Home Daynotes Home Week of 28 October 2002

Photograph of Robert Bruce Thompson Daynotes Journal

Week of 28 October 2002

Latest Update : Sunday, 03 November 2002 08:48 -0500

Click Here to Subscribe Buy PC Hardware in a Nutshell, 2nd Edition: [Amazon] [Barnes & Noble] [Bookpool] Visit Barbara's Journal Page

Monday, 28 October 2002

[Last Week] [ Monday ] [Tuesday] [Wednesday ] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Messageboard] [HardwareGuys.com Messageboard] [TechnoMayhem.com Messageboard]

10:03 - Mozilla Mail gave me a nasty shock yesterday morning. I run it on both my main office system and my secondary den system. Each system stores the mail locally, and they have no connection. The main office system is set to delete mail from the server after retrieving it. The den system is set to leave the mail on the server. That way, my main office system gets and stores all mail messages, while the den system just downloads any that are new since the last time I downloaded mail on the main office system. I can read and reply to new mail in the den, while keeping my mail complete on the office system. The den system is set to BCC me automatically, which means that copies of any messages I send from the den system show up on the main system. Kludgy, but it works.

Yesterday morning I sat down as usual at the den system and told it to go get my new mail. An hourglass appeared and no new mail was retrieved. Hmmm. I figured the mail server was having problems or there was a link problem. So I fired up Mozilla and hit my web site, which displayed normally. Not a connection problem, then. I fired up Mozilla Mail again, and this time it retrieved my new mail normally. I didn't think any more about it.

Until, that is, I went into my office and fired up Mozilla Mail on my main system. Once again, I got an hourglass and no new mail came in. So I fired up Mozilla and hit my web site, which displayed normally. I fired up Mozilla Mail again, and again got the hourglass. Thinking that perhaps the mail server was having problems, I telnet'd over to port 110 on my mail server. I was able to login normally and saw that I had 138 new messages on the server. That's about half the normal number for overnight mail, but it was Sunday morning, so perhaps that was a normal number.

I tried Mozilla Mail again, and again got the hourglass. So I shut down all of the Mozilla windows open on my task bar and then fired up Mozilla Mail again. This time, it started retrieving my mail normally, and I figured the problem was solved. That turned out not to be the case, however.

Ordinarily, Mozilla Mail filters my new mail into various folders, and each folder with new mail in it shows in bold with the number of new messages listed. That didn't happen this time. Instead, as I clicked each folder that I'd expect to have new mail, the folder name changed from normal to bold, and the number of new messages in it appeared. That concerned me, but at least the display was normal after I clicked the folder. 

The real problem showed up when I clicked folders that didn't have any new messages in them. For example, I clicked my Trash folder. One of my filters moves spam to the Trash folder, where they then sit awaiting deletion. But that filter uses the "delete" function rather than the "move" function, and that means that all of the messages filtered to trash are (or should be) already marked as read. When I clicked the Trash folder, the name turned bold and showed that I had something like 40 "unread" messages. That was very odd. Then I clicked Sent Items, which shouldn't have had any unread messages at all in it. The folder name turned bold, and it showed that I had 954 unread messages. I went along clicking on other folders that I knew had no unread messages in them, and in about 2/3 of those folders all messages were marked as unread. The remaining third behaved normally, which is to say that all messages that should have been in them were in them, and all of them were marked read.

So what is going on here? The only thing I can assume is that something on the server caused the problem. These are two completely separate installations of Mozilla Mail, which happen to be hitting the same POP box. 

Barbara and I drove up to the Parkway yesterday afternoon. Steve Childers, one of the guys in our astronomy club, is a professor at Wake Forest University. WFU has a lodge up on the Parkway that's available for use by faculty. Steve had reserved the lodge for last night, with the idea that we'd scope it out as a possible base of operations for the club's Messier Marathon next year. (The Messier Marathon starts at dusk and ends at dawn, with each of us trying to locate all 110 of the Messier Objects during the night. It can be done only during a new moon in March or April each year.) Even though the forecast was for 100% cloud cover, several of us drove up to check the place out.

