|
Week of 2/8/99 Friday, July 05, 2002 08:08 A (mostly) daily journal of the trials, tribulations, and random observations of Robert Bruce Thompson, a writer of computer books. |
|
Monday |
Monday, February 8, 1999 I spent most of yesterday cleaning house, doing laundry, and working on the web site. For only three adults living in the house, we generate a lot of laundry. Yesterday, it was six loads: white, colored, a load for my mother, a load of towels, a load of blankets, and a load of rugs. This in the largest size washer available to consumers. * * * * * And I also spent some time messing around with my ATX test-bed system. I'd loaded the 1999 edition of the PC Magazine benchmark on it, and was playing around running them under various configurations. Before it runs, the benchmark program checks the system to locate problems that would prevent the benchmarks from being accurate. The only show-stopper was that the Taskbar had the "Always on top" property set, which for some reason ZDBench takes strong exception to. It was easy enough to turn that off. But it also found a non-fatal problem. The program Pstores.exe was running. I'd actually noticed that program running in the past when I'd run Task Manager and looked at the Processes tab, and wondered what exactly it was. I think I had some vague idea that it had something to do with Outlook 98. But OL98 wasn't running on this machine. I was pretty sure that Pstores.exe wasn't there when I'd first installed Windows NT 4. so it must have been installed when I applied the SP4 patch. But I wasn't sure about that. I thought perhaps it had something to do with the fact that I had installed a 10 GB hard disk in this machine that perhaps required some kind of special support. So, I set off in search of information about what Pstores.exe was and what it did. I checked TechNet and the MS web site. I found a few hits, but nothing useful. The one helpful piece of information I did find was that it was associated with the Protected Storage service. I fired up Control Panel - Services and, sure enough, there was a service named Protected Storage running. I could have simply stopped that service, but stopping a service that I don't know the purpose of seemed like a bad idea. So I searched some more, this time for "protected storage." This time, I found enough hits to learn that Protected Storage is installed when one installs Internet Explorer 4. I still don't know what Protected Storage does, but I wasn't planning to use IE4 on this test bed machine anyway, so it seemed safe to stop the Protected Storage service. I did so, and the machine continues to run with no apparent bad effects. * * * * * And my friend Paul Robichaux posted this message to the Computer Book Publishing mailing list:
If this is true, and I have no reason to believe otherwise, this is a complete abuse of trust on Amazon's part. It brings to mind the "Playola" scandal in radio back in the fifties, when record companies paid disk jockeys to feature their products. It seems to me that any reputable person or company takes great pains to make clear the distinction between editorial content and paid advertising. You may be sure that I'll continue to follow this thread. If it turns out that Amazon is behaving this way, I'll remove all links to Amazon from my site and stop buying books from Amazon.
|
|
Tuesday |
Tuesday, February 9, 1999 Well, it appears that the reports about Amazon soliciting payment from publishers for featured treatment are true. Reuters published a report last night that confirms it. Amazon admits to doing this, but says there's nothing wrong with what they're doing. I disagree, and many people feel the same. Accepting payment for prominent placement is one thing. Coke and Pepsi pay big bucks for those displays at the ends of supermarket aisles. But accepting payment to feature a book as "Destined for Greatness" or "What We're Reading" crosses the line by implying an unbiased judgement where in fact that judgement has been bought and paid for. Visitors to Amazon's web site might reasonably expect editorial and advertising material to be kept strictly separate. That's obviously not the case here, although it won't be obvious to casual visitors, so anyone who gives any weight whatsoever to Amazon's recommendations does so at his own risk. I will no longer be inserting links to Amazon for any books I recommend. They pay me a small percentage of the purchase price for books that my readers buy by clicking on one of my links, but I've always been completely up-front about that. I will no longer provide such links, and I will remove existing ones as I come across them on my pages. What I'll probably do is sign up for one of the associates programs offered by Barnes & Noble, Computer Literacy, or one of Amazon's other competitors. * * * * * I see in the morning paper that FreePC is giving away very low-end Compaq PCs to people who are willing to provide their personal information and look at a constant barrage of ads. And I mean constant. Ads run the whole time the PC is powered on, whether you're on-line or not. Half of the supplied 4 GB drive is allocated to local ad storage. The box runs 1024X768 video (on a 15" monitor, yet), of which an 800X600 window is available for use. The remainder is used as a frame that constantly displays ads. Free-PC plans to ship their first 10,000 computers in Q2/99, and have announced that they may ship as many as 1,000,000 units. Actually, looking over their page I see that they aren't really giving away the PCs, they're lending them for two years. Their FAQ page says that "Free-PC was founded in 1998 by Bill Gross to provide the power of personal computing to those who might not otherwise be able to have access to computers and the Internet." Yeah, right. I don't think many poor people are going to end up with one of these units. The FAQ goes on to say, "The first 10,000 customers will be selected from among those with the highest match to the variables of interest to our initial advertisers." I don't imagine there'll be many poor people in that group. Just rich ones stupid enough to take this cheese in exchange for sticking their heads into another marketing trap. I wonder how many people will get one of these things and turn around immediately and format the hard drive. Quite a few, I imagine.
