|
|
 |
|
It was actually good that we were running around all day; the power was out most of the afternoon, and the surge somehow got through my surge protector and forced my new Linksys Wireless-G router to revert to its factory defaults. Worse, it fried the Wireless-G PCI card in my Xeon system, and wiped the CMOS on Carol's old Compaq Deskpro EN. This was more computer mayhem than we've seen in one day in a good long while. All systems still boot, but I have to wonder what was weakened and may go next. We got home late yesterday afternoon to see a strange trampoline wedged against our neighbors' house immediately across the street, and garbage everywhere (mostly shredded Christmas wrappings) from blown-over cans. The closest trampoline Carol and I have ever seen on our walks here was two blocks down and a street over, so where this one came from is a mystery, but I for one would have given much to see it travel. Oddly, in all the writeups on Colorado Springs that I chewed through while researching places where we might relocate, nothing said a word about high winds—and yet we've seen more damaging winds in our nine months here than we have anywhere else we've ever lived. Maybe it'll be more sheltered up the hill after we move. Let us pray. |
|
|
Got much wandering to do today, so this morning I'll have to settle for emptying the odd lots file:
|
|
|
A couple of additional Wi-Fi notes concerning things I learned since yesterday morning:
|
|
|
I spent most of yesterday installing, configuring, and testing the Linksys Wireless-G gear that I bought two weeks ago. There were no surprises, really, and everything worked almost flawlessly. I figured you might be interested in some of my observations:
|
|
|
Several people asked for pictures of "the rock" (see yesterday's entry) so I went up there at lunchtime and took a few shots. The one shown below gives you an idea of how it's resting on the ground on a tapered vertex. It's potato-shaped, and comes up to the top of my chest; figure 4 1/2 feet high and maybe 7 feet long. It's hard to tell from the photo, but the ground at that point is about a 10% grade, sloping downward toward the left. The view is looking down the rock's long axis:
The second shot gives you some idea of where the rock would go were something to nudge it into motion down the hill. It would begin rolling along the street, but eventually it would wander to one side or another, and there are plenty of handy targets along the way. At some point, however, even if it stuck to the street, it would go through some townhomes that you can't see in the photo:
|
|
|
Go shopping today? Are you nuts? I'm staying home where it's safe and getting some work done. In the meantime, some odd lots:
|
|
|
A
blessed Christmas to everyone out there! Carol and I are taking it easy
today, and later this afternoon will attempt to roast our very first duck
in all our 27 years of marriage. The weather is insanely gorgeous (55 degrees
already!) and while it's not a white Christmas, it's a fine day nonetheless.
So stop surfing the Web already and go show your loved ones that they really
are loved ones. Whether you're a theist or not, please celebrate this season
of love and unselfishness, for on that thread hangs most of our claim to
being human.
|
|
|
Christmas Eve. My three favorite saints in the Bible are St. Dismas (the Good Thief), St. James, and St. Joseph. I'll speak of James and Dismas in due time, but this morning, let's consider the story of that guy in the (mostly forgotten) corner of the church, with a saw in one hand, and toddler Jesus under his other arm. The man had an iron devotion to his wife, and a species of courage I don't think most of us understand anymore. We who are not Jewish and don't read the Torah are generally unaware that in Joseph's time (and he was an observant Jew, after all) if your fiancee got herself pregnant by some other guy, you were not only permitted but obliged to turn her over to the local authories to be stoned to death. (See Deuteronomy 22:13-21.) Joseph knew this, but he would have none of it; he genuinely loved Mary and although he considered a quiet divorce, he changed his mind (with some help from an angel who came to him in a dream) and not only married her, but raised her child of mysterious fathering as though He were his own. This is so damned contrarian that I have to grin: My kind of guy! Had he interpreted his Scripture literally, Mary and the unborn Jesus would have been toast. Is there a lesson here? Obviously: The law was made for Man. Man was not made for the law. (This does not mean, as some have implied, that there is no law.) When the law goes awry, love and guts were made to bring it back where it belongs. The story of Joseph is not the story of a Divine Plan being imposed from above, but of a good man following his heart in the direction where he knew his God was leading. Carol and I aren't able to get back to Chicago and be with family this year (oh, the house project!) but we hope that all of you will remember the importance of family amidst the jangle of an overly commercial season. It's not about Stuff. (I'll turn my attention to the issue of Stuff shortly.) It's about love, and the guts that makes love real. |
|
|
No one weapon will win us the war against spam; not even authentication, though authentication will at least make victory possible. I have a suggestion for another really powerful weapon that would be easy to implement and to which there is no conceivable (at least conceivable by me) objection by spam's victims. Companies that register domains, at least with Network Solutions, get unique identifiers. (Mine is QWRBETJKRO.) Currently, you have to go through a robot filter to identify who owns a given domain, and I'm fine with that. However, what if there were a simple lookup that did nothing but associate domains with a given customer ID? In other words, I'd like to be able to search on one of those 10-leter ID codes, and retrieve all domains owned by that ID code. I don't care to know their names or addresses. With a system like that in place, for those cases where one company owns hundreds of spammer domains, if I spot one, I spot them all. Bigtime direct marketers, of course—those who bother to register domains at all—will howl bloody murder. (Chickenboners would not be affected.) But there's really no privacy violation here. Ironically, the entity that will fight this most savagely is Network Solutions itself, for this reason: The more domains people register, the more money they make, and if it no longer avails spammers to register domains by the dozen, well, Network Solutions loses money. That's the main reason I can't see this going anywhere, but if you're ever in a position to suggest it, hey, why not?
