Home » Daynotes Home » Week of 10 November 2003

Photograph of Robert Bruce Thompson Daynotes Journal

Week of 10 November 2003

Latest Update : Friday, 14 November 2003 13:53 -0500

Click Here to Subscribe Buy PC Hardware in a Nutshell, 3rd Edition: [Amazon] [Barnes & Noble] [Bookpool] Visit Barbara's Journal Page

Monday, 10 November 2003

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Forums] [HardwareGuys.com Forums] [TechnoMayhem.com Forums]
{Five Years Ago Today]

8:42 - Heads-down writing this week...

One thing that has always puzzled me about Mozilla is its default data location, buried deep within the Documents and Settings folder in a hidden subfolder of the profile folder. Does Mozilla want people to lose their data? I suppose the theory is that people back up their Documents and Settings folder, but of course that's often not the case. Microsoft doesn't make it any easier, because among the files stored in that directory tree are some that are always open when Windows is running. That means a simple xcopy backup blows up when it reaches the first open file, leaving Mozilla mail and other data unprotected.

I get around this two ways. First, my regular xcopy backup batch file specifically backs up the Mozilla profile directory, but not its subdirectories. The /mail directory is backed up in a separate line in the batch file. I use that method for most of my systems. Second, once Mozilla is up and running, it's possible to specify a different location for data files. I use that method on my den system, redirecting Mozilla data to a server data directory that is always backed up automatically.

One of my subscribers just posted a link to what may be a better solution for some people. Mozilla Backup is a third-party utility that automatically backs up all Mozilla data to a designated location. I've downloaded the program, but haven't tried it yet. Try it at your own risk.



Tuesday, 11 November 2003

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Forums] [HardwareGuys.com Forums] [TechnoMayhem.com Forums]
{Five Years Ago Today]

10:20 - Subscriber Sean Donnelly sends samples of the power supply testers he builds and offers for sale. I haven't had a chance to play with them yet, but they appear to be robust and well-constructed.

pstester.jpg (44531 bytes)

Sean sent me an informative letter with the testers. I'm going to ask him to send me that letter electronically so that I can post it.

SpamAssassin let me down yesterday, not that it's really responsible. For the first time in a very long time, I got a false positive that mattered. My agent mailed me a question about the contract negotiation for the new book, and SpamAssassin flagged it as junk. Fortunately, I caught it while deleting the real junk messages in large batches.

What I need is a quick and easy way to add whitelists of specific email addresses to SpamAssassin. I really wish I could just whitelist domains, but of course that won't work because spammers forge headers so frequently. Brian and Greg have given me a web page that allows me to whitelist addresses, but I've somehow misplaced the URL for it. Ideally, what I'd like is a way to add them in batches. Next-best solution would be to allow me to upload an ASCII file with one email address per line. Best solution would be to allow me to upload my address book periodically and have SpamAssassin automatically extract the addresses and add them to the white list.

I have mixed feelings about Mozilla's Junk Mail controls. I have them enabled, and they do indeed catch most of the spams that SpamAssassin misses. But the downside is that they're also much more subject to false positives. In fact, that's why I've started manually scanning my trash folder. Messages that SpamAssassin flags as spam go to my trash folder marked as read. Those that Mozilla flags as spam also go to my trash folder, but they remain marked as unread. That makes it easy to sort by read status and manually check the ones that Mozilla flagged.

Since I'm doing that anyway, it doesn't take much longer to check the ones that SpamAssassin flagged as spam. Once of the nice things about having so many email addresses is that I usually get multiple copies of each spam. That may not sound like a good thing, but it makes it much easier to delete spams in batches. I simply sort my trash folder by subject and scan 20 or 30 message subject lines at a time. I select them in batches and delete them, which takes only a few seconds per batch of 20. This morning, for example, I had more than 300 spams in my trash folder. It took less than a minute to review the subject lines and delete all of them.

11:41 - Thanks to Roland Dobbins for a heads-up on this Techworld article. The subject line of Roland's message was "The Big Lie", which pretty much sums it up. The article is entitled Microsoft prepares security assault on Linux, which is a poor choice of words. In actuality, Microsoft is preparing a FUD assault on Linux. From the article:

"Microsoft’s aim is to undermine critics and place a question mark over Linux’s security by revealing that, on average, Windows poses less of a security risk."