|
|
Wednesday |
Wednesday, February 10, 1999 The Amazon brouhaha appears to be over. Amazon, while still saying they had been doing nothing wrong, agreed to stop doing it. Or at least they announced that, starting March 1st they will label all material for which publishers have paid for placement. The following is the text of a form letter that Amazon began mailing yesterday to people who emailed them to comment about the situation:
* * * * * The following from Scott Kitterman [kitterma@erols.com]. Scott also included the text found at that URL, which I've removed here for brevity:
I agree that it's a satisfactory conclusion, and I will not remove my links. Frankly, I'm glad they did the right thing. I've been buying from Amazon.com almost since they opened, and I've never had a bad experience with them. I'm going to consider this a momentary aberration. I've had numerous messages from people who tell me I'm naive, and that what Amazon was doing was no different than Coke and Pepsi paying big bucks to supermarkets for those displays at the end of aisles. Well, I'm not naive. I've known about such end-cap payments at least since we did cases about them in business school fifteen years ago. But there's a big difference between this and what Amazon was doing. What they were doing was the equivalent of a supermarket accepting payments from Coke, putting up the end cap, and then adding a big sign over it that said, "We put this display here because we honestly believe that Coke is better than Pepsi." At any rate, the problem appears to have gone away, or at least gone back into its hole until later... * * * * * This from Bo Leuf Bo Leuf [bleuf@algonet.se]:
I'm not surprised that this phenomenon is ubiquitous. Advertising resembles weeds. We had something similar in our schools starting several years ago. I can't remember the name of the company, but they provided "free" televisions and VCRs to school classrooms, with the understanding that all students would be forced to watch a 15 minute (or whatever) program each day, which was laden with ads sold by the company that provided the "free" equipment. There was quite a firestorm at the time. Many parents believed, as I do, that this was an abuse of the mandatory school attendance laws. To force children to attend school is one thing. To then force them to watch commercial messages is quite another. As far as formatting the hard drives, come now. I can't imagine that there's anything to prevent someone from booting a floppy and running fdisk. I'm sure you're correct that there will be contractual terms forbidding it, but how is the company to tell the difference between someone who's formatted their hard disk and someone who simply doesn't ever log in to the bundled Internet service? I do hope a lot of people do that, that the advertisers eventually find out about it and demand their money back from Free-PC, and that Free-PC goes bankrupt. It's not like we need any new ad delivery mechanisms. We're drowning in ads now. * * * * * This from John Bartley, PC Sys Admin [usbcpdx@teleport.com], sent both to me and Jerry Pournelle:
Thanks. I'll contact O&O about getting an eval copy. As far as Diskeeper munging a disk, I suspect your colleague may have used the Directory Consolidation feature, which can indeed trash data per the discussions over the last couple weeks on this site. But that problem is due to a bug in NT rather than in Diskeeper. I've used Diskeeper for years and never had a problem with it. In fact, I ran it on an SP4 machine with Directory Consolidation enabled and still didn't lose any data. I've used three of the four products you mention, and my strong personal preference is still for Diskeeper. But I'll check out the O&O product.