So there. This album maxes out the CD format at 70 minutes, and contains not only Karen and Richard, but superb orchestral and studio chorus work as well (coordinated by the estimable Peter Knight of Steeleye Span and Shannon) which could stand on its own even without the Carpenters as leads. The album is thus not 70 minutes of two people singing different songs, but rather a remarkably diverse collection of approaches to Christmas music, from pop to classical to (soft) rock to a sort of Fifties night-club ambiance. All are Christmas standards, both sacred and secular, though one or two (like "Little Altar Boy") hover on the holiday fringes. By now I've heard it over a hundred times, I'm sure, and somehow I never get tired of it. I'm not entirely sure how they did it, but they did—and I doubt that anyone else will ever do it again. Not everyone knows this, but Karen Carpenter died twenty years ago, in 1983, having starved herself to death in a lifelong battle with her perception of her own body. It's easy enough to say that she was just crazy, but having seen some of the viciousness that we heap upon overweight people, particularly women, well, in an industry that elevates looks to the highest good, I can almost understand her desperation. (She was "chubby" as a pre-teen.) Maybe someday we'll come to understand that none of us were born to be beautiful, but all of us were born to get results, in accordance with our talents, whatever they might be. |
|
|
It's snowing...and we're going out for a little last-minute Christmas shopping. So if I can't be profound this morning, pray that I can be at least interesting, with a few odd lots on various topics:
|
|
|
One of my friends fell for the "Free Smileys!" spam scam that peaked a few weeks ago and still trickles in from time to time. The smileys are free, but by installing them you also install a piece of spyware that's extremely difficult to remove—and it will report your Web surfing habits back to some server company that sells the statistics to spammers and other Internet marketers. Needless to say, never respond to spam and never click on a popup ad, and if you're smart you'll keep your machine behind a NAT firewall and Zone Alarm Pro as well. (You have deselected "Install on demand" in IE's Advanced Options, haven't you?) But that's not my point here. Given all the spyware, malware, spam, worms, and RIAA lawsuits that have been haunting the Internet recently, I'm guessing that over the coming years, the Internet will more and more "go dark"—that is, will descend under a cover of strong encryption and become a fragmented tangle of almost entirely insular mini-networks, passing messages and trading files among their own members, who only come out from undercover to read an occasional Web page from behind a very strong firewall. Anonymity will mutate into something entirely different from what it is today: Email will become strongly authenticated, and file transfer will become strongly obfuscated, though protocols like those emerging in projects like WINW (WINW Is Not WASTE) and MUTE. I've heard quite a bit of moaning and robe-rending about this, and how in the future the Internet will be locked down, and no transaction will occur without the authenticating (and recorded) permission of government. Oh, the irony: What this actually means is that the Internet will be returning to its roots as a peer-peer architecture connecting a bunch of privileged insiders. The Internet was not intended to be a client-server system (as it quickly became after the Web conquered all in the mid-90s) and what we're seeing may be in fact the hand of evolution placing Internet functionality back into an appropriate architecture: C-S for e-commerce, and P-P for interpersonal communication and, um, "sensitive" file transfer. You can argue whether this is good or bad, but I'll argue that it's inevitable. People will allow the government to regulate e-commerce and to some extent email routing, at least in terms of authentication. Everything else (including email content) will be very dark. It will be a very different world than the one we see now (and safer) but how its details will play out I suspect will surprise us all.
