Monday, 17 June 2002
09:33 - The ATI web site is still hosed. A tracert encounters timeouts near the ATI server, so it appears there's some sort of datacomm problem. I'd been trying since last week to download the CATALYST RADEON (7.72) Reference Drivers for Windows 2000. One of my readers finally got that driver, and was kind enough to put it on his server so that I could download it. I accidentally put it on my own server, but I'll probably notice it's there sometime in the next couple days and remove it. I haven't tried installing it yet, so if you do you're on your own.
The Inquirer reports that Microsoft may release WinXP SP1 sooner than expected, possibly as soon as the end of this month. As was the case with Win2000 SP1, the first SP is going to be huge, perhaps 120 MB. It sounds like there are a lot of bugs to be fixed.
I just read an excellent DistroWatch article entitled Is RPM Doomed?. Having been bitten before by Dependency Hell, which makes DLL Hell look like a minor problem, I plan to re-read this article in some detail and think about alternatives. I'll probably stick with Red Hat 7.3 for now, but I suspect I'll eventually migrate to one of the other distributions.
Here's the press release about Wal*Mart/Lindows PCs. What I don't understand is how Wal*Mart can sell PCs with what the Lindows web site describes as an alpha release of the OS. Also, given that Lindows will apparently sell for $99 and that one copy is licensed for use on all PCs in your home, does that mean that Wal*Mart will sell the same models without a Lindows license at a discount from the price of that model with Lindows? I also find the Click-and-Run Warehouse concept interesting. Users will pay $99/year for the right to download what appear to be free applications. Apparently, Lindows is charging $99/year for the ease and convenience of a single-click download and install.
I don't much care for the end-user license agreement of Lindows, and Lindows has already drawn the ire of the OSS community for their failure to post their source code, but this business model may in fact turn into something. If there's a problem it's going to be with the alpha nature of the OS being distributed, with the over-promising of compatibility with Windows applications, and with the scarcity of Linux applications that average home users will want to use. My own feeling is that Wal*Mart would have been better advised to install Red Hat or Mandrake, but it'll be interesting to see what happens. For now at least, Wal*Mart is offering these systems only on their web site, but I suspect we'll eventually see them in the stores as well.
I have lots to do before the new edition of PC Hardware in a Nutshell hits the bookstores next Monday, and not enough time to do it. Updates here will probably be sparse.
13:47 - This is very strange. I installed the updated ATI drivers on my main system. Doing that apparently fixed my FrontPage problems, none of which had to do with video. A week or two ago, FrontPage started coming up without a default web, even though I told it repeatedly to "Open last web automatically when FrontPage starts". So this morning I installed the updated ATI drivers. I did nothing else unusual to my system. When I fired up FrontPage 2000 just now, the first sign that something had changed was that it came up maximized (as it used to) rather than windowed. The second sign was that a box popped up to tell me that FrontPage was not my default web page editor and did I want to change that. I told it yes, and noticed the third sign that something had changed. Instead of FrontPage being open but with no web open, the last web I'd edited (ttgnet.com) was open. Then I clicked on File -- Recent Files and noticed the fourth difference. My recently-used file list was a couple weeks or so old, showing 2002-22. html as the most recently edited file. Okay, something apparently rolled me back to the way FrontPage was two or three weeks ago. I have no idea what happened.
I've also spent some time on Barbara's system. I installed Norton Internet Security (which includes NAV 2001) on her system a year ago tomorrow. Norton's AutoUpdate feature for virus sigs has been bugging Barbara about her update service expiring. I finally decided to do something about it this morning. Sure enough, it was set to expire tomorrow.
I had planned to just pay Symantec to extend the service for another year, even though I have multiple unused copies of NIS and NAV around here. But Jerry Pournelle's bad experiences attempting to pay Symantec to extend his AV sig service decided me against doing that. I decided just to re-install NIS/NAV, but I figured there'd be something in the registry that'd prohibit that from working. So I uninstalled NIS/NAV and went into the registry with fire and sword to eliminate all references to Norton and Symantec. Then I went to my software shelf, where there are a couple of unused copies of NIS/NAV sitting.