|
|
Thursday |
Thursday, February 11, 1999 Well, there's bizarre news this morning. Ron Barker, the local sheriff, has been criticized for years for his hiring and firing policies. He's hired his own relatives--including his son and grandson--a lot of his friends, and a lot of his friends' relatives. If the news stories are to be believed, the Forsyth County Sheriff Department under Ron Barker is one of the most nepotistic fiefdoms ever seen around here. Ron Barker's son, Brian, was hired and promoted quickly, his main qualification apparently being who his father was. Brian Barker, although more than 30 years old, still lives with his parents. About ten years ago, the newspapers reported that he'd barricaded himself inside the Barker home with an AR-15 assault rifle, taken his father hostage, and struck him. The SWAT team surrounded the house and waited it out, although Brian Barker did fire a shot. Ron Barker later claimed that (a) there was no hostage situation, (b) they hadn't known the SWAT team was out there, (c) that the rifle had gone off by accident, and (d) that the cut on his head came from a fall rather than from his son hitting him. A lot of people at the time thought that sounded fishy. A lot more wondered how after that incident Brian Barker could be accepted as a sworn officer. Earlier this week, the newspaper reported that Brian Barker had made a routine traffic stop and had been shot by the people in the car. Although he was given only a 50/50 chance to survive, Brian Barker was able to talk to police officers, and reported that the occupants of the car had been Hispanic men. Naturally, a major manhunt has been going on since then. This morning, the paper reports that it appears that there was no traffic stop, no Hispanic men, and that Brian Barker appears to have shot himself. Hispanic spokespeople are outraged, although I find it hard to understand why. They've equated this situation to the Susan Smith thing several years ago, when she drove her car into a lake, drowned her children, and blamed a non-existent black man for the murder. Being upset with Brian Barker is certainly understandable. The Hispanic spokespeople ask why he had to accuse Hispanics rather than some other group, and in that they have a point. That may indeed say something about Brian Barker's prejudices, although it may also mean nothing. He did, after all, have to make up some kind of description, and that may simply be the first one he hit upon or the one he believed would be most credible given the circumstances. But being upset with the police for investigating on the basis of the information they had available is surely unreasonable. If Brian Barker had reported that he'd been shot by black or white or Asian men, the police would have been looking for people who met that description. To claim that there is some sort of bias against Hispanics in the police department on the basis of this investigation is ridiculous. They did, after all, discover that Brian Barker had shot himself. If they were looking for someone to lynch that would not have happened. * * * * * My agent, David L. Rogelberg [davidlr@studiob.com] of StudioB posted the following message to the Computer Book Publishing mailing list yesterday. I think it's pretty important, because he touches on a very real concern among authors in general and computer book authors in particular. The industry is consolidating into a few large publishers and a few large bookstore chains. This consolidation worries a lot of us because it will almost inevitably lead to fewer choices for readers. The interview David mentions early in the message refers to his appearance on a radio talk show Tuesday evening. He had never been interviewed on radio before, and so solicited advice from other members of the list who had some experience with doing radio interviews. Bruce Epstein is a list member who is notorious for his sharp wit and the pointed observations he makes about just about anything:
* * * * * This followup from John E. Bartley, III ]John_Bartley@orb.uscourts.gov] concerning his report of lost data when running Diskeeper:
Hmm. That's the first time I've ever heard of Diskeeper damaging data, other than the recent reports of problems running chkdsk during Directory Consolidation. My own experience, and those of readers I've heard from, seems to indicate that Diskeeper is a rock-solid product. In the absence of more reports similar to yours, I'm going to keep recommending Diskeeper. But I would like to hear about it if anyone else has experienced data loss or corruption that is certainly attributable to running Diskeeper. I'm sure that ExecSoft Tech Support would also like to hear about them. * * * * * This from Michael A. Boyle [mboyle@toltbbs.com] regarding the Free-PC issue:
That'd work. Of course, the question is how far someone is willing to go to get what amounts to a $500 PC. * * * * * And still another followup from John E. Bartley, III ]John_Bartley@orb.uscourts.gov] concerning his report of lost data when running Diskeeper:
Okay, the first guy runs it for a week and then has a problem which may or may not be caused by Diskeeper. He asks "Could Diskkeeper have caused this? Should I buy, or avoid Diskkeeper 4.0 for my NT Server 4.0 PC?" The second guy "suspects" that he may have a similar problem. He says "a few days after I installed the Diskkeeper 4.0 [...] Could it be that Diskkepper corrupts the NTFS partition?" The third guy reports a known problem that is not caused by Diskeeper at all, but by an acknowledged bug in NT's chkdsk. The fourth guy says, "I'm not sure if it was Diskeeper to cause these problems [...]" So of the four you quoted, three aren't sure that Diskeeper caused the problem, and the fourth is reporting a known bug in NT. From these messages, no one could reasonably attribute the problem to Diskeeper, or indeed even finger it as the likely cause. This is what is called anecdotal evidence, which is no evidence at all. We have no idea what hardware configuration these people were running and whether their systems were fully compliant with the Windows NT HCL. We don't know what other software they were running, or what their administrative abilities are. Before anyone blames a software or hardware product for causing a problem, that problem should be reproducible under controlled circumstances or, if not reproducible, it should at the very least be reported in essentially identical form by many people who are all running on HCL compliant hardware. I've been running NT for a lot of years, and many of the problems I've seen others have have been solely due to running with non-HCL compliant hardware. Again, I can only say that my own experiences with Diskeeper have been uniformly good, both on my own systems and on systems that belong to clients and former employers. I've never seen a Diskeeper bug that would cause data loss or corruption, nor have I ever been presented with any credible evidence that such a bug exists. * * * * * In conjunction with work on the books, we're going to be producing many more special reports on various hardware and software products. The first of these is now up. It covers the Maxtor 91000D8 DiamondMax Plus 10 GB Ultra-ATA disk drive. There'll be a lot more to come, and the form and layout isn't necessarily final, so we'd appreciate any feedback you care to provide.
|
|
Friday |
Friday, February 12, 1999 Not much mail the last couple of days. Usually I get six or eight messages related to this web site even on a slow day, and perhaps 15 or 20 on an average day. I doubt I've had even half a dozen in the last two days combined. I guess I haven't said anything outrageous enough to prompt people to write. And speaking of prompting people to write, I need some help. I've been working with these Maxtor 91000D8 DiamondMax Plus 2500 10 GB disk drives, and I've run into a problem with running them in DMA mode. It's not the drives' fault, nor is it the motherboard (an EPoX EP-BXT). What I've run into is the 8 GB limit, but only when I attempt to use a bus master (DMA) driver. The BIOS supports extended INT13, and the system recognizes the drives as being 10 GB at boot time. The drives show their full capacity in Windows NT as long as I'm using the default IDE drivers. The first thing I tried to enable bus mastering was running the Dmacheck.exe program supplied with Windows NT 4. I enabled DMA detection and restarted the system. When I ran Dmacheck again, it told me that DMA was not in use on either channel. That's odd, because I know both the motherboard and the drives support UDMA. So I decided to install the NT4 DMA driver that was supplied on the CD-ROM with the motherboard. It's the Intel PIIX Bus Master IDE Controller driver, and it has no version number or other identifier. Just its name. I installed it, restarted the system again, and fired up Dmacheck. This time, it reported "No IDE/ATAPI devices detected" for both channels. Okay, that's odd, but the system was running fine and I could access both the hard drive on the primary and the CD-ROM drive on the secondary. I ran the ZD WinBench tests, and they showed that DMA was enabled and working. I wanted to test Windows NT Server 4.0 software mirroring. It wasn't until I installed a second drive--rejumpering the CD-ROM as Slave and adding the second hard drive as Master on the secondary channel--that I noticed something strange. When I fired up Disk Administrator, the display was screwed up. I'd partitioned the