|
|
|
A few odd lots on this very (climatically) odd (63 degrees here in the mountains!) day:
|
|
|
Within just a couple of days, I ran across this brilliant article from Clay Shirky, and word of MUTE, an encrypted and decentralized file-sharing network designed specificially to avoid RIAA-style snooping and lawsuits. Shirky's article is a must-read; his point is simple and significant: That while the goading of the Cypherpunk movement to get the general public to broadly adopt encryption for all communications came to nothing, the criminalization of file sharing looks to be doing the job beautifully. Shirky makes the point that Prohibition, which criminalized a popular and formerly legal activity (drinking) hugely increased the distrust of government, and accustomed Americans to be private and secretive to an extent that was unthinkable before the passage of the 18th Amendment. Something similar is happening today. As Shirky says, "...to a first approximation, every PC owner under the age of 35 is now a felon." Before, people had little to hide, and thus hid little. Now, with so many online activities inspiring government suspicion or even legal action, ordinary people (especially the young) are adopting encryption for communications in huge numbers. This may not objectively be either a good or a bad thing—what's important to note is how it came about. We can only wonder what its ultimate effects will be.
|
|
|
I subscribe to four email groups on the Yahoo Groups server, and about a week ago I got a note from Yahoo Groups about a change in policy. It took me a few close readings of the densely weasel-worded document to discern that they were basically claiming the right to give my email address to spammers, and if I didn't like it, I could always cancel my Yahoo account. I set it aside for lack of time to deal with it, but I did notice over the week to follow that my spam count went from 650-700 per day to about 850. More tellingly, the number of new spammer domains that I've not seen before exploded: Earlier I was blocking 8-10 domains per day; recently it's been as many as 30 per day. (Poco's junksender.txt file now contains 2,354 blocked senders, virtually all of them domains known to be owned by spammers—an appalling number of them having appeared in the past thirty days.) Apart from the fact that I'm trying to make time to get back to work on my Aardmail spam filtering utility, I did about all I could do: I cancelled my Yahoo account. And oh, the inhumanity: They said (after the fact!) that it would be cancelled sometime within the next 90 days. So let's just say that helium will freeze in Arizona before I deal with them again, and if you have a spam problem, I encourage you to make sure you don't have a Yahoo Groups account, perhaps one you've forgotten about.
|
|
|
Much is being said by others about the Wright Brothers, who first flew 100 years ago today, and about their rival Samuel Langley, who might have flown first if he weren't obsessed with launching a plane by throwing it off the top of a houseboat with a catapault. Less is being said about two of my very favorite crackpots, Alberto Santos-Dumont of Brazil, and Clement Ader of France, who came very close to beating the Wrights at the flying game, and (to my view) had more fun and much more interesting flying machines. Ader, in particular, exemplified the Victorian spirit of Jules Verne in his ornate and bat-like Eole and Avion machines, both of which were completed before 1900. (Click on his link above and scroll down for pictures.) The Eole supposedly flew about 150 feet in 1890, and some in France claim he was the true inventor of flight. Alas, the steam engine he was using (zut alors! steam!) didn't have the power to keep the craft aloft for very long, but most experts I've read think that the airframe itself was sound and with adequate power could have flown long before 1900. Santos Dumont was a playboy eccentric who used to fly around Paris in his personal dirigible, tying it to lampposts and clambering down to the street on a rope ladder to have dinner at his favorite haunts. His oddly-shaped 14-Bis flew as well as the Wrights' Flyer, but he was a couple of years late getting it into the air. My only point here is that that race to manned heaver-than-air flight was a fantastical saga of crackpots, visionaries, and amazing rump engineering, and although I'm proud that the Wrights were Americans, I have to grant that they were dull, grumpy, and secretive, hardly the stuff of legend. For the whole saga told as a ratlin' good story, I recommend (if you can find it) The American Heritage History of Flight (1962) which I had as a kid and read again and again until it fell apart. Ader, Langley, Santos-Dumont, Hiram Maxim, and Glen Curtis were underdog heroes to me, and I dreamed of roaring across the Atlantic in a refurbed and enlarged steam-powered Eole. History is fun! (It takes college professors to make it dull, and they work very hard at it!)
