory caching techniques, supposedly "transparently" but with unanticipated consequences. DogMatic works with MyDataBox V3.55...for awhile. Then something weird happens, memory management burps, and MyDataBox writes a corrupted buffer to disk. Abruptly, poor Roscoe loses everything he has on the puppy side of his business.

If MyDataBox had been embedded in the DogMatic .EXE file, this could not have happened. If application and database engine are inextricably glued together, DogMatic can only use MyDataBox V3.5. Sure, you could argue that if MyDataBox's developers had really ensured binary compatibility between their 3.5 release and their 3.55 release, this wouldn't have happened. Sure, and if we had a time machine we could go back and strangle Adolph Hitler in his cradle. Get real, people!

A great deal of Windows' flakiness is due to unrelated chunks of code heedlessly calling one another whether or not the APIs involved are really identical...or just mostly identical. Syntax matters, but so do sematics. Whether or not the calling conventions and parameters are identical, what is done with the parameters matters crucially. Designer/programmer assumptions matter. All kinds of things matter, and matter in ways that we can't anticipate up front. The only way to shovel all these problems out the door is to minimize the circumstances under which one independent block of code calls another. OS calls probably can't be avoided, but there is no ego-free reason for an application to call anything that isn't a highly standard and well-understood component of the OS.

Back in 1999 I bought DBISAM, an embedded database engine for Delphi, and have used it with great success ever since. Alas, it costs $250. What I want is a simple embedded engine based on the VCL. It doesn't have to be open source, though I'd like something that doesn't have to be paid for. Some readers have sent in some suggestions, and I'm looking into all of them. Will report back in upcoming days.

July 28, 2005: Thoughts on Aardblog

Aardblog is still a live project here, though everything I've done on it so far is conceptual: Database schemas, UI sketches, etc. I made a decision the other day that's pretty contrarian: All content will be maintained in a local database, by a local Win32 client. The server-side MySQL database is literally a mirror. When a new entry is created, or an old entry modified, it's stored locally, and the local entry is marked as "dirty," so it will be uploaded at the next connection to the server-side database. Basically, all content in the blog will be present both on the local system and on the remote database.

A lot of server-side freaks and going to roll eyes at this—duplicate data! wasted space!—but I have strong reasons: I make a fair amount of money writing, and the first thing I always ask is: Who controls my content? Where does it live? Can I get it out of a server or file format once it's in?

There are other issues as well. What happens to all your entries if HypotheticalBlogHost.com goes under and the server goes down? If you don't have it stored locally, it's gone. Could some company purchase a blog hosting site and claim ownership of all the content under some weird small print or interpretation of small print? Thanks, but I want my words and pictures to be right here under my own roof and in my safe deposit box. I may not be able to keep some kleptocorp from using my content under a questionable rights agreement, but I definitely want to have it all in my own two hands, irrespective of where else it may be.

The wasted space argument is bogus. ContraPositive is one of the oldest blogs on the Web, and everything it includes—text, photos, almost 2,000 entries posted since 1998—doesn't even crack 50 MB. I have half a terabyte of disk here here locally. 50 MB of local storage is nothing.

And beyond all that, Web-served content editing sucks. Period.

There are two coding tasks:

  1. A win32 client in Delphi, which handles all editing and stores content in a local database. On command, it connects to the remote server and uploads whatever doesn't already exist on the server.
  2. A server-side PHP program that serves up the blog to the public at large by pulling data out of MySQL and formatting it for delivery to a generic Web browser.

#1 will not be too tough; it's really a pretty simple database app, and I've known SQL for fifteen years or so. The queries that allow selective division of the content database into topic-specific blogs is the easiest part of it, in fact. (I'm amazed at how tricky it is to connect to a remote SQL database from Delphi. Geez, guys!) The PHP app will be tougher, as I have no experience with PHP and will have to get up to speed. However, I have a local server here in the basement with PHP/MySQL/Apache already installed, configured and ready to go. The rest is just practice.

I'm currently entering some entries (real entries, if short ones, from back in 1998) into the MySQL database through PHPMyAdmin so I can begin putting some Delphi and PHP test code together. PHP first; I'll feel like I've achieved a major victory if I can serve up some entries in a simple format.

I don't want to be in the software business, and ideally I'd like to turn this loose as an open-source project on SourceForge, using all free components, starting with the formidable Turbo Power Orpheus suite. The kicker there is the local database. I hate to say it, but I don't like FlashFiler very much, and I'm not sure what other free relational database engines are available. That's another research item.

I'll keep you informed as the project evolves.

July 27, 2005: Virtual Zones in a Hypervised OS

If I were to design a new OS from scratch (not that I would be capable of anything beyond the highest-level concepts) I think I would make heavy use of the new hardware virtualization features to be present in the next generations of both Intel and AMD CPUs.

What I would do is pretty simple: Create an underlying hypervisor on the Xen model (which requires the cooperation of its virtual machines) and then build the OS in several zones, each of which is itself a separate virtual machine running its own mini-OS:

  • A communications zone would contain all apps with network abilities. This would include Web browsers, FAX programs, chat/VOIP programs, and all other TCP/IP based utilities.
  • An office information zone, which is where the user would install office apps for word processing, spreadsheets, presentations, drawings, and so on.
  • A storage management zone, which would contain machinery for backups, defragmentation, and so on. The file explorer would work from this zone; however, nothing in any zone would write anything to disk except through the services of the daemons; see below.
  • A daemon zone, in which special services would run that would monitor movement of information among the other zones. Nothing would move across a zone boundary except by the intervention of an appropriate daemon, which would do the moving. The daemons would work according to very strict rules, and would be designed to anticipate malware exploits. Executable items would not be moved until they were first copied to a sandbox (in a hidden zone inaccessible to the user) where they would be run and observed. Anything that attempts any of several actions, from scanning for files to tapping on the network port, would be quarantined pending further research. I don't think that relying on virus/worm signatures is enough. (The daemon zone might better be built into the hypervisor. Hey, I'm just brainstorming here; OS architecture is something I've never studied in depth.)
  • Additional zones could be created by the user, for the installation of specialty apps of whatever sort. I would create one for astronomy software, another for programming, and probably yet another for media work (video editing, etc.) Each major programming tool would probably go in its own zone. I do all my Delphi work now in a VMWare virtual machine, and that's where I'm building Aardblog.

Each zone would contain an independent operating system kernel, and all zones could run independently—though if the daemon zone goes dark, I think it would cripple the system. (One reason to exile the daemons to the hypervisor.) This is a little like how I understand the Mach microkernel works, but I don't think that Mach makes any use of virtualization. The user would not be constantly made aware of this zone partitioning, though I think it would be cool to have the screen divided into different colored areas by zone, with icons belonging to a given zone in a distinctly colored area of the UI.

