January 31, 2003:
For the past couple of months, I've been chasing a mysterious message box that pops up at irregular intervals, generally in groups of two or sometimes three within a few minutes. After that I may not see it again for a day or two. All of my apps appear to be working correctly, and I'm wondering if there's some spyware lurking in my system somewhere that Zone Alarm is interfering with sufficiently to cause the box to pop up. AdAware sees nothing, and Zone Alarm is not reporting any peculiar attempts to access the Internet. Have any of you seen a box like this under similar circumstances?
January 30, 2003:

I don't spend a lot of time looking at the guts of the elaborate HTML messages the spammers are sending out, but recently I've examined a few to try and get a sense for how POPFile is making its decisions. I'm still clueless about that, but I've seen a lot of really interesting things. To defeat na´ve word filters, spammers are inserting HTML comments in the middle of words known to be spam-filter triggers, like "penis," "enlargement," and "unsubscribe." I suspect this is how some spam kept getting past my OE filters even though it appeared to contain text present in my delete rules. Here are some examples, the first two from the same message:

Male Performance Enhancer AND Pe<!--jeff--><!--jeff-->
<!--jeff--><!--jeff-->nis Enla<!--jeff-->rgement

To u<!--jeff-->ns<!--jeff-->u<!--jeff-->bs<!--jeff-->cr<!--jeff-->ib<!--jeff-->e

POPFile nailed this one anyway, but I suspect it would be a lot easier if it did an initial pass to strip out all HTML comments. I got a grim laugh from an image tag included in another spam that actually fooled POPFile, probably because it contained so little actual text:

<a href="http://www.zsupper.com/cl5/">Mortgages under 4.12<!--[!Anitspammers are faggots!]-->% ?</a>

I had to wonder if he misspelled "antispammers" deliberately, or if he's just a dumbass, albeit a lucky one. I suspect he won't be lucky twice.
January 29, 2003:

As of today, POPFile (see my entry for January 21, 2003) has had a week and a day to do its thang and demonstrate the power of Bayesian spam filtering. Just under 2000 messages have come down the pipe since then. My reaction?

Wow. It nails 99% of the spam I'm now getting, including the hard-to-catch "chickenboners" who use disposable free mail accounts on MSN, Hotmail, et. al. and hide their "spam-ness" by using as little text as possible and relying on graphics for the ads.

I get a lot of email, of which over 90% is spam. My email address is very old, which contributes, and as a third-shelf "celebrity" of sorts (due to my various publications) people occasionally post my email address on their Web sites, where the spam bots find it. I also got caught in the infamous Yahoo Groups "mistake" in which Yahoo flipped a bit allowing paid advertisers access to all Groups members' email addresses. So I need a spam filter way more than a lot of people. (Carol, for example, gets almost no spam at all, even though her email address is as old as mine.)

POPFile requires training, and you do the training by scanning a list of emails that POPFile presents, along with its decisions. You can override any decision it makes by clicking on a link, and the program then recalculates its statistical model to take the new classification into account going forward.

It was interesting to watch POPFile learn. The initial settings caught about 80% of spam out of the box. It improved steadily after that. There was a noticeable drop in spam activity over Superbowl Weekend (I laughed out loud when I made the connection) but it's climbing again, and I'm getting about two false negatives per day on volume of about 240 messages per. Most of what it misses now are messages with virtually no text to analyze, but which instead present their entire message in one or more bitmapped graphics.

There are still anomalies. Two days ago, I got two copies of precisely the same (typical) "horny housewives" porn spam message. One was tagged as spam, the other as legitimate mail. The first one to come in was tagged as spam, the second legit. Why? I just can't tell. And yesterday morning, a message from Michael Covington came in as spam (admittedly, the statistics were almost a coin toss, only 60-40 in favor of it being spam) for reasons that simply defy analysis. Last night Susan Hughes sent me her recipe (hand typed, not copied from somewhere) for Chicken Kiev. Spam. False positives like that still trouble me, and as good as POPFile has been in nailing the chickenboners, I still can't trust it to filter without supervision.

I'm going to give it another couple of weeks, if for no other reason than the fact that if it only misses two messages per day, there are few opportunities for it to learn. Improvement is thus on a sort of asymptotic curve, because the better it gets, the longer it takes for it to get better. Clearly, the program needs a whitelist feature (one that can import all the common address book formats) and perhaps some way to manually weight certain factors against spam. But overall, I'm astonished at its performance, and how easy it was to maintain, compared to my constant efforts blacklisting commercial spammers and my losing battle against the chickenboners. It's free, and I don't think you'll be disappointed, at least if your mail volume is as high as mine. If you want it, get it here—and let me know how it works for you!
January 28, 2003:

For various reasons I set programming aside in late 2001 and haven't done much coding since then. I hadn't been up on SourceForge since (I would guess) mid-2000, and haven't looked closely at what's going on there since 1999 or so. I went up there yesterday sniffing out possible Delphi implementations of Bayesian spam filters, and while I came up empty on that score, I discovered that TurboPower has already begun posting their code libraries under an open source license. As you'll recall from my January 9, 2003 entry, TurboPower Software is going out of business, and they're contributing over a million lines of Object Pascal code to the global open source community. I own about half of the TurboPower Delphi libraries, and they're simply stunning. Now I guess I can own the rest too, though I really wish it could have gone the other way. This link will take you to the Delphi trove index (great term!) from which you can scroll down to many of the TurboPower libraries, and search for the others. (Items are ranked by popularity. Most of the TurboPower tools rank very high.)