|
|
|
As I begin work on the second edition of Jeff Duntemann's Wi-Fi Guide (we're dropping the "Drive-By" for various reasons that I'll discuss at some point) I went shopping for 802.11g equipment last week, expecting only to note prices. Wow. I bought. CompUSA had matched Amazon's prices to the penny, and I scored a Linksys WRT54g wireless router for $89.99. The Linksys PCI client (WMP54g) and PC card client (WPC54G) were likewise going for Amazon's price of $69.99, so I threw 'em in the cart. I like buying locally if the prices are good, so that it's easy to return faulty merchandise (and perhaps keep the retailer in business) and I hadn't seen significantly lower prices from any online vendor that I trust. These prices are amazing, and I think this is a damned fine time to upgrade an 802.11b network—not for the additional speed so much as for the WPA (Wireless Protected Access) security technology that most of the new 802.11g products are sporting. More on the new gear as I implement it on my network.
I'm very partial to multi-artist anthology discs, since it's pretty boring to hear one singer apply the same vocal heuristics to the Christmas canon for over an hour. This very full (63 minute) CD is admittedly dominated by the King's Singers, but also includes the Taverner Consort and the Empire Brass, both of whom are superb. The music is all traditional, and there are no Santa Claus songs here: They're all Baby Jesus Songs, as somebody put it, which in my view is not a bad thing. Alas, this 1989 title is out of print, but I've seen it many times on the used sites. |
|
|
|
|
|
Gaudete Sunday. Advent is an odd season in the Catholic world these days, though not as odd, I might argue, as the shapeless mess they call "Ordinary Time." When I was a kid, the mood in Advent was somber and almost penitential, like another, slightly shorter Lent. As with Lent, when you were most of the way through all the somberness and penance, they let you out for one Sunday. The vestments and altar trimmings were a bright rose instead of pentitential violet, and the Intriot (opening prayer) began with the uncharacteristic word, Gaudete! (Latin: Rejoice!) We kids would joke that "Gaudete!" was Latin for "Hang in there, guys—Christmas is almost here!" (Great economy of expression, that Latin stuff...) There was a similar Sunday most of the way through Lent, called Laetare, where rose vestments were also used, and we started searching the house for hidden Easter eggs. These days, the Catholic liturgical calendar, and the reliable march of liturgical colors through the seasons of the Church, is in disarray. Churches no longer use rose vestments and altar dressings—why buy a set of expensive fabrics that you use only twice a year?—and instead of purple, some Catholic and Episcopal churches have begun using deep blue vestments, representing the color of the sky just before dawn. That's actually kind of cool, and a little more indicative of expectation rather than penance. (Lent is penance enough, thank you!) Unfortunately, the Vatican never approved the use of blue as an Advent litugical color, and there have been some shouting matches over that. Me, I won't quibble over blue or purple for the bulk of Advent, but I miss those two flashes of joyful rose.
|
|
|
Egad. Just a couple of minutes ago, they got on the news and told us that they snagged the Ace of Spades—Saddam himself, sitting in the bottom of a hole somewhere twiddling his thumbs. It wasn't a firefight, nobody got shot, and somehow it seems almost anticlimactic. So...what now? Who gets to try him? And what penalty can a court exact? I'm not a big fan of the death penalty, but there's a complication with people like Hussein, who have rabid followers: As long as he's alive, his thugs still at large can take Western hostages—ideally, Americans—and try to bargain for his release. That being the case, I might be inclined to let his fellow Iraqis have him, as I suspect their idea of justice in this case would not be something I'd want to watch on TV...and if I were Saddam, I'd be begging the American forces to ship me back to the US. |
|
|
I've been following the recent success of the Lindows distribution of Linux. European and Asian governments have been adopting Lindows right and left, which is interesting, since as best I can tell it's not as sophisticated a desktop as XD2. (I'm about to start evaluating both distros in detail, and you'll see my comments in this space.) Microsoft has been going after Lindows very aggressively, at the behest of MS wildman Steve Ballmer, who seems to think of it as a particular threat, one greater than Red Hat or any other distro. So what's going on here? I have a theory: Lindows isn't free. It's cheap—and that has made all the difference. Software that costs money is a familiar model. Software that's free doesn't fit in many people's model of how the world works. Government functionaries, in particular, seem not to understand how people can give away something that in the past has been so terribly expensive. They probably consider it some sort of trap (especially if they've read anything recent from RMS) and would prefer to stand on more solid legal ground, even if it means money going out