Anyway. Just an odd idea. I'm trying to decide how narrow the zones should be: Should Web browsers and other Internet clients that deal in executable content be isolated in their own zones? Should every major app be its own zone? Or only those apps that can open executable content? Keep in mind that every zone needs an OS kernel, so if you have ten zones running at once, you're going to be losing a lot of memory to code duplication. Maybe that's a Who Cares?—especially once ordinary PCs routinely carry 10-20 GB of RAM.

The interesting thing about this concept is that it wouldn't have to be coded entirely from scratch: We already have Xen for a hypervisor, and both the Linux kernel and the BSD kernel are open source. For all I know, somebody's working on it already. I hope so.

July 26, 2005: Odd Lots

  • I've added Loren Heiny's Incremental Blogger to the blogroll at left, and I recommend it as a superb place to get the latest scuttlebutt on Tablet PCs. I'm researching tablets right now, as I'll be acquiring a Thinkpad X41 Convertible shortly.
  • One of the cooler things I found scanning Loren's blog is the Otter tablet PC cases. The company makes waterproof, ruggedized cases for PDAs, tablets, and other things, including (egad) cigars. The selection is limited right now (basically to the Fujitsu Stylistic 4000/5000 tablets) but I'm hoping they'll eventually get one together for the X41.
  • Yes, the big audio blocking caps in my 6T9 stereo amp (see yesterday's entry) are new old stock paper-dielectric Spragues, but the amp is fused and will not run especially hot, so I'm not as worried about the thing catching fire as some people might be.
  • Also, really and truly, there is no particular advantage to using coax as audio cable. I just do it because I have a lot of coax, and don't need audio cable very often.
  • If anyone would like to build a similar tube stereo amp without as much fuss, the guys who designed the 6T9-based amp featured in the August 2004 Nuts & Volts offer PC boards and parts kits. The circuit isn't quite the same, but in truth there's only so many ways to make a 6T9 amplify at audio, so there is a good deal of similarity between their circuit and mine.
  • Roy Harvey sent me a link to what may be the largest stitched-together digital photo ever assembled, with more than one billion pixels. It's something to see, and the page has some nice information on handling very large images.
  • I find it interesting, in perusing my Web server logs, that the split between Firefox and IE has not changed significantly since March of this year. Since then it's been 62-63% for IE and 24-25% for Firefox. Prior to March, Firefox's share had been climbing significantly each month, sometimes by two full points. In August 2004, the first full month of stats that I have, IE was at 70% and Firefox at 13.6%. Some of Firefox's gain since then seems to have been at the expense of Mozilla (not unreasonable) and Opera. All of these stats are on hit counts ranging from 120,000 last August to 170,000 in May. (Sectorlink lost some of my stats for June, and July isn't over yet.) You have to wonder if most of the people who are willing to (or can) move to Firefox have already done so. Still, 25% is a lot.
  • Cassini sent back some phenominally weird VLF signals from Saturn, so NASA downconverted them to audio and made a .wav file available. If you're going to do a haunted house next Halloween, just put this one on endless loop, and that's all you're gonna need for sound effects!

July 25, 2005: Building a 6T9 Tube Stereo Amp...Slowly

I've been fooling with a hand-built tube stereo amp project for almost a year now; see my entry for July 29, 2004 for the first mention of the project, and a photo of the amp in progress in my March 5, 2005 entry. The project is interesting for a number of reasons, but the least obvious is that I'm deliberately building it slowly. The reason is simple: When I hurry I make mistakes. I forget to solder connections, I use faulty parts, I leave out components or connections entirely.

So what I've been doing these past several months is spending half an hour on the project every so often, and cultivating the discipline of carefully soldering in one component...and stopping. I don't really need the amp, and I want it to look nice inside and out. I'm trying to emulate the old QST style of every wire and component running at right angles, though somehow, no matter what I do (and it isn't always possible) I can't make it look quite as good as they always did.

To avoid missing connections, I drew out the complete schematic in Visio, and printed it to paper. Then each time I solder in a component or a connection, I run a green highlighter over that component or connection on the schematic. I also take the time to test every single component before I solder it in, and wiggle the joint after I solder it to be sure the joint is sound. I make sure I solder in each component so that its markings are clearly visible from above, so that I will not mistake one cap for another when tracing the wiring or troubleshooting.

The big challenge in a project like this lies in arranging the parts, and anticipating the need for terminal strips. I want the wires and components to be neatly arranged and spread out so that I can get a test probe in anywhere I might need to later on. The fact that the 6T9 Compactron tubes require 12-pin sockets with very small pin spacing makes part arrangement even tougher.

And furthermore, I want no unnecessary holes in the aluminum box.

It's coming along nicely. The other day I got about as far as I could go before wiring in the three dual pots (volume, tone, and balance) which requires 12 separate shielded connections through RG-174/U miniature coax. (I use that instead of shielded audio cable, because I got a huge roll of it cheap at a hamfest years ago.) Because the underside will be much tougher to photograph with the three pots wired up, I took some pictures of the circuitry as it is now, knowing it will never be quite as tidy.

I don't know when I'll finish it. I deliberately didn't set a deadline. I'll post the schematic once I know the circuit works as designed. I've tweaked the design a little, especially the input network (which now has a balance control) and I don't want to publish something with bugs. When it works, I let you know, and tell you how I did it.

July 24, 2005: Review of Spielberg's The War of the Worlds

Most of my feelings about any film adaptation of Wells' The War of the Worlds can be summed up this way: Film versions of Victorian novels should be set in Victorian times. (When will we learn?) I saw Spielberg's new film last Friday night, and while it had its moments, they were few, and in general, it was a disappointment. There are a few spoilers below, but there isn't much to spoil; anyone who's ever read Wells' novel (or the wonderful Classics Illustrated comic book version) knows precisely how the plot plays out.

About half of my disappointment stemmed from Wells' Victorian concepts being set in 21st century New York. Slow-moving, philosophical British novels need to be set in a time when the people, culture, and situations in those novels don't violate our understand of that time. The Victorian world knew relatively little about biology and Mars, and had no radar, no jet fighters, and no nuclear weapons. Because today we could have seen them coming if they tried to land in conventional spacecraft, Spielberg has them sneak in under the cover of weird storms, having hidden fighting machines on Earth long before we evolved from lower animals.

Much of the film simply doesn't make sense for this reason. Judging by their mode of travel, Wells' original Martians had never been to Earth before and were clearly desperate and ignorant of conditions here. In Spielberg's retelling, the aliens have been visiting Earth for what may have been millions of years, burying thousands of 25-story metal tripods just underground (Huh? Nobody ever hit one digging a mine, a well, or a subway tunnel?) and waiting until we evolved a technological civilization to attack us. Knocking over classical Greece would have been a lot easier, no? And if they wanted us for food, as the film implies, establishing caveman farms could have been done with far less work and capital equipment 100,000 years ago. Worst of all, if they were a star-traveling species and had been here for so long, how in hell could they not know about the dangers of microbial life? And (in an objection I also have to Wells' original story) why didn't alien microbes kill us as thoroughly as ours killed them? How the aliens got from the stormy skies down into the still-buried tripods is another puzzle, explained so briefly and unconvincingly as to seem like an afterthought glued on to plug a plot hole. "Sheesh, guys! We forgot to explain how the aliens got down into their buried fighting machines!"