I almost picked a NIS/NAV box, but then I noticed an older copy of NAV 2001 Professional. I really don't need the NIS functionality on our desktop machines, because everything is behind a proxy server anyway. So I decided to install NAV 2001 Pro rather than NIS/NAV on Barbara's machine. That installed fine, and had the added benefit of adding the Norton enhancements to the Recycle Bin. The only thing NIS provided that I care about is ad blocking/script filtering, and WebWasher does that very well. When I get a moment, I'll install WebWasher on Barbara's system and be done with it.
Tuesday, 18 June 2002
09:00 - As I sat there last night in the den, throwing the ball down the hall over and over and over, with Malcolm bringing it back each time and placing it carefully in front of me, one of those weird thoughts came to mind: "Note to self: Never take Malcolm to hand-grenade practice"
Which reminds me of a discussion that took place years ago at the table where I used to hold court in the college snack bar (called the Gedunk, which is indicative in itself). This would have been in 1972 or thereabouts, and at the table were Al Brower, a professor friend of mine who'd landed as an infantry sergeant at Normandy and carried his Thompson gun all the way to Berlin, and two other gentlemen whose names I forget. One of them had been a junior officer in the 1st SS Liebstandarte (LSSAH), and the other a junior officer in an elite British division, exactly which one I don't recall. The former had fought the Americans, and the latter had fought beside them.
Somehow the subject got onto hand grenades, and all of them agreed that the Americans used grenades more profusely than any other combatant. Al said they never entered a building without grenading it thoroughly first, whereas both the others said that they seldom had enough grenades available to use them that liberally. So they talked about supply and logistics for a while.
But then they started talking about grenades again. The German guy commented that he and his buddies were always amazed at how far and how accurately the average American soldier could throw grenades. The British guy concurred, and said that a typical American soldier could throw a grenade farther and more accurately than could even specially trained Brit soldiers. They speculated for a while about why that should be so.
I thought their conclusion was both accurate and interesting. Nearly every American soldier in WWII had grown up playing baseball all summer long, so they were in effect pre-trained grenadiers. I was going to comment that Brits and Germans throw like girls, but neither of those guys looked like they had much of a sense of humor.
Nowadays, not many American kids play baseball at all, let alone as compulsively as American boys did back in the 30's and 40's. I wonder if American soldiers are still any good at throwing grenades.
Here's something you don't see very often. A major Apache vulnerability. The Register reports full details, and speculates that this is an attempt to damage the reputation of Open Source Software. Apache, of course, yelped that the bug was made public before they'd had a chance to fix it. Superficially, this looks similar to Microsoft's typical attempt to blame the person who discovers the bug. Someone discovers a bug and announces it to the world before the software vendor has a chance to prepare a fix.
But this is actually significantly different. With a Microsoft bug, typically someone discovers a bug, tells Microsoft about it, waits a month with nothing happening, and then announces the bug to the world to force Microsoft to do something about it. In this case, apparently someone discovered a bug in Apache and announced the bug to the world without bothering even to tell Apache about it. If I were a conspiracy theorist, I'd wonder if Microsoft had paid this company to search out a bug in Apache and announce it to the world before giving Apache any chance at all to fix the bug.
The real test will be how long it takes Apache to fix the bug and distribute the fix. My guess is that it won't be long at all.
People are always talking about "killer applications", but my vote for Real Killer App of the Year goes to Dead Man's Switch. If you install it on your system, it periodically prompts you to enter a password to make sure you're really there and you're really you. If you fail to enter the password on time, the software can take various user-defined actions on your behalf, including "posting to web pages, sending e-mails, and encrypting files".
I suspect a lot of drug dealers and others who don't want their PCs to be accessible to law enforcement personnel will use this program. With a little ingenuity, it could probably be enhanced to control a relay inside a PC that would close when the program activated and set off a small thermite device sitting on top of the hard drive. Even the NSA can't do much with a hard drive that's been converted into a pool of molten slag.
12:14 - There's a fascinating article on software quality on MSNBC, entitled Why software is so bad ... ... and what’s being done to fix it. Here is a quote from that article:
This article is definitely worth reading, and is just another argument in favor of OSS.
And speaking of patches, one of my readers, who wishes to remain anonymous for obvious reasons, sends the following:
Wednesday, 19 June 2002
10:22 - The Apache bug was fixed and an updated release posted sometime yesterday. I'm not sure when they fixed it, but the new version was available when I checked the Apache web site around dinner time. I posted the following on LinuxMuse.com at 5:50 p.m. EDT.