the door. Lindows CEO Michael Robertson is handling all of this brilliantly. Every time Microsoft goes to court against him, he gets more publicity and sympathy. He probably has a new name in his briefcase that he'll slap on the product once he's run out of appeals (he's already lost the first round in Sweden and Finland) but he won't trot it out until he absolutely has to. By then, Microsoft will have "paid" for a PR campaign that Robertson could never have afforded on his own, and the whole world will know the product, even if it changes names. (The sign on that fast food place says "KFC," but we all know what it stands for, right?) Names are cheap, heh. Publicity is priceless. It may strike some as ironic that Microsoft fears a product that costs $50 more than one that costs nothing at all. No irony here at all: Roberston is taking on Microsoft within Microsoft's own business model, and is making headway. The jury is still out in many places as to whether the Open Source idea will wither or prevail, but paid software—well, we know that that works. Maybe MS should be scared. My guess is that the Lindows people have way more pricing room than they do! |
|
|
Deal! (I'm easy.) Mr. Byte's life spanned the truly formative years of personal computing: He took us from the CP/M era to Windows NT, and I wrote a good deal of my first Turbo Pascal book with him curled up around my feet in chilly Rochester, NY. He didn't quite make it to 15, but his life was a pretty good run, and I wrote about him in PC Techniques, Dr. Dobb's Journal, and all of my books. The photo at left was taken early in 1983, when he was just beginning to settle down from puppy rambunctiousness and learn a few tricks. Carol and I always celebrate Mr. Byte's birthday by making a pan of lemon bars and devouring them. Why? Well, on his first birthday, December 11, 1981, Carol had made a pan of lemon bars, and because we had hoped to have them for dessert that evening, she put them on the floor of the mud room to cool and closed the door. (The mud room was unheated, and Rochester was cold in December.) Mr. Byte smelled the lemon bars coming under the door, and so he nonchalantly nosed the mudroom door open and ate the middle of the batch right out of the pan. (He left us the edges, but we decided not to partake.) Carol was about to swat him, but then she remembered that it was his birthday, and we had totally forgotten. He was spared the swatting, and thereafter, we never forgot his birthday, and while he was alive, we always made lemon bars and shared them with him. A couple of years ago, we decided to remember Mr. Byte on his birthday with lemon bars, and it's been a sacramental occasion for us ever since. Will we ever get another dog? I think so, but it won't be until we get settled in the new house and decompress for awhile. Carol gets to name this one—I'm hooked and she doesn't need to bargain anymore—and whether it's a he or a she, the name we've chosen is Abergavenny, or Abby for short. A bichon, of course; I doubt we'll ever have any other breed. And when the project begins, you'll read about it here, of course. |
|
|
Through all the discussion over the past few years about controlling spam, the direct marketing lobby has adamantly insisted that it will only support "opt-out" for email marketing. In other words, they demand the legal right to spam you once, and then it's your job to tell them to take you off their list. Astonishingly enough, no spam law yet passed makes it illegal for spammers to sell email addresses to other spammers, and this is what makes opt-out less than useless: A spammer will sell your address to other spammers, or else set up countless shell companies or simply domains claiming to be separate companies, and will continue to spam you, just not from the same domain or company twice. (At least, for those spammers who will respect the law, which is likely to be a small group indeed.) The direct marketers know that opt-in would mean the end of direct email marketing, and they'll fight to the death to prevent opt-in from being part of any law. I think we could compromise—though I admit the compromise leans a little in the direction of consumers. How about pitching a new form of opt-out to Congress? I call it "blind opt-out," and the gist here is that it allows consumers to opt-out without ever revealing their email addresses to spammers. There are two ways to do it:
There are complications, and of course the main one is defining just what spam is. Spammers will dress their pitches up in any costume they think might skirt the law; for example, claiming a political statement protected under the First Amendment, with three paragraphs of editorializing followed by the pitch for the penis patch or whatever else they might be selling. Fake newsletters have long been a spammer tactic. Legitimate newsletters would have to encourage people to deliberately sign up and be able to prove it. This shouldn't be too onerous a burden; Visual Developer had to prove to the Post Office that we were really a magazine worthy of second class mail privileges, and there were standards to meet that literally counted ad pages accurate down to small fractions of a page. Marketers will howl that this isn't opt-out at all, but it really is, and if it puts an end to direct email marketing, few of us would cry very much. I'm going to write a note to my legislators asking them to pursue this, and if any of you have the inclination, who knows? It might help. Legislation really isn't the answer—there's not enough money in the world to pay cops to chase down illegal spammers—but until we get off our asses and authenticate all email, this might be a reasonable step in the right direction. |