As it turns out, it'll be an excellent base for operations. It's a reasonably dark site with reasonably good horizons. We may be able to do all our Messier Marathon observing from there. Barbara and I left the dogs at home, and planned to return late last night. Once we got up there, though, it was pretty obvious that Barbara wanted to spend the night, as several others were doing. So I told her I'd drive back last night and take care of the dogs. She was concerned about me driving back by myself at night, why I don't know. It was easy enough. I just put it in neutral at the top of the mountain, turned off my headlights, and free-wheeled down the mountain with my eyes closed, shouting "Yee-haaaaaaaaaaaa" and chugging a bottle of bourbon.

Oops. Steve just dropped Barbara off, so I'd better not let her see this. (Barbara now has some pictures of the place up on her page.)

I've gotten mail from several readers asking about the Leonids meteor storm, so I'll try to answer them here:

First, everyone should be aware that the Leonids meteor shower (as opposed to "storm") spans a week or so, centered on the night of 18/19 November (EST). The two peaks are forecast to occur at about 11:03 p.m. Eastern Standard Time on the evening of 18 November and about 5:46 a.m. EST on the morning of 19 November. Each peak spans a couple hours or so, centered on the stated time, with the frequency of meteors expected mapping as a bell curve with the maxima at the stated times. In short, if you look up at the sky from a dark site at any time during the week or so spanning 18/19 November, you're going to see some Leonid meteors. If you're looking at around 0403 UT or 1046 UT on 19 November, you're going to see the largest number of meteors--probably a bright one every few seconds, with many dimmer ones--assuming you're in the correct geographical location.

Second, the radiant is in the scimitar of the constellation Leo (whence the name). If that is below the horizon for you at one or both of the forecast maxima, you'll see meteors, but not nearly as many as those for whom the radiant is above the horizon. For the first forecast peak, the scimitar of Leo is near the horizon for us in Winston-Salem (actually, it's just below the horizon at the 11:03 local, although it will be just above the horizon later in the peak.) That means that the first peak will be best viewed well East of our location, such as in Western Africa and Western and Central Europe. We in Winston-Salem are on the very Western boundary of where that peak will be visible. Any farther west, and the scimitar of Leo is too far below the horizon. The best viewing of this first peak will be in the extreme Northeastern part of the US and Eastern Canada.

The second peak is a different matter. It should be visible from all parts of the US and Canada, although for this peak we in Winston-Salem are on the extreme Eastern edge of darkness. We'll just be getting into twilight then. Astronomical twilight for us in Winston-Salem ends about 0536 local time on 19 November, about 10 minutes before the forecast peak. Nautical twilight ends at 0607 local, about half an hour into the second peak. After that, the increasing sky brightness will make it harder and harder to see meteors, although I expect some will be bright enough to see well into twilight, and a few may even be visible after sunrise.

For those of you in the US and Canada much West of the Eastern seaboard, the first peak arrives too early (although you'll still see many meteors throughout the night). The best bet for you is the second peak, which centers on 0546 EST (0446 CST, 0346 MST, and 0246 PST).

Last year, the peak was on a weekend. This year, it's on Monday night and Tuesday morning. A lot of people will skip it because they have to get up for school or work on Tuesday. I think that's a mistake. This is a celestial show the like of which none of us will live to see again. It's definitely worth losing some sleep to see it.

If you do plan to view it, check the Sky & Telescope article for details about how, when, and where to do so.

13:28 - Congratulations to Jerry and Roberta Pournelle. They have a new puppy, which must be a strong candidate for cutest puppy on the planet. He has several pictures of her posted here (after next Monday, here), including this one, which shows her just after she's conquered what appears to be a stuffed Border Collie. Sable looks ferocious, and the expression on the face of the stuffed BC is classic.


Tuesday, 29 October 2002

[Last Week] [ Monday ] [Tuesday] [Wednesday ] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Messageboard] [HardwareGuys.com Messageboard] [TechnoMayhem.com Messageboard]

8:18 - Bureaucracy gone mad. The morning paper reports that a North Carolina National Guard unit arrived home from the Middle East only to be presented with a $13,000 bill for wear and tear to their equipment. It seems that the soldiers had used indelible markers to write their names on their personal equipment, as they'd been ordered to do. Also, some of their uniforms showed wear. Imagine that. So the US Army billed them for the replacement cost of the equipment they'd been wearing and carrying in a war zone. The assessment has since been reduced, but still averages $41 per soldier. One guy, for example, was charged about $20 for having written his name on his canteen.