One of my New Year's resolutions was to go back and work on the Aardmarks Alpha version, and this has renewed my resolution to use more TurboPower components and dump some of the one-off freebies I've worked into the project over the years.

One interesting and impressive Open Source Delphi project is Phoenix Mail Roundabout, which is a very polished mail client written in Delphi that lacks few major features. (The only one I'd really miss is an HTML preview pane.) I had thought of writing a client called Aardmail at one point, in order to experiment with various things, and this may make that unnecessary.

If you don't haunt SourceForge on a regular basis, you're missing out.
January 27, 2003:

Another busy week begins, and I've got too much on the calendar this morning to go on at length about anything. So let me offer a link that my friend Frank Glover sent over, Explore Mars Now, which offers a beautifully rendered 3-D model of a proposed Mars habitat. You can move around in the model (within certain limits; it's not as "free" in terms of movement as Quake, say) and get a good sense for the details.

It's nice to see this. In the very last Breakpoint published by Visual Developer I proposed "RAD Mars", and it was something very like this on a much larger scale. I'm glad to see that near-photorealistic graphics are being done in this way, on this topic. If we don't have real Mars missions to evoke in us that irreplaceable sense of wonder that the Apollo missions provided in the late 60's and early 70's, perhaps CGI will have to do it. Maybe it's just a consequence of being old (and a little tired) but, hey—I'll take it and be glad of it.
January 26, 2003:

On my regular wardriving runs, I've noticed with concern (as have a lot of people) that probably 75% of wireless networks do not have encryption enabled. About 25% of the networks retain the default SSID (Service Set Identifier) value, and of those virtually all are unencrypted and "wide open." Backing out the 25% who clearly don't understand anything about what they're doing, there is still the peculiar 50% of wireless networks where the user knew enough to change the SSID but not enough to enable encryption, even though the two options are often side-by-side on the configuration dialog.

Maybe it's not so mysterious after all. A recent poll on ZDNet shows that, of people who have wireless networks, 50% cannot make them work with encryption, and thus leave encryption off.

50%. Hmmm.

The 802.11b wireless networking standard defines a 64-bit encryption system for Wi-Fi networks. Wi-Fi manufacturers, responding to the inherent weakness of that standard 64-bit encryption, offer 128-bit encryption systems and (more recently) 256-bit systems. The snag is that these "stronger" encryption systems are proprietary extensions to the 64-bit standard. There's no guarantee that 128-bit encryption from one vendor will work with 128-bit encryption from another vendor. (The situation is even worse for 256-bit encryption.) I've tried a lot of different Wi-Fi products from several different vendors in the process of writing Jeff Duntemann's Drive-By Wi-Fi Guide, and they all work happily together...using 64-bit encryption. My results using stronger encryption have been spotty.

My hunch is this: People price-shop for Wi-Fi gear and get whatever is cheap, even if it means assembling a network using devices from five different vendors. Then they try to get everything configured and working at the strongest level of encryption, and find that the encrypted connections won't happen. Rather than trying 64-bit encryption, they assume that "encryption doesn't work" and give up. I have noticed that on some Wi-Fi devices, the encryption level defaults to 128 bits. It's possible that people don't know what level of encryption they're trying to use, and of course mixing key lengths in a single network won't work even when everything comes from the same vendor.

My book will probably be criticized for my insistence that it's always best to build a network using devices from a single vendor. But it's the truth: Security is crucial, but strong security is non-standard. Until the IEEE 802.11i security standard finally happens (and that may take until the end of 2003) standard 64-bit encryption will be readily breakable. (Wi-Fi Protected Access may happen sooner. I'm still waiting for it.) But even after 802.11i, larger keys will remain non-standard. It looks like it's still going to be a mess for awhile.
January 25, 2003:

Were my father still alive, he would be 81 years old today. Alas, he died in 1978, after a ten-year battle against cancer that still gives me nightmares. All because of cigarettes. If I never discuss the tobacco settlement, smokers' rights (is this what they mean by, "the right to die?") and other related issues, it's because I won't even pretend to be unbiased. I'm not a good hater, but I hate tobacco like I hate almost nothing else.

I went to the Big Box of Jumbled Family Photos looking for a picture of Frank W. Duntemann that I could post here. I didn't want just any picture. I have lots of them when he was a young man, when he was overseas during WWII, and when he first married my mother. That, however, isn't as I remember him, and I wanted something from the mid-1960s, while he was still healthy and doing his best to make sense of the boy that I was. Trouble is, he had an expensive twin-lens reflex camera that my mother wouldn't touch, and thus for the multitude of photos we have from that period, he isn't in any of them. So the photo you see, from 1942, when he was 20 and in the Army, will have to do.