|
|
As I related in my December 6, 2003 entry, PocoMail corrupted my mailbase a few nights ago, and I've begun rebuilding it from backup copies, one mailbox at a time. I'm doing it that way to make sure I knock out as many unnecessary messages as possible from my mailbase so that I don't pass that way again any time soon. (The problem appears to be due to a signed integer overflow in counting messages in my mailbase, which topped 32,768 about the time trouble began.) It's been a strange, sobering experience to go through so many messages. I had saved a lot of company email from the Coriolis era, going back to 1994. Some was obvious cannon fodder: I had about a thousand messages in the "Author Chatter" folder, where I stored traffic with PCT, VDM, and Coriolis book authors. More than half of that is now gone, and I'm still working on it. There's little point in keeping messages containing attached articles that I published in 1997, but just seeing them was like watching a movie of my career over the past ten years. I miss having daily interaction with such a brilliant crowd, and I still chew nails over the death of the technical magazine industry, which I loved. I never had to think too hard about the issue at hand here: What information is worth keeping, and what information can be abandoned and forgotten? One thing that took me by surprise was the number of messages that did not carry their own context; for example, an author replying tersely that "I'll have those contracts back to you next week. Bye." Why I kept those at all I don't know, unless it was simply that I had all the context stored in the back of my head at the time, and the obvious truth that the context would be completely forgotten in five years never occurred to me. Even where the context was obvious—as in a whole passel of messages concerning a cover glitch we had on one of the High Performance books, back in 1997—most of the messages have no value right now, assuming that an insanely detailed history of The Coriolis Group will never need to be written. In fact, I realize that I'm doing something perilously close to writing my memoirs by choosing which messages to keep and which to toss. What's left will be an interesting picture of my activities over time since 1994, and until I shoveled out the irrelevant and context-free traffic, that picture could not emerge from the noise. It's an issue with larger implications. The question of what human knowledge actually is is a subtle one, and the question of what to remember and what to forget depends on human judgment that almost always has biases and agendas. I'm sure I deleted a lot of messages that remind me of things that I'd sooner forget. Is that good? Will I regret it someday? I'm not sure, and I won't know until I get there. History is supposedly written by the victors; well, if I write my own history does that mean that I Win? |
|
|
Ouch. It's a whole different ballgame now. Ten years ago, I had a book called Upgrading and Repairing PCs by Scott Mueller. It was a sort of combined Gray's Anatomy and Handbook of Surgery for Intel PC systems, and I'm pleased to report that Scott has kept it current. I recently bought the 15th anniversary edition of the book, and have been spending my spare moments reacquainting myself with PC hardware. The good news is that there are a lot of strong standards now that we didn't have in 1993; the bad news is that they are much deeper standards, and there are a lot more of them. The book is almost 1600 pages long, set in very small type with narrow margins. I doubt I've ever seen a technical book that packed more between two covers, and it's lucidly if tersely written, though definitely not for beginners. I am currently working through the chapters on motherboard standards, and I think when the time comes to build myself a new main system early next year, I'll know the dependencies among and the order in which I have to make certain decisions. It's a $60 book, but for once I'll say it's worth all of that and more. If you ever decide to go down the path I'm going—or even if you go down that path routinely—this book is a must-have. |
|
|
A few odd lots here on the Second Sunday of Advent:
|
|
|