Yes, I'm prepared for emails bearing tortuous explanations, but Occam's Razor applies to film scripts as much as anything else.

Beyond that, lots of little inconsistencies insult our intelligence. An EMP that takes out every single car and truck throughout the Northeast leaves digital cameras and camcorders unscathed and functioning. If the EMP destroys alternators installed in cars, it would also destroy spares sitting on garage shelves. Hundreds of birds roost on what look like still-sealed fighting machines, as if they know there's dead alien meat inside and just can't figure out how to open the can. Etc. The guy thinks we're idiots.

The other half of my discontent is simpler to explain: We see far, far too much of Tom Cruise looking anguished and Dakota Fanning freaking out, and way too little of any sort of "war". We see the well-designed tripedal aliens (not Martians; the film never states where they come from) in poor light for all of five minutes, and the iconic tripod fighting machines for perhaps fifteen or twenty at most. Spielberg supposedly spent an immense amount of money on this project, and I think he got taken. The effects are well-done, but compared to something like The Lord of the Rings, they occupy a minor position in the film and mesh badly with the human drama in the foreground.

The film is tense, but the tension gets old after awhile. The little girl Rachel was annoying and completely unnecessary, and the rest of the human drama was overwrought. A seeming eternity is spent in the gloomy basement of an old farmhouse, dodging first alien tentacles and then the aliens themselves. We see a lot of American military hardware, but rarely in the same frame with the alien tripods. Hordes of people trudge slowly around rural Connecticut like extras from Night of the Living Dead. I kept wanting something to just happen.

I had the same general reaction to M. Ramalamadingdong's pretentious and excruciating Signs: More (and smarter) aliens. Less Mel Gibson and other bad actors chewing on the curtains.

The great tragedy of Spielberg's The War of the Worlds is that it has evidently buried the authentically Victorian period adaptation of Wells' novel. It's been consigned to DVD and will see no action in theaters. Supposedly I can rent it, and I'll begin looking for it. I may be disappointed, but I suspect I won't be insulted.

(Now, does anybody want to know how I would have scripted a modern-day setting of WOTW?)

July 23, 2005: Odd Lots

  • Frank Glover sent me a pointer to a blog that suggests persuasively that not all suicide bombers may know that they themselves are going to blow up too. The recent London bombers had paid up their parking fees and had the bombs in backpacks, suggesting that they were planning on leaving the bombs somewhere to blow up after they escaped. (Read the item; there's more than I can summarize here.) Whether or not this is true, I think we should all circulate the possibility until it becomes a meme, which might make it just a little harder to recruit future bombers. Not everybody believes in a cause enough to die for it—and there may be a sort of "bombers' remorse" effect that might generate a few more cold feet if we fed it.
  • The next major version of Windows, now code-named Longhorn, is going to be called Windows Vista. Was XP, now Vista. Ya gotta wonder what's wrong with Windows 2002 or Windows 2006. Actually, what's wrong with "Windows 2006" is that it doesn't sound like there's much there worth upgrading to, which is the case more often than not. I still think Windows XP is just Windows 2000 in a clown suit.
  • I spent some late night nostalgia cruising through Oddball Comics, a blog that dips into the stacks of long-past comic books for the most peculiar examples. I was surprised to see how many I recognized, considering that I only saw comic books in my cousin Ron's basement or at Boy Scout campouts in the early-mid 1960s. I was never much for conventional superheros, and preferred things like Monsteroso from Amazing Adventures #5 or Metal Men, who were robots whose personalities mirrored the characteristics of various metals. (Anthropomorphizing gallium—what a concept!) Lotsa fun. Although he doesn't mention it, I should also point out the Catholic comic book Treasure Chest, which ran a famous series called "Godless Communism" when I was in third grade. You might consider that oddball (or just surreal) but it has nothing on a story arc about a bear and a wainwright that ran in the mid-1960s that I have been unable to find. (Somebody, somewhere, has a stack of these things in the basement, I'm sure.)
  • People who prefer clicky keyboards (technically called "buckling spring" keyboards) should read this page from Dan Rutter, which is a good summary of your IBM-centric options, though remarkably, Dan doesn't even mention Northgate. Most keyboards were clicky until IBM lost control of the personal computer market in the late 1980s, and those of us who learned computing then for the most part prefer them as they were. I like Dan's style; in closing, he says: "Computer users are used to hardware that's worthless in three years and useless in five; clicky keyboards aren't like that. You could leave one of these things to your children in your will. Or be buried with it, like some kind of nerd Pharaoh." Heh. Indeed.
  • More nostalgia: Backyard Artillery cites a lot of things we knew and loved in the 50's and 60s, like cap bombs and ping-pong ball rifles. I'm dubious about the fully automatic rubber-band machine gun (you have to crank it and then it shoots rubber bands) but it's a marvelous concept nonetheless.

July 22, 2005: Update on My Personal Spam Wars

I've been having intermittent problems with both Poco Mail and POPFile, and I've been thinking about whether a major email strategy change is in order. If I can figure out how to get my mailbase out of Poco and into Thunderbird I may jump. Maybe this is a good opportunity to lose about, oh, 10,000 messages from the mailbase...

POPFile's accuracy has come down significantly in the last year, from 99.6% to its current 97.75%. I can still live with that, but I'm not sure why the drop, except that it corresponded roughly to my move from Interland to Sectorlink and the corresponding 80% plunge in my daily spam count. It may be that when 95-97% of your mail is spam there are just fewer opportunities for false positives. Now only about half of my mail is spam, so each individual message's butt is statistically a little more likely to get hit in the crossfire.

POPFile needs a "teaching magnet" but it's really the best weapon we have right now. I just completed a three-week test using keyword filtering to see how well I could block spam based on "payload" domains in the message body; that is, the destination URL or (much more rarely) phone number that the spam was intended to convey to the recipient. The payload is the only thing the spammer can't obfuscate, or the campaign would fail, so it's a reasonable thing to block on.

However...

Over the past three weeks, I discovered that the daily rate of new payloads was between 30 and 40 percent. In other words, every day in the last three weeks, one third of my spam was pointing to domains I had never seen before and therefore had not blocked. I didn't see any new obfuscation tricks. I like those, because they're spammer-specific and can be filtered on. Nope, the sole strategy of the zombie spammers is now to rotate payload domains on a near-daily basis. This doesn't leave us much except Bayesian filtering, of which POPFile is the most highly evolved example.