It's appropriate in one sense, actually: He was a capital-S Soldier, like so many of his generation, who had to face down Hitler and Tojo. He wanted me to be a soldier as well, and a baseball fan, and a marksman, and an engineer. He thought (as so many think) that sons are like fathers, and daughters like mothers. Alas, God cuts the cards and deals them His own way, and what Frank W. Duntemann got for his firstborn was not anything like what he expected.

He did his best with the spindly, daydreaming, unaggressive kid who spent most of his waking hours pounding on an ancient Underwood typewriter that Frank's own mother (my grandmother) had given him. He was certainly proud of me, and I think he eventually understood that his only son was far more like his mother than like his father. Nonetheless, of all the things he tried to teach me, two statements that he made still ring in my mind as having been pivotal in shaping the man I became:

"If you're lucky and smart, you'll marry your best friend."

"When you build 'em right, they fly."

Talk about nailing it! Everything good and useful in my life has proceeded from that. I only wish he had lived long enough to see just how good his advice had been.
January 24, 2003:

One of the joys of purging your book collection is actually looking at books you haven't touched in, well, ages. I had a number of Fifties-vintage hardcover SF novels targeted at teenagers, which (I dimly recall) I had gotten in a box back in Rochester for a dollar at a garage sale. All were pretty awful as works of literature, but I was struck by the diptych present on the front and back boards and flyleafs. Alex Schomburg deftly captured virtually every Fifties popular SF motif: flying saucers, astronauts, giant robots on the march, multi-armed aliens, finned spaceships, hapless civilians fleeing a city being destroyed, and over it all the brooding presence of the Mad Scientist. I'm dumping the novels (which really are awful!) but I scanned the diptych and show it above.

Alex Schomburg was the signature artist of Fifties SF, and to see more of his work, check out The Alex Schomburg Art Gallery. He's less well known as a comics artist, but in fact did more comics work than SF covers and illos. And while you're out and around the Web, see this speculation as to why SF went into eclipse, which certainly aligns with much of my own view.
January 23, 2003:

Odd coincidences happen all the time. PopTronics Magazine (the last descendent of Popular Electronics, which so many preteen Boomer geeks lived for in the 1960s) ceased publication not long ago, as several readers have written to tell me. My old friend Harry Helms was one of them, and he quoted Forrest Mims (who used to write cool little electronics books for Radio Shack) as saying that the hobby electronics era is over, and exists only in our memories.

It sure seems that way sometimes, but then the very same day as Harry's message arrived I got email from a gentleman who posted a largish site explaining how to build a one-transistor FM receiver using Radio Shack parts. It's similar to a circuit I designed eight or ten years ago, to prove that you could in fact slope-detect broadcast FM on the (broad) slopes of a superregenerative detector. Clearly, English is not the author's first language, but the site is nicely done and indicates that people are still out there tinkering somewhere. Nuts & Volts is still publishing, too, so I guess despair would be premature. What exists only in our memories is certainly the vibrant community of electronics and ham radio magazines that existed until fairly recently—there were a clean dozen or more worthwhile paper publications even through the early 80's—but there's still a lot of good stuff out there on the Web. Don't give up!
January 22, 2003:

Roe vs. Wade is 30 years old today. The January 18-24 issue of The Economist ran one of the most insightful articles I've ever seen on the subject—probably because it's a European publication, and the American press has basically placed the abortion topic off-limits for thoughtful discussion. The gist of the article is that the problems America faces regarding abortion stem directly from our failure to address the issue through the national democratic process. We left the matter to the decidedly un-democratic courts, and in doing so energized opponents in a way that would not have occurred had those opponents lost on the field of the legislative process.

The article's secondary point, that in America nothing is ever debated except as a death-match struggle between good and evil, might be the more significant one. Commentator David "Bobos" Brooks recently wrote (in the Wall Street Journal) of liberals with kitchens the size of Texas declaring jihad against conservatives with SUVs the size of Colorado.

What I've never seen anyone mention is that history in general is largely the cyclical swing of events and reactions to those events. The USSR was so traumatized by Nazi depredations during WWII (10 million or more Russians died) that it held the rest of the world at bay for 40 years. Only when a critical mass of younger people who never knew the devastation of the War came into power could Russia let go and join (to an extent that I admit is subject to argument) the rest of the global community. Modern Israel would not have happened except for the Holocaust, and many younger Israelis are increasingly objecting to the hard-line stance of Israel's older conservative leaders.

The almost incomprehensible rabidity of the pro-abortion camp is a reaction to the historical exclusion of women from American public life until mid-century, and their exclusion from meaningful work outside motherhood, nursing, and teaching until 1965 or so. The most violent supporters of abortion are those graying women who came of age in the late 1950s and just missed the boat of unlimited opportunity that younger women now take almost for granted. As they gradually pass away over the next twenty years or so, gruesome back-room abortions will gradually be forgotten, and younger women cannot be counted on to keep the fire stoked. Already, support for unlimited legal abortion is lowest among women in their 20's.