first drive (under the standard IDE driver) to a 1,028 MB primary partition formated NTFS as the C: volume and a 8,511 MB extended partition, all of which was allocated to a logical E: volume, also formatted NTFS. The total 9,539 MB represents the full 10 GB capacity of the Maxtor drive, because NT is using binary megabytes and Maxtor decimal megabytes. But when I installed the DMA driver and then the second hard disk, Windows NT reported both drives as having total space of 8,056 MB rather than 9,539 MB. The first drive was showing 8,056 total capacity, but its bar showed a 1,028 MB C: volume and an 8,511 MB E: volume, the bar for which was chopped off by the right border of the Disk Administrator window. Removing the DMA driver and re-installing the default IDE driver returns the display to showing both drives as 8,539 MB. So the problem is clearly the bus mastering driver, but I can't find a later version that will support drives larger than 8. I've been all over the Intel site, the Microsoft site, and the EPoX site. Does anyone know where I can find a Windows NT 4 bus mastering IDE driver that supports drives larger than 8 GB? I even tried installing the original Intel PIIX Bus Master IDE Controller driver and then reapplying SP4, hoping that Microsoft had an updated version. No joy. I'm afraid to try to use more than 8,056 MB under DMA because it may "wrap" or something and destroy data. If worse comes to worst, I could install the DMA driver and just use the drive as an 8,065 MB drive instead of a 9,539 MB drive. It'd be worth it for the dramatically increased performance that DMA yields. But I'd like to have both the speed and the full capacity. Any ideas? * * * * * And, as long as I'm asking for help... I use FrontPage 98 to maintain my web site, and I've run into a couple of issues I'd like to fix, but I can't figure out how to do it:
* * * * * And now I'm starting to get feedback from readers about the the Maxtor DiamondMax Plus review. I have a Reader Feedback section at the bottom of my review template, so the messages and my responses obviously go there. The question is, do they also go here. I think for now at least, I'll post them in both places. I want to avoid have a Dungeons and Dragons maze like Pournelle's. There's too much chance of people missing good stuff that way. Doing it this way means the same material ends up more than one place, but that's probably a small price to pay for the benefits. Some tells me I'm going to have to go to a separate mail page eventually. This from Robert Morgan [robert@morgan.pair.com]:
All those numbers look pretty reasonable, given your ad hoc environment and the fact that you're using an AMD K6-2/333 versus the Pentium II/450 used in my testing. And this follow-up from Robert Morgan [robert@morgan.pair.com], which he sent before he received my first response:
Yep, there's no doubt that the Pentium II/450 doesn't offer much bang for the buck relative to other processor choices. If you had a Slot 1 board, you could get most of the performance of that CPU by installing a Celeron for under a hundred US$, particularly if you get one of the overclockable ones. In Socket 7, the AMD K6-2 currently offers pretty good bang for the buck. You might want to wait a bit longer, though. The K6-III (Sharptooth) should be available shortly, and will be a better processor than the K6-2. * * * * * This from Alberto S. Lopez [Alberto_Lopez@notes.toyota.com]:
Thanks for the kind words. My first thought, as yours, would be that you had a defective memory module. Since the system behaves the same way with different memory, the next thing to check is the cables. First, obviously, make sure that all cables are connected to the right things and that they're oriented properly--pin 1 to pin 1. If all the cables are connected properly, the next thing I'd suspect is a bad cable. I recently had a system intermittently exhibit exactly this behavior. I eventually tracked the problem down to a bad floppy cable. It had been folded too tightly, although it didn't look crimped, and apparently at least one of the wires had broken. Moving the cable around while working on the PC would make or break contact, so the PC would sometimes boot and sometimes not. Replacing the floppy cable made the problem go away. Another possibility is that the motherboard may be defective or its CMOS settings wrong. Before you assume the motherboard is bad, make sure to try clearing CMOS to default. There's probably a motherboard jumper that you can set to clear CMOS. If not, you can remove the battery to clear CMOS. The last thing that immediately comes to mind is that you may have a failing power supply. If anyone else has any other ideas, they can mail you directly. Good luck.