By the way, reader Andrew Colbeck confirmed what Darrin Chandler suggested in my July 1, 2005 entry: That spammers do not generally cache DNS lookups. In other words, once a spammer has an IP for you, he'll use that IP without further verification, to avoid the time cost of a DNS lookup. About 80% of my spammers are therefore probably still hammering the IP address of my old POP server at Interland. I'm a little surprised they don't "freshen" their lists now and then. I would expect something more than about a 10% increase in my daily spam count in six months, but that's all I've seen.

If we were all using IPV6, we could retire IP addresses every month or so, and stay ahead of these "optimized" spammers that way, but our current IP addresses are too scarce for that. Bulk domain sales have made domain names essentially disposable. My suggestion that public DNS records contain a unique code for each domain owner would be extremely useful: With that, we could just tell our software to block every domain owned by the same people who own shitheadmarketing.com. They can buy domains in bulk, and we could block domains in bulk, and the best thing is, it wouldn't even violate domain holders' privacy. The code would contain no information, but would simply allow us to identify which domains are owned by the same someone—and we wouldn't have to know who that someone is. Alas, even if the political issues could be solved, there are practical challenges in that no single agency registers domain names, and thus managing a unique domain owner code across all domain registrars would be a huge logistical problem. Damn, I can dream, though.

July 21, 2005: War Against the Weak, Concluded

The big question that Edwin Black's War Against the Week fails to address is simply, What were these people thinking? How could eugenics have gotten so far and stayed "legitimate" for almost fifty years? Some of my thoughts here:

  • Racism itself was legitimate (meaning legal and accepted by ordinary people) until well into the 1950s. Our sensitivity to race issues is in fact a very new thing.
  • Eugenics played to a visceral fear that "our tribe is being outnumbered by the other tribes." This fear is still with us (it's in our genes, I suspect) and we sense it today in a lot of discussions running from political parties to immigration to the rise and fall of religious traditions—but people are no longer seriously talking about murdering or sterilizing the other tribes. (Not in public, at least.)
  • Utopianism was a very big thing in the century1850-1950. Most utopian schemes are both elitist and coercive or even totalitarian, and none of them work for long, if ever. Eugenics was very much a utopian idea, and just as lame as all the others.
  • Eugenics was a favorite idea of the cultural elite among the urban moneyed classes and the universities. (Ordinary people in the white middle class did not widely embrace eugenics and often protested vehemently against it.) The ones who pushed eugenics the hardest were the same ones who dominated what passed for mass culture at that time. (How many Black folks had seats at the Algonquin Round Table?) Thus, eugenics may appear to have been more widely embraced than it actually was because those who embraced it were those who did most of the writing and defined most of the culture.

However, my favorite personal theory involves a strange psychological shortcoming present in many people: They get fixated on an idea, and can't put it into perspective before they experience it directly. Years ago I read an article in the Rochester, NY Sunday paper about some fool who moved from Manhattan to Rochester, and one day decided that he absolutely had to have a bag of onion bagels from his favorite grimy little deli in Brooklyn. It wasn't until he had driven most of the 300 miles to New York City that he realized what a total idiot he was.

Ugly ideas often sound compelling when they can be embraced only in the abstract. Much of the big noise in the Libertarian movement comes from addle-brained anarchists who have no idea what the consequences of genuine anarchy would be. Wave a blood-smeared real-world example in front of them and they're likely to object, Those people in Somalia just don't know how to do anarchy correctly! (We need an Anarchy Corps so we could send our home-grown anarchists over there to show them how it's done. One-way tickets will suffice.)

Sterilizing "defectives" sounds great until you (or someone close to you) gets classified as "defective" and ends up a statistic. (It happened to between 70,000 and 90,000 Americans between 1900 and 1960.) I keep thinking of upper-middle class liberals whose mindless chant is "Soak the rich! Soak the rich! Soak the rich" until the Feds say, "Sure thing. Guess what? You're rich!"

Ideas have consequences. Always. The inability to imagine consequences is a tragic human failing. I think that that failing was the reason that people who would never consider clubbing a handicapped person to death themselves would in all earnestness nod approvingly when they read some rant by some university racist talk about "lethal chambers" for the "genetically unfit."

Anyway. War Against the Weak is required reading for people who think that we live in a morally debased era. Not true. Piddly things like promiscuity or flag burning vanish into insignificance next to ideas like eugenics, which ultimately led to Auschwitz. The more I read history, the more I appreciate our own era, which even with its flaws is the most humane era humanity has ever seen.

July 20, 2005: Odd Lots

I just got back to Colorado Springs, and the heaps of suitcases in the middle of the living room floor whisper that there will obviously be no time today to sum up on War Against the Weak, so clearing a few odd lots will have to suffice:

  • I did some significant work to my Tom Swift, Jr. page a few weeks ago, including some edits and a few more cover scans. I'm getting close to being able to replace the ancient duntemann.com front page; the page is done and just needs the creation of a few more things before I can post it.
  • Pete Albrecht sent me a link to a roll-off observatory created using 80/20 extrusions and connector components. Although the observatory shown is a commercial product, there's nothing in it that a savvy tinkerer couldn't do with hand tools. 80/20 is very cool (they call it "the Industrial Erector Set", which is bang-on) and I hope to do something with it someday.
  • From the cautionary tales department comes Doomed Engineers, a page summarizing brilliant men who came to bad ends, mostly because they were not as wise as they were smart. The poster child here is Gerald Bull, who was so obsessed with the idea of shooting projectiles into orbit with big guns that he tried to build a monster artillery piece for Saddam Hussein and got himself rubbed out by the Israelis. Bull didn't invent the concept, which goes all the way back to Jules Verne, in his fine old 1879 novel The Begum's Fortune, which should be online somewhere but I haven't had the time to find it. Thanks to Pete Albrecht for the pointer.
  • Finally, while Carol and I were crossing from Terminal 1 to Terminal 2 at Chicago O'Hare yesterday, we ran across a little kiosk where Lenovo (to whom IBM recently sold its Thinkpad line, along with most other IBM PC stuff) was doing hands-on demos of the Thinkpad X41 Convertible tablet PC. A young woman took me through it, and I was most pleased, pleased enough to feel like I'll grab one once my X21 is fully depreciated in a few months. If ebooks are going to happen, machines like this are going to have to go mainstream first. More as I learn it, but it was nice to actually hold an X41 in one hand and my venerable X21 in the other hand. They weigh the same, and are almost exactly the same size, which is small.

July 19, 2005: War Against the Weak, Continued

(Continuing my review of Edwin Black's book, War Against the Weak, begun July 18, 2005.)

Among the photo plates in War Against the Weak is a still from The Black Stork, a low-budget 1917 film written by a Chicago newspaper reporter and produced in Hollywood. The still shows a mother and father standing before a dead newborn, whose soul is depicted rising into the arms of a scruffy-looking, miasmal Jesus. Typical silent-era tearjerker material, with a crucial difference: The parents decided that the infant was not fit to live, and killed the poor thing. A glossy poster promoting the film read, "Kill Defectives, Save the Nation, and See The Black Stork."