The movie Memphis Belle was on MoviePlex last night. Not the original 1944 documentary about the first B-17 crew to complete 25 missions, but a 1990 fictional remake. The movie guides gave it a mediocre review, but I decided to watch it to see if it was really as bad as they said.
It was a typical war movie, but I was disappointed in the numerous technical errors and anachronisms. The B-17's were of a later model than flew in Spring, 1943, when the movie was set. The stuff that happened on their inbound leg was simply stupid. B-17's tested their guns over the English Channel, not over eastern France, when it was too late to do anything about a gun problem. The escort fighters in the movie were P-51 Mustangs, which didn't even arrive in theatre until a year later. Had they really had P-51s escorting them, the fighters wouldn't have had to turn back due to low fuel. P-51s escorted B-17s all the way to Berlin and back. In fact, when Hermann Goering first saw P-51s over Berlin he turned to his companion and told him the war was lost. The Memphis Belle would have had P-47 escorts, which indeed could escort them only part way.
And the German fighters weren't stupid, although you'd think so from watching this movie. They stage their first attack while the American escort fighters are still present. As my dad told me, it didn't happen that way. The German fighters' job wasn't to mix it up with American fighters. It was to shoot down B-17s. My dad said that as soon as their escort fighters disappeared, they'd clear their guns because they knew the German fighters would appear literally within a couple minutes of when the P-47s headed for home.
And the fight scenes were pretty ridiculous. They had the German fighters moving in slow motion. They'd approach the B-17 at a leisurely pace and then pull up (!) showing their bellies to the B-17 gunners. Geez. In fact, any German fighter pilot with any instinct for self-preservation was scared to death of a box of B-17s. (Well, I don't think they'd started flying the Combat Box then, but the idea was the same). A real German pilot would line up on the B-17's while still well out of range of their 50 caliber guns and then push his throttle to the firewall. He'd blast through the B-17 box, balls to the wall, shooting when he got within range and then blasting through the B-17s at top speed until he was again out of range of their guns.
When approaching from the side, the closing rate would be something on the close order of 400 MPH. Neither the fighters' nor the bombers' machineguns were effective at much more than 200 to 300 yards, which meant that neither plane had much more than a second or two to shoot. Only newbie gunners on either side wasted ammunition shooting at long distances. And the head-on passes they showed were wrong for the time. As I recall, German fighters didn't start using head-on passes until late 1943 or early 1944. The problem with head-on passes was that the closing rate was nearly 700 MPH, which meant that the fighters' machineguns were within range for only a fraction of a second. That wasn't long enough to do significant damage to a B-17 using machine guns, so by the time the Germans started using head-on passes they were emphasizing cannon (and many of the B-17s by that time had quad-50's in the nose to discourage head-on attacks. The movie did include some original footage, which made it very clear just how badly they'd simulated German fighter attacks on B-17s.
The bomb run itself was pretty stupid too. They had the American pilot refusing to drop his load because they might hit a school near their target. Nothing like projecting 1990 sensibilities onto a 1943 battle. In fact, no bomber crew in 1943 would have even thought about that school. The bomber crews in 1943 knew that the "precision bombing" concept was a sick joke. The Americans, who bombed during daylight, were lucky to get their bombs within half a mile of the target. That's why they started using 1,000 plane raids--to make sure they got what they were aiming at by saturating the area. The Brits, who carpet-bombed at night, aimed at whole cities, and sometimes missed even that large a target. And what was really stupid was that the primary was clouded over, so the lead pilot (in the Memphis Belle) decided to take the bombers around and make another pass. Yeah, right. This with a 20 mile long string of other bombers following him in to target. That would have worked. In his dreams.
Then, on the return run, we have a crippled plane. In that situation, the pilot would have ordered his uninjured crew to bail out near the airfield. If he thought he had a chance to bring the plane back safely and particularly if he had someone too badly wounded to bail out, he might have tried. Otherwise, he'd have bailed out along with his crew and ditched the plane in the English Channel. But all of our heroes in the Memphis Belle ride a crippled plane in to land, getting the jammed wheel (and was it the left or the right that was jammed--they seemed confused) down just as the wheel touched the tarmac. Give me a break.