The real hazard of political action through the courts is that it falls mostly outside any kind of deliberative process, and gains can be erased as quirkily as they are sometimes achieved. There really isn't any explicit guarantee of privacy in the Constitution (as much as I feel that such a guarantee is desperately needed, to protect freedoms much more urgent than that of abortion) and if a couple more strict constructionists make it to the Court, abortion could be suppressed as quickly as it was elevated to the dubious touchstone of women's freedom.
January 21, 2003:

This evening I installed POPFile, a Bayesian spam filter. (See my entry for January 15, 2003.) The program is free, and comes with source, in Perl. POPFile is a mail proxy, which means it installs as an intermediary between your mail client (like Outlook Express) and your POP3 incoming mail server. You configure POPFile to talk to your incoming mail server, and you configure your mail client to talk to POPFile. What POPFile does is perform statistical analysis on the words in an email message, and calculate the likelihood of a message being spam. If the chances of a message being spam are greater than 50-50, POPFile inserts the spam tag (which you define; I use "[spam]" as suggested) at the beginning of the subject line. Other messages it can either lump together as legitimate, or further break out into categories as you choose.

I configured Outlook Express's mail rules to look for "[spam]" in a subject line, and route those messages to a special folder I created for spam. For a few weeks I'll watch it and see how well it does. I'm terrified of "false positives" and it will take a lot of experience before I can fully trust this thing.

But I have very high hopes. What is remarkable about POPFile (and the whole Bayesian idea) is that the filter learns, as you correct its mistakes. POPFile's UI contains a screen in which email messages are listed, along with its decision as to what category a message belongs in. If it decides wrong, you correct it by placing the message in the correct category. POPFile then recalculates and updates the filter for that category, taking the newly re-classified message into account.

In the two hours since I installed it, POPFile has classified 24 messages. It made two mistakes: It called one message spam that wasn't, and called one spam legit. The other 22 were bang-on. This sounds bad, but recall that POPFile is a statistical mechanism: It needs lots of data points before it classifies with "confidence." That will happen tomorrow morning. I typically wake up to 100-150 spams, often more. The spammers are getting exceedingly clever, and what little I can do in Outlook Express's mail rules they're increasingly getting around. I am very interested to see what it does with spams that consist almost completely of downloaded images. I'll report how it does in coming days. Cross your fingers; this could be the silver bullet.
January 20, 2003:

I just learned that Robert A. Heinlein's widow Virginia Heinlein passed away last week. Coincidently, I'm halfway through the video of the recent film Starship Troopers, which I should finish tomorrow. (I rented it, and I walk the treadmill two nights per movie when for whatever reason I can't walk the roads here.)

I met the great man once, at the SF world convention in 1976, and stood in line for a good long time to shake his hand and simply say, "Thanks." I never met Virginia, though I almost did, and for rather weird reasons: Carol's manager's wife was the realtor who listed the Heinleins' home near Bonny Doon, California (in the hills above Santa Cruz) in early 1989, and we thought about buying it. We could almost afford it, but I had just lost my job with Borland and it wasn't a good time to take on another $50,000 in debt. The house itself was spectacular: Round, all local redwood, with a bonsai garden under a dome at the center of it. The entire inner periphery was floor-to-celing bookshelves. We passed, and I'm glad we did, since it would have been a distraction on the path to Coriolis, which began to coalesce in the air between Keith Weiskamp and me about that time.

As for Starship Troopers, well, it aggravates me that RAH's weakest novel should have gotten such expensive treatment. Where are the film versions of Double Star or Citizen of the Galaxy? It's funny how people I've spoken to comment on "gratuitous breasts" in the film, but nobody bitches about the horrific blood, gore, and dismemberments. (Heinlein liked breasts; I'm not sure he would have approved of the gore.) The concept behind the novel is an eerie shadow of what actually happened to us in Vietnam: The old guys who had won The Big One were running America at that time, and tried to use the techniques that had won WWII in a conflict completely unsuitable for those techniques. So it was in Starship Troopers: The madness of sending foot soldiers onto a desert planet to attack giant insects that take twenty point-blank rounds to kill is breathtaking, especially since the Bugs were a hive mind, and there were no Bug "civilians." We could just as easily have turned the Bugs' own weapons—asteroids—against them, and melted their planet's crust to fifty meters' depth. But the vets of earlier wars had taken over, and they don't listen to ordinary citizens without military service. Although the novel ends with the war continuing, in my mind I always assumed that the Bugs would win ten or fifteen years down the road, because Earth would have spent all its wealth, will, and young people at the behest of graying ex-warriors who knew they were right and would tolerate no dissension.
January 19, 2003:

We've never lived anywhere as long as we've lived right here, on Lowden Road, and in consequence we've accumulated a great deal of Stuff. Carol and I have decided to literally reboot our lives, and thus we're shedding as much old stuff as we can bear to let go of. We're selling over half of our furniture. The hot tub and glider swing are already gone, as is our camping gear (we haven't camped out since 1986) and most of my classic (read here: junker) radio gear.

Our biggest push, however, is to purge our library. Both of us love reading, and we've never felt inclined to thin out the collection much. This time, however, we're putting the steel up our backs and thinning out the stacks with a vengeance. I have already set aside about three hundred books for donation to the Visiting Nurses for their annual book sale.