The pathetic thing is that Microsoft may actually be able to convince non-technical decision makers that this Big Lie is true. Anyone who knows both operating systems understands that Linux is an order of magnitude more secure than Windows, and that what security flaws are discovered in Linux are patched in hours or days, versus weeks to months (or even years) for Microsoft flaws. But that won't stop Microsoft, with its huge war chest, from trying to convince everyone from Joe Sixpack to Aunt Minnie to non-technical decision makers that Microsoft Windows has fewer security flaws than Linux and patches those flaws more quickly. It's disgusting, of course, but there it is.



Wednesday, 12 November 2003

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Forums] [HardwareGuys.com Forums] [TechnoMayhem.com Forums]
{Five Years Ago Today]

8:58 - I really must do something about my batch-file xcopy backup routine. I run it several times a day to copy new and modified files from my working directory to other volumes on the network. The idea is that if something happens to a file or a hard drive dies, I have reasonably current copies of my data distributed on other hard drives throughout the network. The problem is that as the amount of data in my working directories continues to grow, running that batch file takes longer and longer each time. I do what I can to keep the size of the working folders small by archiving infrequently used data to other folders, but that's not a real solution.

This morning, as my batch-file xcopy backup was running, I typed "rsync for windows" into Google, which returned this page. The advantage of rsync is that it copies only changed bits rather than changed files, making the copy process a lot shorter. I think I'm going to play around a bit with rsync for Windows. I know I can use Windows Task Scheduler to automate the process, for example by running it every ten minutes. I can also run rsync for Windows on multiple machines, which would allow me to replicate my data store automagically to multiple systems. What I don't know at this point is how rsync handles exceptions such as open files, and whether I can exclude such things as the temporary files Word opens while I'm editing a file.

Several people expressed interest in the ATX power supply testers that Sean Donnelly makes. He'd enclosed a hard-copy letter with the testers, so I asked him to send it to me electronically. Here it is, edited only for formatting, and posted with Sean's permission:

Here is a check for subscription renewal, along with a couple of freebees.

First the niceties

1- thanks for the web site. I only access once a week on Mondays for the previous week. Then I print it out for bed time reading Monday nite. The good stuff I read to my wife, as she gets a kick out of the trials and tribulations which sometimes are somewhat similar to ours.

2- Sympathies concerning your Mom. Mine passed away 15+ years ago suddenly while having an afternoon nap. I don't know how I could have handled a situation like yours.

3- I have a wonderful view from my apartment, and saw/see Mars in all its glory straight out my living room window. Didn't even have to fire up my crappy Meade with a 12x eyepiece. My wife jokes that Mars is really my Home planet.

4- BTW yours and Dr. Pournelle's site are the only 2 daily sites that I visit.

Now the Testers and Blatant marketing

I started building computers seriously about 5 years ago, having tinkered with the hardware for the last 15 years. Building computers dovetails nicely with my other interests, namely electronics/electricity/photography/2 way communications and Telephony.

Anyway as far as computers go, once you have rolled your own, you don't go back.

After smoking an ATX mobo with a defective power supply, I bought a PC Power and Cooling ATX Power Supply Tester.

Soon realizing its limitations, I re-designed it, adding proper informative LED's and additional pins for accurate Digital Voltmeter testing.

Then I upped the ante with my Heavy Duty model for serious "burn-in" duty.

I also designed the Test Sheet/report.

I manufacture sell a handful of them weekly.

Use them well.

The very first steps in "roll your own"

So for the "built it yourself" stuff, here's my $0.02 for the box/power supply/mobo

1- cooling, sharp edges, looks

2- if the box comes with a power supply, take it out, weigh it, open it up, look at the components. The warranty of the PS at this point is irrelevant as the down side is disastrous. Heavy with big fat Caps, and I mean big and fat, usually mean a decent PS.

3- the PS alone should be worth at least $45, and be traceable to a manufacture's web site. There are plenty of name brands other than Antec /PC Power Cooling.

