Daynotes Journal

Week of 25 December 2000

Latest Update: Friday, 05 July 2002 08:16
 

Order PC Hardware in a Nutshell from Fatbrain.com

Monday, 25 December 2000

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

Barbara and I were awakened in the middle of the night by the roar of the ZSU-23-4 antiaircraft gun cutting loose at Santa, and the tinkle of hot brass falling to the ground as it was ejected. I ran outside, expecting to find Santa's sleigh shot up and burning nearby, surrounded by dead reindeer, but there was nothing but a fading "ho-ho-ho" sound. I later found a small gift-wrapped box sitting on top of the ZSU-23-4. After X-raying it and putting it through the explosives sniffer machine, I opened it, not knowing what to expect. All I found was a small lump of black, rock-like substance. I'm not sure what to make of that.

Barbara's family is coming over this morning. We'll have brunch, gift exchanging (ugh), and then dinner. Have I mentioned how much I truly hate Christmas? At least we're having baked Cindi Lu Who for dinner tonight.

Click here to read or post responses to this week's journal entries

[Top]

Tuesday, 26 December 2000

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

I am sorry to report that PC Hardware: The Definitive Guide has been cancelled by O'Reilly. Neither Pournelle nor I is happy about that, but such are the realities of publishing. The good news is that O'Reilly is so happy with PC Hardware in a Nutshell that we're already discussing plans for the second edition. In fact, according to O'Reilly, the quality of PC Hardware in a Nutshell is the major reason why they decided to cancel the second book. They just didn't see a place for that book in the market. I don't agree, but it's their call. 

When I called Pournelle to tell him what had happened, he was disappointed, obviously. But Jerry was kind enough to tell me that if he had seen the completed PC Hardware in a Nutshell before starting our book, he wouldn't have bothered signing on for the project because he didn't think we could do much better together than Barbara and I had already done with PCHIAN.

So Barbara and I are spending some time now planning the second edition. It will be updated, of course, but it will also be larger and cover some additional topics. Don't let that dissuade you from buying the current edition, though. The second edition probably won't hit the bookstores until late next year. If you haven't bought a copy of PC Hardware in a Nutshell yet, you can do so by clicking on the link at the top of the page. We get a small commission from Fatbrain if you buy directly from that link.

Click here to read or post responses to this week's journal entries

[Top]

Wednesday, 27 December 2000

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

There is something purely evil about a government which admits that it cannot protect its citizens and yet forbids those citizens from protecting themselves. All the government can do now is prosecute and imprison the maniac. That must be a great comfort to the families of those who were murdered.

Thanks to everyone who wrote expressing condolences about the cancellation of Pournelle's and my book. The most common question was whether it could be published elsewhere. Unfortunately, the answer to that is no. Our book was based upon PC Hardware in a Nutshell, to which O'Reilly holds the copyright, both for the original book and for derivative works. Even if Pournelle and I rewrote the entire thing from scratch, we still wouldn't be able to publish it because that would mean I was going into direct competition with my own book. That's prohibited by contract as well as by common sense.

The cancellation of that book means my work schedule for 2001 is going to look very different from what it otherwise would have been. I'll do the PC/Nut update, of course, and I may do another PC hardware-related book for another publisher, perhaps an A+ Certification title or something. My agent is also looking into possibilities for me doing a PC hardware-related column for one of the on-line publications. Who knows? I might end up doing a column for BYTE.com. And then there's the book project that Barbara's been working on, which we will publish ourselves, although we've not yet decided whether we'll publish it as a traditional book or as an e-book. More on that later.

Click here to read or post responses to this week's journal entries

[Top]

Thursday, 28 December 2000

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

The scanner, an HP 6200C, is also a problem. I had it working via the SCSI interface under Windows NT 4, but getting it working under Windows 2000 is likely to be a problem. When we crawled behind the machine to connect the scanner yesterday, I was forcibly reminded that I'd installed an Adaptec 2940U2W host adapter in that machine. The external connector on that card, of course, doesn't mate with the SCSI cable from the scanner. The 6200C also has a USB interface, but I've had nothing but problems attempting to use that. I may give it another try anyway.