Much of what I'm dumping is terrible science fiction, some of it dating back to 1967. Most hadn't been read for thirty years or more; several of the paperbacks literally crumbled to yellowed flakes in my hands as I flipped through them.

This flipping-through was purposeful. Years back I had a bad habit of using credit card slips as bookmarks, and I don't want that sort of stuff going out with used books to VNS. So every book we're dumping gets flipped for bookmarks—and what a collection we found! Business cards, sales receipts, scribbled notes, homework papers, photos of Carol, sheets from the Lane Tech school newspaper, a nail file, a piece of a balsa wood glider, newspaper clippings, even a 5" floppy disk. At left is my best find: A copy of my very first ham radio QSL card, which I used for about ten months in 1973 and early 1974, when I was a novice. It's literally the only copy remaining; the box with the rest got dumped when we moved away from Chicago in 1979. I was ashamed of it then, but now, well, like a lot of klutzy old things, it's crossed the threshold from junk to personal history, and it will live on in our scrapbook.
January 18, 2003:

Earlier this afternoon, we got an offer for our house here in Scottsdale, and we accepted it. So the house is basically sold, though we won't be leaving for a few more months. A month or so ago we mentioned to our next-door neighbors that we would be listing our house for sale, and Don's jaw dropped. "Don't list it yet," he said, and a few days later he brought by an old friend of his who had wanted a place up in our neighborhood for a good long while. Our house is among the oldest and cheapest in this (newly) upscale neighborhood of $500,000 Santa Fe mansions, so it was not the kind of place that comes up for sale very often.

Our buyers' names are Frank and Vickie. This is eerie, as those were my parents' names, and I've never heard of another couple named Frank and Vickie...until now. It seems to be a good omen, as they love the house and will probably take great care of it.

I'm going to miss this place, where we've lived since the beginning of 1994. Our third-floor walk deck is pretty unique, with 360° mountain views and stunning sunsets, and I had a fabulous workshop in our detached garage. I had a 180-foot longwire antenna, though in truth my ground system was terrible (no soil, much less water in the soil) and the antenna didn't work very well. Most of all, I'm going to miss the poured concrete telescope pier, with electric service, that my friend Pete Albrecht and I put together back in 1999.

However, the astronomy is getting bad here, between the air pollution and the light pollution. Five thousand homes and a new high school (complete with kleig-lit sports fields) have been built within a couple of miles of us in the past few years, and I can no longer see the Milky Way on a regular basis. The traffic is appalling, and "going south" (to Phoenix and Scottsdale) is now an ordeal that we undertake only when we can't avoid it. Whether it's global warming or not, the summers are getting longer (or perhaps we're simply getting tired of them) and my dust allergies continue to be a nuisance.

Time to go. If we come back to Arizona, it will be to a condo somewhere in the middle of things. But one thing at a time. Colorado beckons, and getting from here to there will not be trivial.
January 17, 2003:

I don't know if this is a trend, exactly, and if it is a trend I don't know whether it's a good thing or not, but cigar smokers are developing a taste for zinfandel wines. There is a tradition going back to G. K. Chesterton or beyond, linking cigars and port wine at the end of a long day. Port wine is a strong, and (usually) sweet red wine, with alchohol up in the 16-18% range. You can get cheap port, of course, just like you can get cheap anything, but good port is expensive, and (assuming you like port at all) worth the cost. Some cigar freaks are now drinking zinfandel instead of port, for reasons I haven't seen explained. They favor the sweeter zins (I wonder if any have discovered Coturri's Chauvet vinyard, see my entry for September 23, 2002) and some labels are starting to cater to the cigar crowd. I picked up a bottle of one of these new and somewhat controversial semi-sweet, pretending-to-be-a-port zinfandels the other day...and damn, I like it!

The wine is Cosentino Winery's Cigarzin 2000. 14.6% alchohol. Very dark red, unfined, extremely fruity (which is good in my book—as I've said before, wine is made from fruit and that's what it should taste like) and about as sweet as a white zinfandel, a degree of sweetness that makes some people gag, but which I still consider semi-sweet. (The problem with white zin is not its sweetness but its utter gutlessness.) I had planned to take a picture of the bottle, but it's one of the least distinguished wine labels I've seen in some time, so if you want to give it a shot, just call around. I don't think it's quite as obscure as Coturri.

A lot of wine fanatics are rabidly anti-smoking (as I readily admit that I am) and decry the linking of "good" wine with something as base as tobacco. On the other hand, if it brings a touch of sweetness back to strong, complex red wine I'm all for it. I like dry wine too, but guys, a little diversity is a good thing! Cosentino Cigarzin 2000. $25 range. Highly recommended.
January 16, 2003:

Michael Covington sent me a pointer to a site dedicated to the Klein Bottle concept, and sells blown-glass Klein Bottles. (Ok, ok, you extradimensional math freaks, I know it's a three-dimensional projection of a four-dimensional object!) I loved the Fourth Dimension as a high schooler (I'm sure my teachers thought I spent a lot of time there) and the site is delightful. My favorite is less the glass bottles than a Klein stocking cap (photo at left) of all things. My mother used to knit me stocking caps while waiting for me to come home from dates when I was a teenager, and I wore them a lot. As I'm about to return to a cold climate, maybe I should get one.