4- Always burn-in the PS before attaching to the Mobo.

5- Always burn in the ps/mobo/cpu/mem overnight before adding any other components to the system build. Use a scrap Video card during this burn in as you don't want to toast your latest nVidia/Ati card in this process.

6- Mobo choice is the most important decision at the onset. As the old Fram Ad goes, "Pay me now or pay me later" . I prefer Aopen, Msi, etc with intel chipsets only because they work out of the box, and continue to work later. You can wave the flag for AMD/Via/Ali and on and on, what counts is : can I do my work/play on the computer reliably after its up and running ? I paid my dues with the others and have seen the light.

7- Don't buy cheap memory. IMHO speed is less relevant than quality. I've settled on Crucial for years. Market rates, free shipping, guaranteed for life. Works for me!

8- Over clocking is now irrelevant, even a "value" system is light speed fast. I think this happened around 600 speed, when the hardware got faster / better than the software OS and Applications. So go for stability, as speed is now built in!

Finally , it amazes me when I decode computer ads, how they show just the outside of the box, and by the time you read the fine print on that $499 Dell/Gateway etc. that you "get less" than what you pay for. For example many of the cheap ones have integrated graphics, and NO Agp slot for future upgrade. In other words you are "done", right out of the box.

I could go on for another 10 pages in the build it yourself.

Currently I build about one a week, custom orders for profit, and I don't have component failures .

"Burn it In before or Burn it Out later" That's my motto !

Good luck on the book, would be happy to contribute.




Thursday, 13 November 2003

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Forums] [HardwareGuys.com Forums] [TechnoMayhem.com Forums]
{Five Years Ago Today]

9:05 - We had an interesting astronomy club meeting last night. The club meets at SciWorks. Duke Johnson, who was a member of the astronomy club and worked at SciWorks, has taken a new job in Salt Lake City. His replacement apparently forgot about our meeting, so we all arrived at SciWorks only to find it dark and locked up. Fortunately the weather was warm and dry, because we ended up having our meeting out front on the sidewalk.

Our speaker, Mike Castelaz from Pisgah Astronomical Research Institute (PARI), had driven three hours to get here. Fortunately, he had a projector with him and we were able to find an outdoor power receptacle for him to plug in to. He ended up projecting his presentation on the side of the SciWorks building. Mike is working on a project to learn more (actually, anything at all) about the stars that are at the center of the protoplanetary discs (proplyds) in the Orion Nebula. Proplyds resemble our solar system soon after its birth, and Mike is working to discover the nature of the stars that these proplyds surround. His presentation was very interesting. He's probaby a bit hoarse this morning from having to shout over the wind so we could hear him.

In the past, the Forsyth Astronomical Society (FAS) in conjunction with SciWorks has been doing four public observations a year at Pilot Mountain State Park. There had been some discussion of expanding that to six sessions a year, but last night when Steve Wilson announced our 2004 schedule it turned out that we'll be doing only four again next year. The problem is in getting enough people with scopes to participate in these public observations. Apparently, it's hard enough to get out the membership four times a year, and six times would really be pushing it. That's a shame, because these public sessions fit well with the public outreach mission of FAS and are also an excellent way to drum up new members. The rangers at Pilot Mountain State Park like us to do these sessions, because they draw a lot of people to the park and support the public outreach mission of the park itself.

Neither Barbara nor I particularly like attending these sessions, but we do it because it's a Good Thing to do. Neither of us is particularly comfortable around large groups of children, and there are children galore at the public sessions. Scopes have come close to being knocked over, and periodically one of the members has something stolen. There are also other dangers, including women wearing too much mascara and people putting their fingerprints all over eyepiece lenses. We always use inexpensive eyepieces and clean them afterward.

But, despite the aggravations, it's a worthwhile pursuit. I'm always conscious that one of those little kids might be a future Nobel Prize winner, and that public observation session may just be the thing that gets him interested in science. I think back to when I was one of those little kids running around, and realize that the adults that were introducing me to science may not have been having a particularly good time, either. Jerry Pournelle sometimes tells the story of his friend Robert Heinlein helping him get started as a science fiction writer. When Jerry asked Mr. Heinlein how he could pay him back, Mr. Heinlein replied that he couldn't pay back, he could only pay forward. So attending these public observations is one way I pay forward.