|
|
Saturday |
Saturday, February 13, 1999 Thanks to everyone who sent suggestions about my FrontPage questions. I've included several representative ones below. Several of those I didn't publish pointed out that FrontPage is a poor tool, and that I could do everything I wanted by switching to HotMetal and hand-coding HTML, writing CGI scripts, and so on. I'm sure they're right, but the problem is that I don't know how to code HTML or write CGI scripts, and I don't have the time to learn. The rule around here is that if FrontPage won't do it, it ain't gonna get done. What I'm trying to do is focus on the results rather than the process. I don't have the time or the inclination to learn much about the process of creating web sites and pages. I just want an easy user-oriented tool that will let me do what I want to do. Perhaps FrontPage 2000 will be a step in that direction. Or perhaps I'm being unfair to FP98. I haven't really made any effort to learn to use it. There are still whole broad functional areas of FP98, Cascading Style Sheets for example, that I know are there but have no idea how to use. The point is, I want the program to do all the heavy lifting for me. If I have to install a gig of RAM or SMP to run that program, fine. But I want to concentrate on writing content rather than the mechanics of creating and maintaining pages and the site. I guess it'd make sense to get a good FrontPage book and read it, but I don't really have time to do even that. * * * * * This from Robert Morgan [robert@morgan.pair.com]:
I blush to admit that I hadn't thought about using the FrontPage global search and replace. So I went out and did a global search for the string "<meta name="Microsoft Border" content="b">" which should appear on every page. FrontPage searched all 146 of my HTML pages, but I got zero hits, so apparently the global search and replace works only for the body text, and not for HTML tags. Just to confirm that global search really worked, I just saved this page and re-did the search. Sure enough, I got one hit for this page. As far as PERL, that'd be a good suggestion, except that I don't know how to write PERL scripts and don't have the time to learn. You're probably right that I'll eventually find that protecting parts of my site is server specific, but there's no reason why it should be. The FrontPage Extensions already support a validation subroutine and password protected login, so I'm going to try wangling that into doing what I want. As Pournelle says frequently, it's only me, and I simply don't have time to do anything except things that my existing tools will allow me to do easily. If they won't do it, it's not going to get done. I wish I had time to learn a lot of new things, but I don't. You're not the only person who's suggested making the site database-driven. I haven't done so (and probably won't) for two reasons. First, the "new thing" problem. Second, and more important, I think I'm correct in believing that making a site database-driven eliminates having the content indexed by search engines. I want my material to be searchable on AltaVista, Northern Light, HotBot, etc., so I wouldn't want to do anything that isolates it from those search engines. * * * * * This from Rick Boatright [boatright@vocshop.com]:
Thanks. I may give that a try. I think I even have a copy of MultiEdit floating around here somewhere. * * * * * This from Bo Leuf:
I understand that HTML isn't Word, and that the formatting available with it is much less powerful. And I also understand that rendering is browser specific. I guess what I'd like to see is FrontPage doing the kind of things you mention, but doing so behind the scene. The last thing I want is to hand-edit HTML code. I don't particularly care exactly how my pages render, so long as FrontPage and the browser do "reasonable" things. I guess what I was objecting to was that FrontPage does offer the option for a user to do things that FrontPage itself can do. For example, for the following bulleted list, I entered the text "Bullet item 1", set the style to "Bulleted list", and then presssed Enter. Front Page inserted a new line with a bullet. It appears on screen as single spaced. I typed the second bullet item, pressed Enter, and typed the third. On my screen, at least, I end up with three bullet items with minimal line space.
For the following bulleted list, I typed "Bullet item 1" and pressed Enter twice, "Bullet item 2" followed by Enter twice, then "Bullet Item 3" followed by Enter twice. I then applied the "Bulleted list" style individually to each of the three lines. I then put the cursor on the "extra" line between items and pressed Delete. On my screen, at least, I end up with three bullet items with double-spaced lines.