And Barbara mentioned that I hadn't installed Plextor Manager 2000, which she used to listen to audio CDs. No problem, I told her. Windows 2000 has a decent CD player application built in. So she inserted an audio CD, which started to play. But no sound came from her speakers. As I was checking audio settings, volume controls, and so on, Barbara said, "I'll bet you forgot to install a CD audio cable again, didn't you?" 

Yep. I always forget to install a CD audio cable. I muttered some excuse about assuming the Plextor UltraPlex Wide supported digital audio and so didn't need a CD audio cable, but the truth is that I just forgot to install the cable. I don't know why I have this mental block, but I don't think I've ever remembered to install a CD audio cable on any system I've built. I always have to open the system and install one after the fact. This also means I have to find a CD audio cable. I know I have a bunch of them around here. That's another item that I always buy two or three of whenever I need to buy one. I then stick them on the shelf so that I'll have one the next time I need one. When that time arrives, I can't find the spares I've bought, so I go out and buy three more. Somewhere in this house lurks a tangled bunch of CD audio cables, but I sure don't know where. They're probably with the spare IDE, SCSI, and serial cables that I can never find either.

Yesterday afternoon, we visited the library, where I picked up several books to add to my to-be-read stack. A couple of those were Harry Turtledove alternative history novels. I can't believe how much work that guy turns out. He has about three different series running, and he seems to turn out at least a book a year in each of them. Not small books, either. Typically 500 pages or more. At any rate, if you like alternative history novels, his are worth reading.

After the library, we went to dinner at the newly-opened Rose & Thistle restaurant. We used to go there frequently, when it was located on the other side of town where we then lived. Then, about five years ago, the R&T was forced to close down when the state decided to re-route Interstate 40 through down town Winston-Salem. It didn't reopen elsewhere, and at the time many people assumed it was gone for good. So it's nice that they've re-opened, and near where we now live at that. The new place has a small stage for performers, and we're hoping that the R&T goes back to their old custom of having live music. Barbara used to belong to the Fiddle & Bow Society, whose members frequently performed traditional Celtic music at the old R&T.

Click here to read or post responses to this week's journal entries

[Top]

Friday, 29 December 2000

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

It has been my habit not to install the Microsoft games on my main systems, or to uninstall them if they were already present. That way, I'm not tempted to play games when I should be doing something else. Yesterday, I noticed that when I ran Windows 2000 Professional setup on my new main system it had installed these games. So I went to Control Panel -> Add/Remove Programs to uninstall the games. That's not an option, nor do I remember being offered the option not to install the games when I ran Windows 2000 Professional setup. I appears that Microsoft really, really wants you to have these games installed.

So, lacking a formal mechanism for uninstalling the games, I fired up Explorer and changed to the \WINNT\system32 folder, where the executables (Freecell.exe, Sol.exe, and Winmine.exe) reside. I deleted all three of them. But, as I said, Microsoft thinks these games are critical system components, so it promptly put back the files I'd just deleted! Okay, I know how to deal with that. I changed to the \WINNT\system32\dllcache folder and deleted the master copies of those programs from that folder. I then changed to \WINNT\system32 and again deleted the executables. 

Up popped a horrifying warning dialog, telling me to insert my Windows 2000 Professional CD so that Windows could copy these critical system files from it. I clicked Cancel to tell it that I really wanted to delete these files, and it popped up another horrifying warning dialog before finally accepting my decision. If I believe what that dialog says, my system may now be unstable, and all because I chose to delete those games.

So the question is, why does Microsoft consider these games to be critical system components? It would certainly have been easy enough to install those files in a folder that wasn't delete-protected, or alternatively simply not to include copies of those files in the dllcache folder. Come to that, it would have been easy enough not to install those files by default, or at least to offer an option not to have them installed.