The site's proprietor is named Cliff Stoll, and though it's supposedly the same Clifford Stoll who used to dis the Internet as a second career, I've never met the man and don't know what he looks like. (The guy wearing the hat above is kleinbottle.com's Cliff Stoll.) I gave away my copy of Silicon Snake Oil or I'd match the author photo. The site itself says nothing about whether the hat-wearer is the "Snake Oil" Cliff Stoll. (I'd guess that it is—after Silicon Snake Oil, I'd guess being taken seriously in computing is harder than it used to be...)
January 15, 2003:

The American Radio Relay League (ARRL) has put together a task force to adapt 802.11b Wi-Fi technology for use in amateur radio communication on the 2.4 GHz band. The ARRL's Web description of the project is here. Most non-hams (and a lot of hams, for that matter) have no idea that there is a ham band overlapping the Wi-Fi frequency allocation at 2.4 GHz. The idea is to take ordinary off-the-shelf Wi-Fi gear and use it for ham radio packet communications. Because Wi-Fi Channel 6 (2.437 GHz channel-center) is out of the way of existing ham radio activity on the band, it's the channel of choice for early experiments by the ARRL task group participants.

I'm generally a rabid supporter of ham radio, but this one has me a little nervous, for the simple reason that Wi-Fi channel 6 is the default channel for many access points and wireless residential gateways (including those of market leaders Linksys and D-Link) and my own wardriving research indicates that Channel 6 is by far the most popular Wi-Fi channel of all. The Amateur Radio Service is not govered by Part 15 of FCC rules, as Wi-Fi is, and hams are allowed as much as 10 watts of power (!!) on the 2.4 GHz band. Crank up a 10-watt packet station on Channel 6, and there goes every Linksys or D-Link-based Wi-Fi network (which would be most of them) within half a mile or so.

Gonna be an interesting project, and worth watching.
January 15, 2003:

Our best hope against spam may be trainable Bayesian filters, as brilliantly explained by Paul Graham in his much-linked paper on the subject. The idea is simple: Have two "delete" buttons on your mail client. One deletes ordinary messages, and the other deletes spam. Each time you delete a message as spam, the filter "learns" by performing a sort of statistical analysis on the text, and creates a statistical model to compare future messages against. Graham's code is in LISP, which ties my head in knots, so I'll freely admit I don't understand how it works. However, the idea sounds really good, and according to Graham, the module currently misses 5 spam messages out of 1000 spam messages, with 0 false positives. 0!

I'd buy that for more than a quarter!

Trying to sell a mail client is dodgy, since so many are free and the rest are old and dug in, like Eudora. Nonetheless, spam is getting to be such a nuisance that I think a commercial client that specializes in spam management might be worth a shot. In addition to the Bayesian filter, it needs to optionally switch off graphics fetching (which is used in email address "beacons" to verify emails because the mail client fetches an encoded graphic embedded in the message) and maybe a few other things. But I can't imagine it would be all that hideously hard to clone Outlook Express in Delphi and add the filter and associated logic.

Certainly, if anybody sees an implementation of Graham's Bayesian filter logic in Pascal I'd appreciate a pointer. Graham's site has pointers to a few implementions in "normal" languages like Python and Perl, and even one in VBScript, so if I get ambitious I may try the Pascal port myself.

And while we're talking spam, can anybody explain why spammers are increasingly placing hundreds of random characters somewhere in the message (usually at the bottom)? I assume they're trying to evade some sort of spam trap, but I don't know which one.
January 14, 2003:

Amazon has done it again: Today I discovered the Amazon Honor System, a way for people to request small donations via their Web sites, routed through Amazon's 1Click payment system—with Amazon taking a cut, of course. The cut is 15% of the donation, plus 15 cents. Unlike Amazon Marketplace, which I used so successfully to sell my own books (and a few others) this fall (see my entry for October 31, 2002) I have no current need for Honor System, but if I were to try and make money distributing something like my Aardmarks bookmark manager program (assuming I finished it) I might give it a shot.

In fact, I'd be tempted to finish Aardmarks just so I could "sell" something via Honor System to see how well it works.

As with Marketplace, it's a brilliant move by Amazon. They have this magificent e-commerce system, honed over years of experience, and now they're letting small-time operators use it, and shave an effort-free slice off each transaction. It's not quite micropayments as I'd like to see them, but it's a helluva move in the right direction. As I already have an Amazon account for Marketplace, I may sign up for Honor System anyway, to get a sense for it. It's certainly easy enough.