As appalling as it might seem to us today, The Black Stork was very popular in 1917. Chicago's LaSalle Theater played the movie continuously between 9 AM and 11 PM, and it ran intermittently in theaters around the nation for almost ten years. It was the brainchild of Dr. Harry Haiselden, chief of medical staff at Chicago's German-American Hospital. Haiselden was a man of gleeful coldness who makes the most indifferent abortionist seem like Santa Claus. He not only admitted infanticide (though neglect and refusal to treat or provide basic human needs to infants) but declared that all physicians did it, and that it was a necessary step to keep defective individuals out of the human gene pool. He laughed at people who expressed concern for the euthanized infants, and quipped that "Death is the best disinfectant." Several prosectors attempted to convict him of murder or medical malpractice, but none succeeded.

One of the great strengths of Black's book are its recall of minor incidents, mostly forgotten by history, like Dr. Haiselden and The Black Stork. Eugenics was not freakshow stuff in 1915. It was the material of public debate, and the eugenicists were taken seriously, even when they suggested, as did the author of the popular textbook Applied Eugenics, that genetic defectives (a term never crisply defined but often assumed to those of include low intelligence and lacking moral fiber) be killed.

On the negative side, Black goes to great lengths to exonerate Margaret Sanger of racism. Sanger was an early feminist and the person who coined the term, "birth control." She's become a feminist hero as being the first person to demand that women have the freedom of choice about sex and childbearing. And although she was not primarily a racist (far more of a Malthusian, actually) she expressed belief throughout her long life that people who could not prosper in society should be forcibly sterilized, even into the 1950s, knowing that many or even most of the very poor were nonwhite. (See this article.) She surrounded herself with some of the worst American racists of her time, including Lothrop Stoddard, author of The Rising Tide of Color Against White World Supremacy, and approved of most of the elitist thought coming from her friend H. G. Wells in the last years of his life. It's impossible to understand either Sanger or Wells without understanding the Fabians (a British lefty utopian clique most active between the Great Wars) but Black does not mention them even once. When Sanger finally broke with organized eugenics, it was because leaders of the eugenic movement (virtually all of them men) could not abide her strident feminism and threw her out. I'd rather have honesty, and won't insist that Sanger's support of eugenics casts doubt on her feminism. Nonetheless, no one who had fingers in the eugenics movement came away with clean hands.

The most disturbing facts among the many presented in War Against the Weak are accounts of state laws passed in the early 1900s providing for (or even requiring) sterilization of the unfit, and regulation of marriages to prevent unions between whites and those of other races. (Unions between partners both of nonwhite races were not illegal and not discouraged.) Between 70,000 and 100,000 Americans were forcibly sterilized between 1900 and 1960, and countless interracial marriages were prevented, their intended partners harassed and sometimes charged with felonies. Although enforcement became uncommon after World War II, some of those laws remained on the books well into the 1970s.

As most people know, the most enthusiastic proponents of eugenics were the Nazis, and until they provided real-world illustration of eugenics in action (rather than merely university lounge conversation) eugenics retained its place in American and world thought.

I'll conclude tomorrow.

July 18, 2005: Review: War Against the Weak

For this year's beach reading I chose Edwin Black's War Against the Weak, which I stumbled across in my ongoing quest to understand the roots of the Roman Catholic Church's peculiarly intense (and suicidal) prohibition of preventive contraception.

War Against the Weak is a solid and very readable (if slightly shrill) history of eugenics, which in turn is nominally the quest to improve the human race through selective breeding. Most people who have looked into the topic even briefly know that the whole thing was a sham, but few, I suspect, understand just how deep and how ugly a sham it was. Eugenics is perhaps the purest example in all history of bad people using bad science as an excuse to impose their own biases on society at large.

Very shortly after an obscure Augustinian monk named Gregor Mendel described some simple rules of inheritance that he had observed in peas grown in a monastery garden, the world's educated elite seized on the concept as evidence that all human traits were inescapably heritable. Some of it may have been the when-all-you-have-is-a-hammer-then-everything-looks-like-a-nail effect, but the greater part of the enthusiasm was far simpler: It was the perfect excuse to declare any group not in favor with the university elite "genetically defective." Prior to Mendel, bias against the poor, the unsophisticated, and the nonwhite was just that: Bias. After British mathematician Francis Galton popularized the notion of heritable human traits in the 1880s, he coined a new word, "eugenics," to stand for the goal of improving humanity through breeding. Galton did not know of Mendel's research, and though he described the "what," he was clueless as to the "how." Mendel provided the "how," and that's when the party began. Edwin Black documents the dark party of eugenics, from its origins in Victorian England to its end in the Nazis' Final Solution.

The really nasty part of eugenics is how much of its history occurred right here in the USA. American organizations, including the Rockefeller and Carnegie foundations, funded eugenic "research" and the popularized the concept through books and journals. American universities, never able to resist an intellectual fad that eventually makes them look like mean-spirited idiots, provided researchers and academic credibility. Some of the most influentual men of that era, including Alexander Graham Bell, Robert Yerkes, George Bernard Shaw, and H. G. Wells, bought into eugenics and argued for government measures including mass sterilization of the "defective" and rigorous regulation of marriage to prevent the "unfit" from reproducing. No sooner did the British invent the "lethal chamber" as a humane way to kill stray dogs and cats circa 1900 than the worldwide eugenic community (including groups in nations as progressive as New Zealand) began to discuss in their journals and conferences whether the best way to solve the problem of the genetically unfit was simply to gas them.

No, I didn't initially believe it either, but Black provides an immense body of research to support the book; his footnotes and lists of sources take 80 pages alone. I don't much care for historians who insist on telling me how outraged I should feel (as Black does a little too often) but the material itself is so appalling I can understand him getting a little unhinged about it. In truth, Black is not a historian but an investigative reporter, which explains the general tone of the book.

So where did the passion for eugenics originate? That's no mystery, and Black explains it well: The last quarter of the 19th Century saw unprecedented migration from Eastern Europe to Western Europe and the United States, and of rural Blacks from the southern US to the urban north. The educated and moneyed classes saw their own racial group being swamped by great numbers of browner and less educated people, who (of course) "breed like rabbits." Popular books with titles like The Passing of the Great Race and (egad) The Rising Tide of Color Against White World Supremacy fed the fires of racial paranoia. Racism as an unremarkable fact of ordinary life in this period is well-known, but most people think of such racism as a moral flaw within individual human beings. Eugenics systematized racism and tried very hard to give it the glint of scientific integrity and the force of government authority.

The scary thing is the degree to which the eugenicists succeeded, and I'm not even talking about the Nazis. Depressed yet? If not, patience: The worst is yet to come. More tomorrow.