Several people have suggested alternatives to rsync, including Microsoft's robocopy, Pixelab's xxcopy, and Unison (that link returns a 403 Forbidden today, but worked yesterday and presumably will work in the future). I've actually used the first two at one time or another. In fact, I'm sure I've recommend xxcopy in these pages in the past. But for some reason, probably one of the periodic network rearrangements we do around here, I'd "lost" those programs and was no longer using them. I'll have to check all three to see which I prefer.



Friday, 14 November 2003

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Forums] [HardwareGuys.com Forums] [TechnoMayhem.com Forums]
{Five Years Ago Today]

9:35 - One of my readers tells me that there can be an amusing side to spam. Her husband, Richard, is 6'9" (206 cm) tall, weighs 280 pounds (127 kg), and is "built like a lumberjack". One of the spams that made it through her filters and into her inbox the other day gave her the giggles all morning. The subject line? "Want a Bigger Dick?"

Stuff is starting to accumulate for the new book project. I have Intel processors, including a matched set of Pentium 4/3.2G's, one standard and one Extreme Edition (with 2 MB L3 cache), Intel motherboards, AMD processors and motherboards, Seagate and Maxtor P-ATA and S-ATA hard drives, various DVD writers, video cards including an ATi RADEON 9800 XT, various cases, and so on. Much more is on the way or will be. Poor Barbara. All of the horizontal surfaces in the house will soon be covered with computer "pieces parts" as she calls them.

One of the very interesting components I have is an Intel D865GRH motherboard, which as far as I know is the first shipping motherboard with embedded DRM hardware. I intend to work with it quite a bit. I realize that the knee-jerk libertarian/slashdot reaction to DRM-enabled hardware is unfavorable, but DRM truly is a two-edged sword.

Just as a padlock can be a Bad Thing if it's keeping you in a jail cell but a Good Thing if it's protecting your property, I can see that DRM-enabled hardware can have advantages and drawbacks from the perspective of us, the users. It's all a question of who has the keys. If the Bad Guys like the RIAA or the MPAA or the federal government or Microsoft have custody of the keys, DRM is nothing less than the chains of slavery. Same thing if DRM can be enabled or disabled only system-wide. For example, if running a new Microsoft operating system requires that DRM be enabled and if enabling it in order to run that Microsoft OS also enables it for everything on that system, then DRM is unacceptable.

On the other hand, if we can pick and choose where and how to enable DRM, it could be a very Good Thing. We could, for example, enable DRM to protect the directories where we store our own data while disabling it for the directories where we store audio and video files we've ripped from CDs and DVDs. It's all a matter of who has the keys and who controls the locks. That's part of what I need to find out.

13:53 - And there's yet more stuff on the way in, including a bunch of stuff from Logitech. I've recommended Logitech speakers for years, but I haven't looked at some of their new models. It's also time to take another look at Logitech keyboards and mice. I'd tried those in the past, but always gravitated back to Microsoft keyboards and mice. However, Logitech has some interesting new stuff out, including optical mice with resolution that matches or exceeds Microsoft models.

There'll also be some interesting new stuff coming from Antec, including their new P160 aluminum case, which turns out not to be made from recycled fighter jets. When the product first appeared on their web site a couple of weeks ago, there wasn't any asterisk near the headline, "Made from recycled fighter jets." Call me gullible, but I believed them. I figured they'd contracted with a salvage company to buy the aluminum from crushed MiGs or something, just as a marketing stunt. As it turns out, they were joking, but to their credit they immediately added the asterisk and note when it became clear to them that some people believed that the new cases were really made from recycled fighter jets.

Barbara really will be delighted. She just loves it when this place starts looking like the back room of a screwdriver shop.



Saturday, 15 November 2003

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Forums] [HardwareGuys.com Forums] [TechnoMayhem.com Forums]
{Five Years Ago Today]




Sunday, 16 November 2003

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Forums] [HardwareGuys.com Forums] [TechnoMayhem.com Forums]
{Five Years Ago Today]




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