If any of you have used Honor System, I'd be interested to hear how it works for you. I'm going to post a pointer to it on some of the Old Catholic listservs, as it sounds ideal for use as a virtual collection plate.
January 13, 2003:

I was out digging around in the garage earlier this fall, when in a box of old papers I came across a song sheet from a Catholic high-school retreat I had attended as a senior in 1970, at Divine Word Seminary in Techny, Illinois. I had been searching for the words to a song called "Easy Come, Easy Go" for some time. After posting a query here in my November 19, 2000 entry, someone sent it to me in an email, and then I lost the email in one of my periodic purges. The song sort of defines The Sixties for me, back when church retreats were more hootenanny than theology, and I'd gladly buy a recording if I knew of one, since all I can do is remember unsteady teenage voices singing it in a drafty old red-brick seminary, not really understanding how completely young and carefree they were.

Well, I pulled the sheet out of the box, and there it was. And even though it was on a fragile and yellowing sheet of cheap pulp stock, it slapped down on my scanner and came in without a struggle, thanks to Fine Reader Sprint. So let me post it here. It's got a very nice melody, and if anyone knows its author, please drop me a note.

Easy Come, Easy Go

Refrain:
Easy come, easy go, through summer and through snow.
Up and down, all around, this universe I go.
And I'll walk upon your waters, move mountains from your path,
With a smile for my companion, I'll teach you how to laugh.

'Cause I've ripped off my mask and kicked down my walls,
Strapped up my boots to answer His call.
You say that I'm a roamer, got no place to call my own,
With a thousand places I call home, I know I'm not alone. (Refrain)


I've thawed out my body, I've put blood into my veins,
'Cause people were cryin', they were callin' out my name.
You say that I'm a poor man, got no one to care for me;
With a million brothers and sisters, how rich can one man be? (Refrain)

My sister is in Boston, my brother rides the rails,
My cousin lives in China, my uncle brings the mail.
I've met with kings who've passed away, but I know they still care.
My father lives in heaven: Baby, He's a Millionaire! (Refrain)

Saw a light in your window, thought I'd pause and take a rest,
Say hello and good-bye and wish you all my best.
I'm just here on a visit, I may not be back your way,
'Cause when my journey's over, I'll be going home to stay. (Refrain)


January 12, 2003:

It's done. Really. Well, ok, the writing is done—I still have to read page proofs and do all that other stuff, but this grueling schedule of 4000 words a day, every day, is over. I opened a bottle of expensive wine (for me, that's $25) and treated myself to a bratworst, and (later tonight) a good night's sleep.

It's sobering to ponder that this is really only the fourth book I've written solo. (My unpublished SF novel doesn't count.) I've done portions of a dozen books at least (lost track of them) but only four are mine alone:

  • Complete Turbo Pascal has been through four editions, and was in print from 1985-1994. (The fourth edition was titled Borland Pascal 7 From Square One, but it was the same book with a makeover.)
  • Turbo Pascal Solutions is my least-known book, and was around for a little while in the fall of 1987. It died because it was about Turbo Pascal 3.0, and appeared just as Turbo Pascal 4.0 came out, and changed everything. I guess that's just the game we play.
  • Assembly Language Step By Step is by far my best seller, now in its third edition, though the first edition was called Assembly Language From Square One. It first appeared in 1989 and it's still in print, and selling 5,000-7,000 copies per year.
  • Sometime about February 15, 2003, Jeff Duntemann's Drive-By Wi-Fi Guide will roll off the presses and into history. It's much loopier than any of my earlier books—I explain how to build networking equipment out of kitchen trash!—and also the first one of interest to the general computing public and not strictly the programmer nerd crowd.
I guess I'm not Steve Holzner. On the other hand, with one exception (poor Turbo Pascal Solutions) I've defied the conventional wisdom that computer books live for six months and vanish. We're gonna do it again, too. This one's going on national TV. You just watch.
January 9, 2003:

Only a couple more days—oremus. I've been heads-down here, doing three or four thousand words a day, every day, for about a week, and that doesn't leave me much energy to be brilliant in this space. However, this morning reader Richard Biffl informed of something that caused me great sorrow: TurboPower Software is going out of business.

TurboPower founder Kim Kokkonen (who told me that people often assumed he was "a small Asian woman") was one of the very first industry figures I ever met, in early 1985 as a cub editor at that great metropolitan nerds' magazine, PC Tech Journal. He and TurboPower staffer Brian Foley taught me a lot about how Turbo Pascal worked internally, knowledge I later applied in writing my shortest-lived and least-known book, Turbo Pascal Solutions. We used to hang out with them in Scotts Valley, when they were there and we were there, and I had had hopes of hanging out with the TurboPower gang again (I know it's a radically different gang now) once Carol and I moved to Colorado Springs later this year.

The sole ameliorating factor in all this is that TurboPower will be releasing many or most of their products into the open source free software arena, under the Mozilla 1.1 license. There are some (unexplained) legal hurdles to releasing some of the products, but they're going to turn loose as many as they can. The products will be hosted by SourceForge and they truly will be free.

It amounts to a cool million lines of Object Pascal source code. Wow.

As I've said many times, elsewhere and elsewhen, there has never been anything like TurboPower. From the beginning, they gave the lie to the BS about "we can't reveal our source or people would copy it!" All TPS products developed in-house have always come with full source. It was from TurboPower's example that I became convinced that people hide their source because they're ashamed of what a mess it is, not because they want to "hide their secrets."