July 17, 2005: Vacation Recovery

Carol and I got back to Chicago O'Hare a few hours ago, and now begin the process known as "vacation recovery." We're taking a day or so here to visit, and then will be flying back to the Springs this Tuesday afternoon.

We had a great time, and have half a suitcase full of salt-water soggy clothes to prove it, but as I've mentioned here and elsewhere before, I have become very impatient about getting answers to even minor factual questions that occur to me. While sitting on the beach, I recalled an old New Colony Six novelty song from the late 1960s, and I wanted to look up the lyrics for "Wingbat Marmaduke" right now. Later on, it came up in conversation that Tommy Lee Jones had a role in Love Story, which I found very hard to believe. Where's IMDB when you need it? While poking through the local Anglican cathedral, a hymn popped into my head, but I was missing one of the lines of the melody. I wanted to bring up Cyberhymnal and play the MIDI file, with a sort of fiery impatience that startled me.

None of these things are important (and there were dozens of others) but they're also precisely the sorts of things that come up when you empty out your brain and refuse to think of anything important in an organized manner beyond when the next high tide (or maybe lunch) will occur. So part of my vacation recovery will consist of emptying out my mental file of trivia to research, just as soon as I can get over to Panera Bread—though probably not until tomorrow morning.

And if you're reading this, I've already verified Tommy Lee Jones' role in one of the worst movies ever made (his first film appearance, in fact) and can play "In Babilone" in my head in its entirety. (The Wingbat Marmaduke has eluded even the long reach of the Web. I'll play it when I get home.) Little itches matter the most—especially when you can't scratch.

July 12, 2005: The Art of Photochemicals

Bob Halloran wrote to suggest (a little bit sadly) that ours may be the last generation that knows how to use a darkroom. I agree. In fact, even among the Boomers darkroom tech was getting pretty geeky by the early 70s, because of the availability of cheap and quick snapshot services down at the drugstore. My dad tinkered with color-slide darkroom techniques when it became possible circa 1950, but set aside that (and most of his several other hobbies) when the kids started coming along. I lacked an enlarger, but I had a small bellows camera that took 120 film, which is a large enough format to produce contact prints of a useful size. I played around with b/w development (Dektol!) in the basement, particularly with lunar and planetary photos taken on my crude 8" Newtonian scope.

What bothered me then about darkroom work was that there was a lot of black art in it. I had a hard time duplicating my successes and avoiding my failures. There were broad book-learned techniques to be followed, but there also seemed to be a lot of wildcards in the process, and eventually I set it aside, not finding the fascination for photography that I had found for telescopes or electronics.

I think what attracts a lot of people to digital image processing is that it's repeatable. There's some black art in it to be learned, but no wildcards—if you study what's going on closely enough, you can quantize every adjustment and transition and do it again identically. This is especially true of color work. A color chart is purely conceptual in chemical photography. In digital photography, the chart maps to real numbers (which produce real pixels) inside a real image. Given the same image, the same manipulations will create the same results, which is much more repeatable than holding your thumb over a dark part of the image for a second while you turn the enlarger light on.

Maybe this is a loss, but maybe not. I think that chemical photography will continue for a good many years, rather like tube audio, even if 98% of photography enthuasiasts go digital. And just as tube audio hobbyists have benefited by better passive components born of solid-state science, so chemical photography guys will benefit from better treatments of color theory in the press, better optics, and perhaps other things as well. We just know more than we used to, in almost every facet of every field, including optics, color, and image manipulation.

To the contrary, maybe it's one of those increasingly rare win-win situations. I hope so. And I'll be glad to let the chemical photography guys have their Dektol (or whatever it's called 40 years later) and their black art. Me, I just want the pictures.

July 9, 2005: On Vacation for a Bit...

...so if entries here are a little sparse, bear with me. I'll do what I can. I never much liked writing entries in airports. We'll see how it goes in the hotel.

In the mean time, my current laptop depreciates fully early next year, and I think I know what my next laptop will be.

July 8, 2005: The Two Megapixel Difference

I've had a two megapixel Canon Digital Elph camera since Christmas 2000. It has worked flawlessly all this time, producing images that, when printed using any of several printing technologies, look as good as almost any of the photographic prints that came out of my several "automatic" 35 mm film cameras over the years. This past Christmas, I gave Carol a four megapixel Kodak digital camera and matching printer dock. It's been interesting seeing the prints that her camera produces, on its own printer and other print technologies.

They seem fundamentally different somehow. They're sharp, razor-sharp, so sharp that my eye/brain partnership looks at them and tells me, That's not a snapshot. That's something entirely new. It's a whole different way of remembering, because there's so much more "remembered" in the photos. We can see things in the prints that we could never see before: The moist glisten and texture of Q-Bit's little black nose, or contours in Carol's hair that had always been ever so slightly blurred out in older prints.

The improved resolution is, of course, a partnership between a camera that captures finer detail, and a printer that can express finer detail. It's not just the camera. Chemical photographic film and paper have inherent limitations of resolution, especially the less expensive and faster kinds. I'm sure there are photographic ways of delivering resolution like that. I'm also sure that they're neither easy nor cheap.

I can't compare my Canon's prints and Carol's Kodak prints here in Contra for you because computer screens don't have that kind of resolution, but it's striking how striking (metastriking?) the difference is. There's a quantum jump between two and four megapixels that makes me wonder if there will be a similar difference in perception between four and six megapixel cameras, coupled with new printing technology that we don't have yet. Four megapixels almost looks weird, as though I have a space-warp window into another time and place. Will six or more megapixels rendered in a high-resolution display technology be so lifelike as to be disorienting? Either way, it looks to me like a whole new art form, and it will be interesting to see what the professionals do with it.

July 7, 2005: The Sovereignty of God

I read a fair bit of theology, and there are times (as one who cherishes the Catholic tradition) that I feel like a stranger in my own country. One of these times happened recently while I was reading a near-rant about the sovereignty of God. The expression means just what you would expect: God is the ruler of the universe, without exception. He is all-powerful and nothing escapes His notice. To me this is a headscratcher. Making the point that God is all-powerful is like saying that the dogcatcher catches dogs, or that magnets attract iron filings. It's built into the definition: If God isn't the all-powerful ruler of the universe, well, he's not God and we need to keep looking.

I see discussions of the sovereignty of God on a regular basis, and in my experience the vast majority of those discussions make the point that God can do what he likes with us or anything else, especially sending us to Hell, but also including striking dead Uzzah, the poor goodhearted slob who tried to steady the Ark of the Covenant (II Samuel 6:3-7), or having Elisha whistle up some bears to eat a bunch of kids who were making fun of his baldness. (II Kings 2: 23-25) (Because I understand doesn't mean that I approve, heh.) Basically, as the term is most often used, "the sovereignty of God" means "God can be a bastard if He wants to."