Nothing like sending our kids off to a war zone and then slapping their faces when they finally get home. Thank goodness none of them were killed. I can just imagine the US Army sending bereaved parents a bill for a uniform full of bullet holes.

We're off to the mechanic to drop off my truck for inspection. I noticed when we were up on the Parkway Sunday that my truck's inspection sticker runs out the end of this month. North Carolina has no sense of humor about expired inspection stickers. I think there's a $250 fine.

I'll be doing heads-down writing the rest of this week, so posts here are likely to be sporadic and sparse.


Wednesday, 30 October 2002

[Last Week] [ Monday ] [Tuesday] [Wednesday ] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Messageboard] [HardwareGuys.com Messageboard] [TechnoMayhem.com Messageboard]

9:12 - Now that it has been announced that the current season of Buffy the Vampire Slayer is the last, Sarah Michelle Gellar is branching out. Thanks to Brian Bilbrey for the link. I wonder what's next.  Willow on string theory?

Patricia Cornwell's latest book, Portrait of a Killer / Jack the Ripper / Case Closed, is due out on 11 November. I plan to rush right out and not buy it, and I hope millions of people will follow my lead. If Patricia Cornwell were honest, she'd have titled her new book, Jack the Ripper / More Unfounded Speculations, or Jack the Ripper / We Still Have No Idea Who He Was.

After spending $4 million to "research" the case in her usual fashion, the most definitive evidence Ms. Cornwell can apparently come up with is:

"At best, we have a 'cautious indicator' that Sickert and Ripper mitochondrial DNA sequences may have come from the same person."

At best? May have come? That means Ms. Cornwell has no evidence at all. On that basis, my great-grandfather might have been Jack the Ripper. Ms. Cornwell's great-grandfather might have been Jack the Ripper. Hell, anyone alive at the time might have been Jack the Ripper. Ms. Cornwell hasn't even narrowed the field. She adds exactly nothing to the evidence, but hopes enough people will be foolish enough to buy her book to make it a paying proposition. The sad part is that her gamble will probably pay off. The two names, hers and the Ripper's, should be enough to guarantee huge sales for the book.

Millions of people will be convinced that Walter Sickert was Jack the Ripper. Poor Mr. Sickert. To anyone who looks at all of the (little) evidence actually available, it's clear that Mr. Sickert is probably the least likely suspect among those usually suggested. Ms. Cornwell adds insult by blaming Mr. Sickert for seven murders, rather than the generally accepted five.

I've always thought the Ripper actually killed only four. Long Liz Stride, one of the two killed on 30 September, 1888, was probably the victim of a copycat murder. There are enough discrepancies--no mutilation, the nearly impossible timing, and the fact that her murderer was almost certainly right-handed, among others--that I've always thought she was probably murdered by Michael Kidney, her common-law husband.

At any rate, there are really only four credible known suspects for the Ripper Murders--Druitt, Kosminski, Ostrog, and Tumblety--and even they are not particularly good suspects. No direct evidence exists against any of them, and good arguments exist against each of them being the Ripper. Sickert just isn't a credible suspect, and nothing Cornwell can write at this late date can make him one.


Thursday, 31 October 2002

[Last Week] [ Monday ] [Tuesday] [Wednesday ] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Messageboard] [HardwareGuys.com Messageboard] [TechnoMayhem.com Messageboard]

12:17 - I sent the following message to subscribers this morning, with the subject "Critical Vulnerability in Windows 2000 and Windows XP PPTP":

If you're running PPTP (server or client) on Windows 2000 or Windows XP, you need to read:


An unchecked buffer in Remote Access Services (RAS) could be exploited by an attacker to launch a Denial of Service (DoS) attack on unpatched systems.

Microsoft ranks this flaw as a critical risk for Windows 2000 Internet servers, a low risk for Windows 2000 Intranet servers and Windows 2000/XP clients, and not a risk for Windows XP Internet and Intranet servers.