We tore down old thor yesterday. Thor is/was an IDE test-bed system built around an Intel CA810E motherboard, a Pentium III/600, 64 MB of Crucial PC100 RAM, a Maxtor 10 GB 7,200 RPM ATA drive, a Smart & Friendly CD writer, an Hitachi DVD-RAM drive, and an OnStream tape drive, all surrounded by an Antec KS-288 case. I've finally decided to get serious about Linux, and so I need a competent system for running Linux as a desktop OS.

After tearing down the box and vacuuming out all the crud, we pulled the DVD-RAM drive and the OnStream tape drive, replacing that with an old Seagate/Connor Travan TR-4 ATAPI tape drive which should be adequate for backups and should be easy to get running under Linux. While we had the case open, I decided to upgrade the memory to 128 MB, which seems a fair amount for running Linux in GUI mode. So I pulled the original 64 MB Crucial PC100 DIMM and installed what I thought were a matched pair of older Crucial PC100 DIMMs to replace it.

After we reassembled the system, we stuck it under my desk, connected it to the Belkin OmniCube KVM switch, and fired it up. The first sign that all was not as it should be was that the memory count showed only 64 MB. The next sign was that the video started pulsing on and off about once a second--half a second on, half a second off, ad infinitum. I'd never seen that behavior before. Given the first problem, I thought it likely that what I'd installed was two 32 MB DIMMs rather than two 64 MB DIMMs, which made me wonder if they were PC100 as I'd thought. 

Apparently not. I haven't looked up the part numbers, but I think what I ended up doing was installing two 32 MB PC66 DIMMs. That's the hazard with my method of salvaging memory when swapping components around. I pull the old DIMMs and stick them in whatever antistatic bag happens to be handy. In this case, I think I stuck old PC66 memory in a bag labeled as PC100 memory. The CA810E motherboard is interesting. It requires PC100 memory regardless of the FSB speed of the installed processor. That is, if one installs a 100 MHz FSB processor, that motherboard requires PC100 memory. If one installs a 66 MHz FSB processor, the motherboard still requires PC100 memory. Even if one installs a 133 MHz FSB processor, the motherboard still requires PC100 memory. Oh, well.

That was easy enough to fix. Old kiwi is still sitting nearby with the side panels off, and it still had one of the two 128 MB PC100 Crucial DIMMs that was originally installed in it. So I pulled that DIMM and installed it in the new system. I pulled the two 32 MB PC66 DIMMs and put them back in the original bag, still labeled PC100. When will I ever learn? With the PC100 DIMM installed, the new machine booted and ran normally.

While all that was going on, I was downloading two ISO images for Linux Mandrake 7.2. Sometime in the next few days, I'll burn those to CD-R discs and build a Linux workstation system on this box. Linux can't yet compete evenly with Windows as a desktop OS, but it's getting closer every month. When the day arrives that Linux is a viable mainstream competitor to Windows on the desktop, I intend already to be using it. The only way that's going to happen is if I start using it now.

Before I install Linux, though, I have some testing and benchmarking to do early next week. Intel is supposed to be sending me one of their new Celeron/800 processors, which uses the 100 MHz FSB. It'll be interesting to see how a Celeron/800 compares with a Pentium III/800, now that the only difference is L2 cache. The Pentium III, of course, has 256 KB of 8-way associative L2 cache, whereas the Celeron has 128 KB of 4-way associative L2 cache. My guess is that the 100 MHz FSB Celeron will end up benchmarking only a few percentage points slower than the Pentium III, which presents Intel with an interesting positioning problem. 