Well, sure. The ranters and I are in agreement on that one. God can be anything He wants to. God can be an Elvis impersonator or a Pizza Hut delivery boy, though I'm sure He has better things to do with His time. A lot of people speak of the sovereignty of God as a cover and an excuse for saying, "God can do what my own mean-spirited, envious, vituperative self wants Him to do, which is mostly cause pain to or consign to Hell the people I disapprove of." (Some few use it as a way past problematic Scripture passages like those cited above, and fewer still engage it in useful discussions of the tension between divine power and human freedom.)

My point is this: Never, not even once, have I seen the sovereignty of God invoked to support the possibility that God will be kinder or more forgiving than our crippled understanding of Him suggests. God can do anything he wants, which includes rehabilitating even the worst of us, while giving us enough leash to be truly free creatures on this Earth, and neither puppets nor pets. The sovereignty of God is paradoxically the best support for both eternal damnation and universalism (simultaneously!) that you could find.

As I mentioned earlier, inside the sock puppet of all these rants on the sovereignty of God are two serious discussions, one about free will and divine power, and the other the interpretation of Scripture. I've thought some about both, but I'm about out of time this morning. More in the next few days if I can squeeze it in.

July 6, 2005: Virtualizing the Internet Experience

Sooner or later, you get stung: You navigate to a dicey Web site and something ugly installs itself on your PC. Or you forget yourself and open an attachment (or are fooled into opening it) and a trojan starts marching on your registry. Ditto if you download something from a Usenet newsgroup that contains a little extra ingredient. Some of this malware is diabolically difficult to dislodge once it's in place, and in many cases you have no choice but to wipe your disk and reinstall.

There may be another way. My recent experience with virtual machine software like Virtual PC 2004 and VMWare Workstation 5 suggests an interesting possibility: Create an Internet suite (something like Mozilla) that contains its own virtualizer, and thereby run all of your Internet software in a virtual machine. I can envision a sort of launch bar floating on your screen, with icons for Web, mail, Usenet, and IM. The launch bar is a window into a virtual machine, with an additional icon to bring up a management window, where you can take snapshots or delete them and configure the overall system.

This could be done by starting with a copy of Linux and the Xen hypervisor, stripping out the general GUI, and building a sort of mini-OS that installs a "toehold" in the form of a Windows service. When the suite is launched, the service loads the hypervisor and creates a virtual machine for the suite apps to run in. You hold a fully configured snapshot of the virtual machine in reserve (it wouldn't have to be more than 100 MB in size, and probably much less) and if the malware bites, you abandon the virtual machine you were using and restore from the "clean" snapshot.

This is a little like using Norton Ghost or other brute-force restore utilities, but with a crucial difference: Your underlying PC is not affected in any way if malware strikes. This is true for two reasons:

  • The Internet suite apps are not running under Windows at all, but under an embedded OS based on Linux. Malware intended for Windows simply won't run, and because the hypervisor would not run its apps as root, even Linux malware can't install itself.
  • The isolation of the virtual machine from the underlying system is extremely strong. I won't go so far as to say it can't be broken (at least until our CPUs have built-in silicon support for virtualization) but it will not be easy for malware to "get out" of the virtual machine.

The weak link in any such system lies in how the virtualized Internet suite shares files with the underlying host Windows installation. That's why I suggest that a Windows service be running at all times: Periodically, the service would "peek inside" the Internet suite's virtual machine (this is easier than going the other way) and sync whatever files the suite changes (mail, bookmarks, newsrc, etc.) to copies in a directory on the Windows side. This would allow other Windows apps to access the Internet data, and would also allow easy restoration of the state of the Internet suite if the currently running virtual instance had to be abandoned. The service could invoke a malware detector on anything it syncs out from the virtual system. (I would suggest not taking along email attachments or anything even remotely executable.)

Just a goofy notion, heh. I now have 2 GB of RAM on my main system, and will soon have 4. Installing a terabyte of hard disk is no longer prohibitively expensive. (It's actually getting pretty cheap.) How are we going to "spend" all of these riches? How about virtualizing the Internet experience? Most of the misery in PC land is caused by bad things that come in through your Net connection. We have the tools to isolate Internet work from Windows itself. All we need is the will to do it.

July 5, 2005: Odd Lots

  • Pete Albrecht sent me a link to Heavens Above, which will tell you when most of the brighter satellites are visible in the sky from your location. This has helped me to spot the ISS, which, while very bright, scoots across the sky in a great hurry and can be easily missed. The only time I've seen an Iridium flare was thanks to its guidance. First rate.
  • Many years ago, I used to listen to avaiation chatter on the AM aircraft band at 108-136 MHz. I don't even have a scanner anymore (I sold it before we moved away from Arizona) but if you're interested in that sort of thing, this page will show you where aviation traffic happens.
  • The S-Meter page cited in the previous item is interesting in itself: It allows you to hear radio signals through Windows Media Player. Art Bell hosts one of the Web receivers, but it doesn't seem to be related to most of the tinfoil hat stuff he presents on his own show. It's fun (assuming you can stomach WMP) and I encourage you to give it a try.
  • There's an interesting conflict going on these days between vendors trying to sell GPS re-radiators and the FCC, which doesn't allow them in the US. Wal-Mart had been selling them and has now stopped, and the FCC is pursuing any retailer attempting to sell the devices, which are made in the Pacific Rim. (Would that the Feds would go after spammers as enthusiastically as that!) GPS re-radiators are niche-y but useful, if you have a GPS-based appliance but don't have good access to the sky from the appliance itself. They simply receive the GPS signal, amplify it a little, and re-broadcast it on the same frequencies. People who own Meade GPS-equipped telescopes sometimes find that the peculiarities of an observing location (trees, buildings, whatever) attenuate GPS signals past usefulness, so mounting a re-radiator under clear sky can strengthen the signal that the scope itself receives. Re-radiators are apparently legal in most places outside the US, so if you want one, you're going to have to pack it home in your luggage and pretend it's just another damned laptop parasite.

July 4, 2005: Applause

Several of my friends have displayed some eyerolling at the line in Episode III, in which Padme says: "That's the way democracy ends, with applause." Yeah, like, well, who would applaud a dictator?

Yeah, like, well, people whose democratic government has become corrupt, incompetent, and dysfunctional. Hitler didn't come out of nowhere. Germany was economically hobbled by the Treaty of Versailles, and the Weimar Republic, created out of the ruins of WWI, was in a state of near collapse when Hitler took power in 1933. The German people wanted somebody, anybody, who could just get the country functional again. The currency was worthless and people were starving. They applauded him. Boy, did they applaud him. Like Mussolini, he made the trains run on time, and got the country functional again. Of course, he didn't stop there...