Regardless of the risk level Microsoft assigns, I consider it prudent to apply the patch if you are running PPTP in any of these environments.

If you're a current subscriber and did not receive this message, please let me know via the subscriber-only email address. Also, I've now trimmed my subscriber list to remove those subscribers whose accounts expired 9/30/02 or earlier. If your account is expired and you'd like to renew, please visit this page.

The Inquirer reports that AMD is really on the ropes. Market share for AMD processors plunged to just 11.6%, a four-year low, and down from about 20% the previous year. PCWorld reports that AMD is losing mind-share, even among the performance/enthusiast crowd which is their base. Delays in shipping their new processors are hurting AMD badly, and it appears to me that AMD is in serious danger of reverting to their former position as just a low-cost, low-performance alternative to Intel. I hope that doesn't happen. The presence of AMD as a serious competitor has forced Intel constantly to improve their products and cut prices, and the departure of AMD as a serious alternative would hurt all of us. Still, in good conscience, I must recommend Intel products to my readers, not because of AMD's financial woes, but because Intel is the superior solution.

13:23 - The Inquirer has posted an update to their AMD story. It now seems that AMD may be in worse shape than it first appeared. At their current burn rate, AMD has enough cash to last them only through next May, which makes a successful roll-out of Barton and Hammer critical to AMD's future viability. Given the current catastrophically bad PC market and their plummeting market- and mind-share, I'm afraid AMD isn't going to make it. Even a successful roll-out of Barton and Hammer may not be enough, given AMD's fragility.



Friday, 1 November 2002

[Last Week] [ Monday ] [Tuesday] [Wednesday ] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Messageboard] [HardwareGuys.com Messageboard] [TechnoMayhem.com Messageboard]

9:34 - Barbara is off to attend a Border Collie trial about an hour from here. Rather than packing up all her stuff and driving back tonight, she's going to camp out in her truck tonight and drive back tomorrow afternoon. Two of her friends are also staying overnight. Nancy is tent-camping, and Fletch has his RV. With Barbara gone, that means wild women and parties for me tonight. Well, there would be, except (a) I don't know any wild women, and (b) I don't like parties. So I guess I'll just hang out with the dogs and read a few mysteries.

Here's something that more and more people who work at home are encountering:

-------- Original Message --------
Subject: Error 650 when using VPN and Roadrunner
Date: Fri, 1 Nov 2002 06:22:33 -0600
From: Traci Terrall (tterrall at houston dawt rr dawt com)
To: <webmaster@ttgnet.com>

Hello Robert. I am trying to resolve an error that I am getting when I use VPN to connect to my employer's network while connected to the internet via Roadrunner. I don't know if you can help me, but.... judging from what I've read, you may be able to explain my problem.

Whenever I click on the VPN connection that I have created, I get the following error:

Error 650: The computer you're dialing in to cannot establish a Dial-up Networking connection. Your server type setting might be incorrect. To fix this problem, contact your InternetService Provider, or network administrator.

I get this error if I am running on my desktop (Window Millenium) or on my laptop (Windows 98). I have a linksys router (or access point) that is wired into my Roadrunner cable modem. If I create a normal dial-up connection using Roadrunners new dial-up option, the VPN connection works. It just doesn't seem to work with the cable modem. Other people at work are able to use high-speed connections (DSL that I know of).

I have called Roadrunner and they will not support my call. And my employer only supports the dial-up connection.

Can you tell me what I have to do to fix this? I sure would apreciate it.

Thank you!

My first thought is that Roadrunner is blocking VPNs. A lot of cable-modem and DSL providers do that nowadays, claiming that VPNs are a business-class service. If that's the case, you may be able to solve the problem by upgrading your Roadrunner service from residential to business-class, although that normally costs a lot more. Even that may not solve the problem, because some cable-modem providers don't allow VPNs under any class of service.

I'll post your message (with your email address disguised) on my journal page in case anyone else has any ideas. If so, they can email you directly.