I've been having a lot of trouble with PocoMail in recent weeks, all of which involve messages arriving with headers but no bodies. This happened most frequently on messages containing forwards or attachments, but it also happened to ordinary text messages as well. I originally thought it was something to do with POPFile, but it continued to happen even after I uninstalled POPFile. Last night was the kicker: I compressed all folders in my Poco mailbase, and in the process the mailbase got seriously corrupted. Headers and messages were being combined willy-nilly, so that the headers from one message would be shown atop the body of an entirely different message. That copy of the mailbase is a goner, though I do have backup copies. Why it should have happened is still something of a mystery: A week ago I downloaded a new copy of PocoMail 3.03 and installed it, and then imported a copy of the mailbase from another, older but still problem-plagued install. No dice; the problem was the same as it had been on the older install. I installed a fresh copy of Poco from the same download, and didn't import the mailbase, but simply started reading mail into it, and for that install the problem has disappeared. So it was something in the mailbase somewhere. I've been beating my head against this all week, and might still be bloodying the wall except that my most brilliant wife suggested that my mailbase might simply be too damned big for Poco to handle correctly. Now, I've been accumulating messages in this mailbase since 1994, and admit that there are a lot of them. How many I wasn't sure, but I guessed around 20,000. Hah. I did a count, and found that I had just under 33,000 messages in the mailbase. That is indeed a lot, but I couldn't see how that would in and of itself be a problem. Then last night at four ayem, it hit me: I had started to have problems just about the time the number of messages in my mailbase passed 32,768. Yikes! Somewhere in Poco's code, something was counting messages with a 16-bit signed integer. I suddenly flashed on the legendary comment made by Bill Gates many years ago, concerning the now-forgotten DOS 32-megabyte "barrier" on hard drives: "Who would have thought that anybody would ever have a disk with more than 32 megabytes on it!" And who would ever be so mad as to accumulate 32,768 emails in a single mailbase? Moi. I haven't told Poco Systems about this yet, pending some additional sniffing, but I will at some point. In the meantime, I'm rebuilding my mailbase, one RFC822 mailbox at a time, dumping the unnecessary messages as each mailbox comes in. I hope to cut the size of the mailbase by half, though it will take some time and work. Several people wrote to tell me that they had had similar problems with Poco. How many messages do you guys have lying around? Will the number overflow a 16-bit integer? This may be the core issue. |
|
|
I show it here more for personal nostalgia than anything else. I miss my workshop badly, and although the workshop I'll have in the new house is coming along nicely, it's going to be a few months yet before I can get back to making metal shavings in a big way. Finally, it's strictly an optical illusion that seems to show sparks leaking out of a hole in my jeans pocket, but the photo did prompt Carol to suggest that I was wearing the safety shield in the wrong place. |
|
|
Those of you who tried to read Contra earlier this afternoon discovered a number of weirdnesses, and the backstory is that I had a serious meltdown here as I tried to upload several days' worth of entries. Only a portion of the diary.htm file reached the server, but enough got there to replace what was already there—and then my flaky Dell had the temerity to destroy the Zip cartridge on which I keep all my Web files. I had to revert to backups, which ended on 11/29. As my note files went up in flames with the cartridge, I'm not even going to try to recreate Dec 1-3. Clearly, I have some serious work to do here on my three desktop systems. One is healthy, one is sick, and one is dead. The dead one should be easy to bring back to life, though it'll take an afternoon of CD swapping and foot tapping. I had intended to use it as my new Linux lab machine, but I need a standby Win2K system now that I can rely on. It's not especially fast (550 Mhz) but it's fast enough for word processing, email, and Web surfing. I may also see about going back to Zip 100 cartridges, which always worked famously well for me, unlike the famously flaky 250s. (I am not planning on evaluating the new Zip 750s. No way.) One way or another, I'll be busy on it for a bit, so if I'm late posting things here, don't despair. I need to have a firm place to stand or I won't get a great deal done. |
|
December 31, 2003:
We're
up to #8 in my top-ten countdown of Christmas CDs. 
Christmas
Eve is now only eleven days off, and with this entry of Contra I have
decided to kick off a countdown of my top ten favorite Christmas CDs,
one each day until December 23. #10 is shown at left: The Christmas
Album, by the Taverner Consort, directed by Andrew Parrot. The collection
is of 17th and 18th century sung carols from many different countries.
None are likely to be familiar to you, and that's one reason I like it.
Hey, when was the last time you heard William Billings' "Methinks
I See an Heav'nly Host"? (And by now are you maybe just a little
sick of the radio playing, "Santa Baby"?) Unfortunately, when
I checked on Amazon I found that this disc (which was published in 1992)
is no longer available new, but if you scout the used music sites you
might score a used copy. It's unusual, and like all of the discs I will
profile in coming days, well worth having, especially if you like religious
choral music.