Even though we're nowhere near as bad off as the Weimar Republic in 1933, there is still cause for concern. Like I've said many times, freedom and democracy are loosely coupled, and not everyone values all freedoms. (I'm astonished at the viciousness with which homeowners' associations prosecute people who display the American flag, which virtually all deed restrictions now prohibit.) Many people, perhaps most people, would gladly give up the freedom to choose government representatives if they felt it would allow them to keep their jobs and some feeling of personal safety. We have entered a period in American history when, for whatever reason, almost no one feels secure, and that insecurity often has no grounding in reality. It's all a little odd; while we seem to imagine an Islamic terrorist behind every tree, it's the Supreme Court who's most likely to seize our houses and hand them over to Donald Trump. The rich people who build the nasty little gulags we call "gated communities" are convinced that everyone except other rich people hate them and would slit their throats in a second given the chance. People who are losing their jobs in today's economic upheaval think illegal immigrants are ruining the country. Liberals think conservative Christians are trying to impose Biblical law on America. Conservative Christians think liberals are trying to outlaw all public expression of faith. Almost nobody expects to receive Social Security except those who are already receiving it.

What's really happening is this: Everyone has a personal worldview, and within each worldview is a smallish slice of insecurity. Trouble is, those slices are all different. So even though the country is working reasonably well in the aggregate, almost everyone has their own little slice of creeping dread, and this slice is the lens through which they see the future. Everybody is afraid.

As long as most people are making a living and paying the bills, the worst that happens is acid indigestion. But if things start getting really bad, the exaggerated terror people feel today will become overwhelming, and even a modestly charismatic leader who pushes the right set of buttons will be handed the keys to the country.

As bad as things were, people still believed in the future during the Great Depression. Almost no one believes in it now. My great fear for America is that we may lose democracy (and most of our remaining freedoms) simply because we are afraid.

I happen to think things are still going reasonably well. On the other hand, I keep one ear cocked for the sound of just a little too much applause.

July 3, 2005: All My Outbound Mail Is Coming Back

I have some interesting things to write about, but I have a more urgent issue this afternoon: Since some time yesterday morning, every single email message that I have tried to send through my Sectorlink SMTP servers has bounced back to me, with several completely irrelevant error messages, some as absurd as the contention that AOL's DNS information cannot be found. I tried to reply to a note from Michael Covington, and the system tossed the reply back in my face, scolding me for attempting to relay through uga.edu. All my outbound mail goes through SMTP servers hosted at Sectorlink and DNSed to my own domains; there is no relaying of any kind going on, and my mail setup has not changed in many months.

I've begun to worry about Sectorlink's corporate health. For several weeks now my Web stats have been utter nonsense, with no visitors logged to my site sometimes for three days running—and then 200+ the day after. Sectorlink tech support doesn't seem to be reading my support tickets. I have no idea what's going on.

Anyway, if you've written to me in the last 30 hours or so and haven't received a response, this is why. I'm going to start replying to email via my GMail account while I try to chase this thing down. Sorry for the messup and do bear with me. (I've begun shopping for hosting services again, sigh. Like I need another project.)


Update, early July 4: Things are working again, at least on the email side. I don't know what went wrong. However, I'm still shopping.


July 2, 2005: "Throw Another King on the Fire..."

Pete Albrecht and I were talking about efficient fuels for home heating a while back, and he mentioned that maybe we should start burning mummies—and not just any mummies, either. "Kings burn better than peasants—Mark Twain said so!"

As Pete remembered it, Sam Clemens was touring Egypt in the late 19th century, where they had begun to unearth so many mummies in so many odd crooks and crannies in the desert that people had begun burning them for fuel. After awhile it got even worse than that: Fuel for steam locomotives was scarce in the Egyptian desert, so engineers would sometimes start their wood or coal fires with chopped up mummies (I'd heard this myself years ago) or just stacked them like cordwood in the tender along with anything else that could conceivably be used to boil water. Supposedly Mark Twain had called out to the engineer of a slow train on which he was riding, "Hey up there! Throw another king on the fire! Those peasants don't burn worth a damn!"

I'd never heard that, but it sure sounded like a Twainism. Pete and I decided to look it up, and found this sage explanation from Cecil "The Straight Dope" Adams. Whoops. As with a lot of outrageous things, it was a very early urban legend—basically, a meme that grew out of a gag that Twain had included in Chapter 28 of his wildly popular 1869 book The Innocents Abroad, which spoke of his travels in Egypt and described what he saw there:

I shall not speak of the railway, for it is like any other railway—I shall only say that the fuel they use for the locomotive is composed of mummies three thousand years old, purchased by the ton or by the graveyard for that purpose, and that sometimes one hears the profane engineer call out pettishly, "D--n these plebeians, they don't burn worth a cent—pass out a King;" (Stated to me for a fact. I only tell it as I got it. I am willing to believe it. I can believe anything.)

It's true that mummies were used here and there for Victorian parlor entertainment, in what was called an "unwrapping." (The idea is so ghastly that I'd just as soon not learn any more.) They were also ground up as patent medicine, and on occasion for pigment used in paint. Sometimes the wrappings were used in making paper, just as any rags were. But as best we know, they were not regularly burned as fuel, for locomotives, home heating, or anything else.

All this leads to another question: How broadly is Cecil Adams known around the country? Pete and I are Chicago boys, and I had forgotten that Cecil is a columnist for The Chicago Reader, the oldest of the several alternative papers in my home town. He's been doing "The Straight Dope" for 32 years now, since I was in college at DePaul University on the near north side. Cecil's actual identity is obscure. My guess is that he is a creature of The Reader itself, much as Victor Appleton (and his son) were creatures of The Stratemeyer Syndicate. We may never know, and that's all right. Santa Claus ceased to be useful when I was seven or so. Cecil Adams is the Answerer of Last Resort, and that's a legend we can ill afford to outgrow.

July 1, 2005: Do Spam Zombies Use Their Own DNS Data?

Darrin Chandler posed an interesting question to me last night, relative to yesterday's entry: Do spammers use their own cached DNS data to speed bulk mailing? In other words, do spammers look up an email address just once, and then store both the email and the resolved IP address together in a database somewhere? This might account for what happened to most of my spam when I changed Web and mail hosting providers last July. If there are widely used spammer address databases containing "old" IP information (that is, the IP address when my domains were still at Interland, as they were up until a year ago) much of my spam may still be going to the mail servers at the now-obsolete Interland IP address I used to have.

This would account for my current hosting provider's insistence that they're not doing any server-side filtering. As I've mentioned here a couple of times, my daily spam count fell by 80-90% as soon as I moved my several domains over to a new hosting account at Sectorlink. I haven't been able to account for the reduction at all until this possibility arose.

Remarkably, I can't find any crisp information online as to whether spammer address databases also include pre-resolved IPs. Darrin said he was guessing, but it was a mighty plausible guess, and if anybody out there knows anything about low-level spammer tactics (especially the nature of the email address databases that they buy and sell among them) I'd appreciate hearing about it.