Clearly, clock-for-clock, the Pentium III can't sell for much of a premium over the 100 MHz Celerons, so it may be that Intel will adopt AMD's scheme of avoiding speed overlaps between their value processor and their premium processor. The problem with that scheme is that the Coppermine Pentium III effectively tops out at 1,000 MHz, which doesn't leave much headroom above a Celeron/800. The problem becomes worse when you realize that Intel plans to release faster versions of the 100 MHz Celerons in the near future to stay competitive in clock speed with the AMD Duron. So where does that leave the Pentium III? Out in the cold, I'm afraid.

I think that's why Intel is pushing the Pentium 4 so hard. They need to maintain a clock speed advantage over the AMD Athlon at the high end, and at least match clock speeds at the low end with the Celeron. Given that the Coppermine core effectively tops out at 1,000 MHz, whereas the Duron has some headroom above that, my guess is that Intel will transition to the Pentium 4 for their performance processor and a 130 nanometer Celeron with copper interconnects for their value processor. That'd allow them to beat AMD on clock speed in both segments. It's going to be an interesting 2001.

The Register has an interesting article today about CPRM, the copy-protection scheme for hard drives I mentioned earlier. It's a FAQ that points out what CPRM is all about and what you can do to stop it.

Click here to read or post responses to this week's journal entries

[Top]

Saturday, 30 December 2000

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

That brings up the problem of adding the 13th month. We can do that the same way that July and August were added (that's why September, the "seventh" month, is the ninth month in our current calendar, and the eighth through tenth months, October through December, are now instead the tenth through twelfth months). I propose that we insert the new month in the middle of the existing calendar, where it will fall between June and July. We need a name for the new month. Just as July was named for Gaius Julius Caesar and August (the "eighth" month) for Octavian, also known as Augustus, so I propose that the new month be named for me. Of course, I haven't conquered anyone lately or been a Roman Emperor in this life, so that may not fly.

Rationalizing the calendar would have many advantages, but there's no doubt that it would also introduce some complications. For example, nearly every computer program worldwide would need to be updated to handle the new scheme. Also, there will be some confusion between old-style (OS) dates and new-style (NS) dates. For example, Monday 1 February 2001 (NS) corresponds to Monday 29 January 2001 (OS). Still, that should be easy enough to get used to. I blame myself for not getting started on all of this earlier. All that time spent on Y2K fixes could have better been spent implementing my cunning plan. 

The new Linux box (which isn't yet a Linux box) is behaving strangely. Not surprising. When I was tearing it down and rebuilding it, I did some swapping around. I pulled the Hitachi DVD-RAM drive and the OnStream DI30 tape drive that were in it (both ATAPI), and installed an old Connor/Seagate TapeStor Travan TR-4 ATAPI drive. The hard drive is primary master, and (I assume) the Smart & Friendly CD burner is secondary master. Since I was in effect just substituting one tape drive for another, I foolishly assumed that both had been jumpered slave (which is what I normally set tape drives to), and I didn't bother to check. Now Windows 98SE Device Manager is totally confused. It tells me that there's no secondary IDE interface present, although it recognizes the SAF CD burner that's connected to that interface. It also sees the tape drive that's connected to that interface, but as an "Unknown device". If you ever encounter a similar situation with IDE devices, it's almost 100% certain that you've misjumpered one or more devices.

And that brings up another interesting point. That TR-4 tape drive had been installed in an old Gateway Pentium/133 full tower system. As I was pulling the drive, I happened to glance at the power supply installed in that system. It was, if you can believe it, a 145 Watt unit! It's simply incredible to me that anyone would install a 145 Watt power supply in a full tower system. What's even worse is that I'm almost sure I paid for an optional power supply upgrade when I bought that system four or five years ago. What on earth would they have installed if I hadn't ordered the upgrade? Oh, well. Gateway is known for installing low-output power supplies in their systems. But a 145 Watt Astec unit? Give me a break.