Saturday, 2 November 2002

[Last Week] [ Monday ] [Tuesday] [Wednesday ] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Messageboard] [HardwareGuys.com Messageboard] [TechnoMayhem.com Messageboard]

9:08 - Judge CKK gave Microsoft a huge win yesterday, or at least so the various news outlets are reporting. Certainly the judge gave Microsoft essentially everything they wanted. But whether that is indeed a win remains to be seen. I am reminded of the old saw about being careful what you wish for. It may turn out in retrospect from a few years' distance that this decision will be worse for Microsoft than anyone can imagine. Not because it places any limits on Microsoft's behavior, but for exactly the opposite reason. By allowing Microsoft to continue behaving as they had been, this decision may ultimately mean Microsoft's downfall.

Microsoft was already arrogant, and this decision can only make it more so. As even we pagans know, pride goeth before an injury, and arrogance before a fall. And it looks to me as though Microsoft is headed for a big fall.

Licensing 6.0 is a big part of that. Certainly, Microsoft's short-term profits showed a major increase because of the artificial jump in revenues caused by the Licensing 6.0 deadline a few months ago. But a lot of companies regarded that as little short of extortion. They paid up this time, but the real tale will be told three years from now when those licenses come due for renewal. During those three years, many of those companies will be testing Open Source alternatives, and those OSS alternatives will continue to mature. Many of the large companies that chose to go with Licensing 6.0--and nearly all of those that didn't--have Linux skunkworks projects working now to devise implementation plans for an OSS roll-out. I expect to see large-scale desktop Linux/OSS deployments start to pop up around the summer of 2003. By the summer of 2004, the trickle will become a flood.

By the summer of 2005, Microsoft is going to be faced with a very different playing field. Windows will by that time have lost the server-space battle to Linux, and desktop Windows will be competing against Red Hat Linux 11.0, not to mention OpenOffice.org 3.0 and a plethora of other world-class Linux applications. Microsoft won't be standing still during that period, of course, but there's only so much they can do to differentiate Windows and Windows applications against the challengers. Linux and Linux applications are already Good Enough for some users. Two years from now, they'll be good enough for most people, and hree years from now, they'll be good enough for anyone. And what will Microsoft do to compete against that?

Unless they can succeed in their current campaign to make OSS literally illegal or to marginalize OSS with such initiatives as Palladium, I can't see much hope for Microsoft. One thing is sure. They're not making any friends now that'll stand by them when they need friends. Oh, they'll still have friends in the various legislatures, friends they've bought and paid for, but even those "friends" may not be able to help them much. And they'll have friends like Disney, for what little they may be worth. In a few years, Microsoft is likely to be very alone, and they have only themselves to blame.

In effect, Judge CKK's decision gave Microsoft the excuse to rest on their laurels and to continue making enemies. And that's the last thing they should be doing.


Sunday, 3 November 2002

[Last Week] [ Monday ] [Tuesday] [Wednesday ] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Messageboard] [HardwareGuys.com Messageboard] [TechnoMayhem.com Messageboard]

8:48 - We're definitely getting into Autumn weather now. Cool days and freezing temperatures overnight. Today is the first time this season that I'll climb up to the roof to blow the leaves out of the gutters. Most of those fallen leaves are still green. We haven't had all that much color change around here, I guess because of the frequent rains over the last month or two. We're still technically in extreme drought, but that refers more to the water table than to the surface wetness.

I'll also put bags over the vent fans on the roof, assuming we have any bags available. If not, Barbara can pick up a set the next time she goes to Lowes or Home Depot. I'll be up on the roof frequently over the next couple months, blowing out leaves, so if not this time I'll do it next time.

I feel a bout of temporary insanity coming on. Periodically, that happens. I'll decide for no particular reason to change something that is working fine. I often have cause to regret that, but I do it anyway.

This time, I'm looking at theodore, our main NT4 Server SP6a machine. All theodore does is serve files and act as the Primary Domain Controller for our network. I'm not sure it's worth feeding him any more. I may take theodore down and just use a workgroup until I get around to bringing up a Linux server to replace it. A Linux/Samba box could certainly do everything theodore does, and would be an interesting learning experience as well. Of course, that would mean setting local permissions on all the shared stuff, so perhaps I should wait until I have a Linux server ready as a drop-in replacement for theodore. Either way, it'd be an interesting learning experience.


Copyright 1998, 1999, 2000, 2001, 2002, 2003, 2004 by Robert Bruce Thompson. All Rights Reserved.