Unless, of course, the secret is that you're a shitty programmer. That's a secret worth hiding. And that is not anything we ever needed to worry about with TurboPower products.

I don't know what this will do to the Delphi development arena. I am sticking with Delphi, even if I have to learn Russian to buy new VCL components. I've been a Pascal guy since 1979, and I do not intend to stop now.

By the way, I have tried and failed to locate Kim Kokkonen and Brian Foley. If any of you know how to reach them, please tell them that Carol and I are moving to Colorado Springs, and if they're still in that vicinity we'd like to get back in touch.
January 3, 2003:

I'm closing in on the book—should be done within a week, whew. This morning I have a question for you: I've heard of SOHO router/gateway appliances with built-in 56K V.90 analog modems, so people who don't have broadband can still share a dialup Internet connection through their LAN. Trouble is, I can't find a real example. Can anybody send me a pointer to such a modem-equipped SOHO router/gateway appliance? Linksys doesn't have one, nor D-Link nor Belkin, at least as far as I can tell. I'd cover such a gadget in my book if I knew it actually existed!

And if it doesn't, well, I guess it means you have to use Windows ICS and leave your main box on all the time, which is a bummer. Or?
January 2, 2003:

My Lane Tech Astronomical Society cohort Pete Albrecht is coming out a little later this winter, bringing his big new Meade "go-to" telescope for us to play with. It's brought back to mind the firestorm that erupted in the astronomy geek press and Web forums when go-tos first appeared a few years ago: Will go-tos completely ruin observational amateur astronomy?

Good question. A go-to telescope is a heavily computerized device, generally constructed on a motorized altazimuth (rather than equatorial) mount. The better ones have a built-in GPS receiver so that the scope "knows" where it is and what time it is when you set it up. You center two easy-to-locate stars in its eyepiece, so that the scope "knows" where it's pointed, and can then locate any other object in the sky with standard coordinates, automatically, no searching required.

Select an object like M33 (the great spiral galaxy in Triangulum) on the LCD control handset, push a button, and whirrr-whirrr-whirrr, there it is, centered in the eyepiece.

Doubtless you can see more objects with a go-to scope than the Old Way, which involved starting with a known star and then guessing how far in each axis it was to a faint galaxy or star cluster, and then frantically sweeping back and forth until you find the object or give up and look for something else. Setting circles help...a little. Having a clock drive helps a little more, since without a clock drive one of your setting circles gets out of date second by second as the world turns. I've logged over half of the famous Messier deep-sky object list (a collection of 104 galaxies, nebulae, and star clusters) doing it the hard way. It's taken me almost 40 years of intermittent observation. With a go-to scope you can log half the Messier list on a dark night in a couple of hours.

So where's the skill? The challenge? Does a go-to scope change the night sky into a kind of slide show? Does it get boring after awhile? I just don't know. Certainly ham radio lost a lot of its fascination for me once the radios got broadband and mostly automatic. I'm nervous about whether a go-to scope would ruin my most truly ancient fascination, which dates back to 1957, when Sputnik happened and five-year-olds everywhere suddenly went nuts for space and astronomy. On the other hand, having tried literally dozens of times, I have never been entirely sure that I actually found M33, which while very large (most of the size of the full Moon) is also very faint and fantastically difficult to see.

In a few weeks I'll find out. I'll keep you posted.
January 1, 2003:

Hey, what the banner below sez! Let us hope—or pray. It will certainly be an interesting new year, on both a personal and a more cosmic level. Carol and I will actually see the house we've designed for ourselves become reality, and we'll move to a whole new city, state, climate and...altitude.

Everybody I know is chewing nails about the war against Iraq. Stop, already. It'll either happen or it won't, and nothing we think or do will change it. All the panic about the "Arab street" rising up against us is nonsense. When we knocked over Afghanistan in about 20 minutes last year, the Arab street hunkered down and began wondering if it was its own worst enemy, which in my view was a damned fine thing. Those guys understand nothing but shows of force. The real reasons not to invade Iraq involve loss of Iraqi life and the risk, effort, and expense of trying to make Iraq into a sane nation afterward. Maybe that has to be done, but I sure wish somebody else would do it.

The real issue to watch in 2003 is health care. More and more people are being dumped out to forage on their own for health insurance, and health insurance can cost eight or ten thousand dollars per year. If you already have some significant medical condition, forget it—you can't get insurance at any price. Pressure is building for a national system to deal with this, and there's no time like a recession to do it.

What may finally kill private health insurance, however, is genetic profiling. More and more tests are being developed for genetic predispositions to certain serious, high-cost diseases—breast cancer, MS, things like that—and insurers are fantastically eager to use them. The idea, of course, is to avoid insuring the people who will most need medical care in the future. This makes excellent business sense, but conservative industry talk to the contrary, medicine is not a business in the same sense that food or clothing is...especially once we begin seeing health as something you're simply born with, as genetic tests imply. The Democrats are desperate to find an issue to use against the Republicans, and this is as good as they're going to get in the near future. 2003 will be a key year in this debate. Keep an eye on it.