Speaking of calendrical issues, I've officially deemed the last nine days of the year to comprise a single week for some purposes around here. Barbara's daily journal page is one example. She started her weekly page on the first of this year, so the last day on each of her pages has been Friday throughout the year. Instead of creating a "Week 53" page for her that had only two days in it, or that had two days of this year and five days of next year, I decided simply to tack the last two days of this year onto her final weekly page. So she'll start the year with a new weekly page that begins on Monday, 1 January. Same deal on the web access reports that I run for Pournelle and me. I started those for the year on 1 January 2000, so they've been running Saturday through Friday. I normally generate reports for the preceding week, the month-to-date (or the full calendar month, once all days' data is available), and year-to-date. This time, I'll generate reports for the full calendar year, the calendar month of December, and a "weekly" report that includes the final nine days of the year. Then I'll start the cycle again with the first day of 2001, shifting from a Saturday through Friday reporting week to one that runs Monday through Sunday.

Click here to read or post responses to this week's journal entries

[Top]

Sunday, 31 December 2000

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

If you want to order the EB on DVD-ROM for $25, you need to do so now, because that offer expires as of today. Make sure you have Javascript and cookies enabled in your browser. If you don't, you see an apparently normal page, but there's no way to order.

I haven't been keeping very close tabs on our SETI@home effort lately, but I did check this morning. Our group was sitting at 37,213 work units complete. That puts us about 400 work units short of appearing on the Top 100 Clubs list, albeit in 200th place. I never could figure out why SETI top 100 lists have 200 members, but I just figured it was an in-joke or something. The Top 100 Clubs list is the tough one. For any of the other lists--educational, business, etc.--we'd already be in at least the top 100, if not in the top 20. The "Clubs" group is where the action is.

At any rate, we've been overtaking the current 200th place team pretty quickly, and we're cumulatively doing something like 250 work units/day, so we should be on that list shortly. At that point, we'll be only about 3.2 million work units behind the current top group, Ars Technica's Team Lambchop. Before long, we'll have passed at least one team that represents an entire state. A lightly-populated state, true, but a state nonetheless. And there's a small country or two that probably hears our footsteps. Overall, we're doing really well.

To help out a little bit, I just added two more machines to our local SETI effort last night. Barbara's Pentium III/1000 and my Pentium III/800. Between them, they should do about 8 work units per day. That's with the 2.04 client, of course, which is considerably faster than the 3.x client, although it does less scientifically useful work, says SETI. I avoided upgrading when the 3.0 client was announced, both because I didn't have time to deal with it and because I don't trust dot-oh releases. As it turned out, there were some problems with the first 3.0 release, so I'm glad I waited. I'll probably upgrade after the first of the year. At some point soon, it will no longer be optional, as SETI will stop accepting 2.x results.

This being the last day of the year, we'll be starting Barbara's patented Deep Solstice Cleaning®. She does one room at a time, and she basically strips that room down to bare metal, reformats, and then reinstalls everything. I mean, we're talking serious cleaning here. She actually removes the cover plates from all the receptacles and washes them, for example. The cover plates, that is, not the receptacles. She'd vacuum out the receptacles, too, if she weren't afraid of electricity. I am not making this up.

Like any man, I can only do my best to appear to be useful. So I try to be useful, which is mostly a ruse to allow me to rescue Valuable Stuff that Barbara considers Useless Junk and is about to discard. Nearly all men, of course, lack the cleaning gene that is present in most women. What most men regard as an environment so clean that it squeaks is viewed by most women as a pit of filth and squalor. 

And speaking of pits of filth and squalor, I need to clean up my office before Barbara decides to do it for me. Pournelle tells me that if he waits too long to clean the Great Hall, Roberta does it for him. The Great Hall is located on the second floor. Roberta literally rolls a dumpster under the balcony, and Jerry sometimes returns home to find her carrying Valuable Stuff out to the balcony and shouting "bombs away" as she drops Useless Junk over the edge and into the dumpster. Shudder.

I'll be back sometime next year.

Click here to read or post responses to this week's journal entries

[Top]