Then there's the small matter that they had the Memphis Belle attacking the wrong target, being damaged, and taking casualties. In fact, the 25th mission of the Memphis Belle was uneventful, with no damage to the plane or injuries to the crew. I know the people who made this movie were attempting a tribute, but they ended up unintentionally insulting the brave men on both sides who fought and died during the air war over Germany.
13:57 - Someone emailed me to ask if the Germans really used cannon in their fighters. The answer is that they did, but not in the sense that many people think of a cannon. Technically, anything below 1" (25.4mm) was considered a small arm and anything one inch or larger was considered artillery (or cannon). In practice, anything up to about 50 caliber (12.7mm) was considered a machinegun, and anything above that an automatic cannon. The Germans mounted 20mm, 25mm, and 30mm guns on various fighters at various times.
The upside to the cannon was that even one hit could do a lot of damage. The downside was that most cannon had relatively slow cyclic rates, and so during the second or two that the fighter was in range, the cannon didn't fire many projectiles. But even the 50 caliber Brownings used by the US threw pretty big projectiles. Just as a basis of comparison, I went down to the basement and pulled the tarp off my quad-50 Anti-Santa gun and grabbed one of the rounds from the belt. That round is shown at the top. It's an M20 50-caliber round with a Silver/Red Tip 619 Grain Armor Piercing Incendiary Tracer (APIT). Immediately below is a 5.56mm ball round (the current US/NATO military rifle caliber) and a roll of 35mm film for size comparison. On most monitors, the full-size image will be about life size.
We're getting ready to head off for dinner around 6:00 p.m. and then to the Forsyth Astronomical Society meeting.
15:00 - Well, this is an outrage. CNN reports that the pilots involved in the "friendly fire" (talk about a really stupid oxymoron) incident in which four Canadian soldiers died may face criminal charges. We may never know who really screwed up. From the published reports, it's not clear to me who, if anyone, was responsible. It might have been the pilot, certainly, but it might also have been any number of other people, from those responsible for coordinating the live-fire practice, to the controller aboard the AWACS, to the folks who briefed the mission. Or, as often happens in war, it might be a combination of small errors that snowballed into a huge error. As the pungent saying has it, "Shit happens."
Criminal charges ordinarily require either malicious intention or gross negligence, and I don't see evidence of either here. This pilot reasonably believed he was being shot at, and under the rules of engagement in effect he was entitled to shoot back to protect himself and his aircraft. He did shoot back, with tragic results. Everyone involved, the pilot most of all, regrets what happened. But using 20-20 hindsight to assign blame to the pilot and then filing criminal charges against him smacks of political expediency.
Sure, many Canadians are upset. If the situation were reversed, many US citizens would be demanding something be done to punish a Canadian F-16 pilot who accidentally killed US soldiers. But if you asked the guys on the ground and in the air, both Canadians and Americans, I'm sure they'd tell you. They don't like it. They wish it wouldn't. But Shit Happens.
And I just got off the phone with a reporter for the Washington Post. I suspect I impressed the hell out of him, because he'd called to talk to me about the FTC action against Rambus, and I'd just finished reading the articles on The Register and The Inquirer when he called. I didn't mention that, of course. I let him think that I magically know everything that's happening in the PC industry. Heh.
The timing was good, though. He asked me how I wanted my name to appear, and it sounds like he's willing to put in a plug for the new edition of PC Hardware in a Nutshell. I now get calls regularly from reporters for the New York Times, the Wall Street Journal, the Washington Post, the Times (of London), and other major papers. A lot of them are deep background, but I get quoted and attributed sometimes. Perhaps one day I'll even graduate to the pundit class.
Thursday, 20 June 2002
8:58 - Oh, well. The Washington Post quoted me. They didn't mention the new edition of PC Hardware in a Nutshell, but at least they spelled my name right.
Speaking of media, those clueless morons at NPR have come up with something unbelievable. They now have a "Linking Policy" that states:
Linking to or framing of any material on this site without the prior written consent of NPR is prohibited. If you would like to link to NPR from your Web site, please fill out the link permission request form.
Someone needs to tell them that nobody needs permission to link to another web site. Oh, well. They've been slashdotted, so presumably they already know.
And destoplinux.com has called for the First Annual wIndependence Day on 4 July. Cute.
Lots to do. More later.
14:03 - Wow! That was close. The BBC reports that we almost got nailed by a passing asteroid the size of an office building. It missed Terra by only about 120,000 kilometers, which in astronomical terms is a direct hit. What's worse is that no one noticed it until after it had passed Terra's orbit. If it had struck, it would have leveled everything in an area of about 2,000 square kilometers. One day, perhaps sometime soon, we're going to get hit, and we're doing nothing to protect against that eventuality. Thanks to reader Michael Hill for the link.
It's official. I've now relegated Microsoft Word 2000 and Excel 2000 to secondary status on my systems. My word processor of choice is now OpenOffice.org 1.0 Writer and my spreadsheet of choice is OpenOffice.org 1.0 Calc. Based on the work I've done with the OOo product since its release, I've decided that it's more than Good Enough for me, and easily replaces the corresponding Microsoft Office products.
What's very bad news for Microsoft is that OOo is getting mainstream mindshare. See, for example, this New York Times article by David Pogue. The timing of the OOo 1.0 release couldn't be worse for Microsoft. Just when they're trying to cram Licensing 6.0 down unwilling corporate throats, along comes a product that's free, full-featured, fast, and compatible. For most people, the only real reason to continue using Microsoft is compatibility concerns. And OOo is quite compatible, with minor exceptions.
Microsoft has always used their Office file formats as a competitive weapon, so perhaps it's time to turn the tables on them. I think that from now on when I send a document or spreadsheet to someone, I'll send it in native OOo format and suggest to the recipient that he download and install OOo, and configure it to use OOo native standards-compliant file formats. If someone is sending me a file, I'll tell him that although I can read Microsoft formats I'd really prefer that he send the file in native OOo format. Wouldn't it be ironic if Microsoft were forced to include an import filter for OOo formatted files in the next release of Office? They'd never include an export filter, of course.
Like most OOo users, I'll keep Word 2000 and Excel 2000 on my systems just in case I need them to read a file in their native formats, but I don't expect that to happen very often. Oddly enough given their focus on OSS, my one concern right now is exchanging documents with O'Reilly & Associates. I've already asked them to produce an author template for OOo to match the one they have for Word, and I expect they'll complete that soon. Eventually I will migrate entirely away from proprietary Microsoft file formats and begin using the native standards-compliant OOo formats. In fact, I'm already doing so for in-house stuff.
It's unfortunate that OOo doesn't have clones of Outlook 2000 and FrontPage 2000. If it did, I wouldn't need Office at all. I find I need Internet Explorer only once in a great while now, and my need for IE will continue to shrink as web sites begin eliminating proprietary stuff that demands IE. Mozilla is my default browser now, and I find I use it for about 99% of my browsing. I'm taking what steps I can to migrate away from proprietary software and proprietary formats toward OSS software and standards-compliant formats, albeit still on the Windows platform. But that'll change, too. I've had enough of Microsoft locking me into proprietary, undocumented file formats. As The Who said, We don't get fooled again.
Friday, 21 June 2002
9:43 - We're getting Pournelle's sites moved over to Rocket, the server that Greg Lincoln and Brian Bilbrey run LinuxMuse.com on, and the same server that I transferred my domains to last month. As expected, the process has been problem-free on our end. Greg and Brian get things set up quickly and properly. The problem is dealing with InterNIC/NSI/VeriSign, which is run like an old-style Soviet bureau. Oh, well. Eventually we'll get everything moved to Rocket, after which we'll get all of Pournelle's existing domains and several new ones registered with someone other than VeriSign. Once that's done, life will be a lot easier.
I need to get some work done. More later, maybe.
Saturday, 22 June 2002
9:56 - Well, it's later. The new edition of PC Hardware in a Nutshell is supposed to hit the bookstores on Monday. I still have lots to do on the web site, so I won't have much time to write entries here for the next several days.
We're still in the process of getting Pournelle's sites moved over from pair Networks to rocket, the server that hosts LinuxMuse.com, and to which I moved my domains last month. The transition has been largely painless, and chances are if you're not already hitting the new server you will be by tomorrow.
As the penguin once said, it's back to work for me...
Sunday, 23 June 2002
Copyright © 1998, 1999, 2000, 2001, 2002, 2003, 2004 by Robert Bruce Thompson. All Rights Reserved.