Looking at the HTML source, the only difference is that each line of the first group is formatted as <li>text</li> while each member of the second is formatted <ul><li>text</li></ul>. Now, I don't know what <ul> or <li> is, but it'd be nice if I could choose the "effect" and have FP98 insert the appropriate codes, instead of it forcing me to do weird things or cut-and-paste from another document. * * * * * This followup from Robert Morgan [robert@morgan.pair.com]:
Thanks for making that clear, and for understanding my position. I used to have a boss who was fond of saying, "we can do anything, but we can't do everything" and that's the case here. If I had the time and desire to learn about HTML, scripting, etc. I'm sure I could do it, but I work pretty much seven days a week as it is. So I have to prioritize, and (unfortunately, some will say) learning more about these things is just too low on my priority list to ever get done. I appreciate you explaining the dynamic versus static issue, and I promise I'll never club you with that dead fish ever again. I knew about the Robots.txt issue, but I hadn't thought through the idea of "static" dynamically-generated pages. Now if only the FrontPage client and the FrontPage Server extensions would do that for me automatically... And I do intend to keep focusing on content. Much as I enjoy playing around with the mechanics of creating, maintaining, and publishing pages (and I do enjoy it), I enjoy the process of researching things and creating new content much more. * * * * * And regarding my two list versions, Bo Leuf [bo@leuf.com] says:
Thanks for explaining how the lists work. As I've said, I know nothing about HTML. I just want to use it indirectly via FP to generate pages that kind of look like what I want them to.... * * * * * While working with the Maxtor DiamondMax Plus 2500 drives, I wondered what kind of performance I'd get running two of them using Windows NT Server 4.0 native software-based RAID1 mirroring. So I ran the tests. The results are now up as Windows NT Server Disk Mirroring Performance. * * * * * This from Mendel Leo Cooper [thegrendel@theriver.com]:
Thanks for the kind words. As it happens, I'm running Netscape under Windows NT, but I can accomplish pretty much the same thing you're describing simply by making the cookie file read-only. * * * * * Late Afternoon: I've spent most of today working with the Promise FASTtrak IDE RAID Controller. I got started on the wrong foot by not reading the manual first, and that cost me a couple of hours. But I went back to square one, and am still having problems getting the the system to boot from the FASTtrak mirror set. I may simply be overlooking something simple, so I'm going to forward my current results to Promise for their comments before I publish something that may be unfair to them. Presumably, I'll not receive a response until at least Monday, so I'll put that review on hold until then. |
|
Sunday |
Sunday, February 14, 1999 I'm not feeling too great this morning. It's probably due to this typical North Carolina weather. Friday afternoon, the temperature reached a high of about 75F (24C). Yesterday evening, the wind chill was at 6F (-14C), with snow predicted. The snow didn't show up, but the temperature this morning was 25F (-4C) with wind chill near 0F (-18C). That's North Carolina. Swimming weather one day, and ice skating weather the next, literally. * * * * * There's a lot of talk over on Jerry Pournelle's site about changing page names. Jerry maintains weekly View and Mail pages, which he updates daily. People are asking him to keep the file names constant for the current View and Mail pages so that they can bookmark them. I'd never thought of that. Duh. I feel particularly bad because I'm the one who suggested Jerry's current scheme. He has a "View Home" page that has a link to the Current View page, e.g. View35.html. Same thing for Mail. So, although readers can bookmark the View and Mail home pages, they can't bookmark the current View and Mail pages. Well, they can, but then they have to change their bookmarks every week. That problem is easy enough to solve. All I need to do on my own site is keep creating weekly documents as I have been doing. Each time I update the current Daynotes file, I'll simply also save it as thisweek.html. I'll have two identical copies on the disk and the web server, but who cares? Then, when the week changes, I'll simply copy the current week's Daynotes page to lastweek.html. So, if you want to bookmark this page, just set bookmarks for thisweek.html and lastweek.html and you'll never have to come through the home page again. * * * * * Last evening after dinner, Barbara and I decided to pull the two unused hard disks out of sherlock. Sherlock boots NT from a 6.4 GB IBM ATA drive, but also had two Western Digital 4.3 GB IDE drives installed. I put them there back last summer when I was writing the MCSE training courses for DigitalThink, because I needed a three-disk machine for screen shots of Windows NT Server mirroring, striping, and striping with parity. Since I finished writing that course, they've just been sitting there not doing much of anything. What motivated me to pull them was actually two things. First, I needed a decent boot drive for the test-bed system, and a 4.3 GB IDE seemed ideal. Second, there was so much dust in sherlock that it was literally coming out around the edges of the drive bezels. The final straw was when I ejected the CD in the drive and watched dust dinosaurs falling off the bottom of the CD tray. So, we decided now was as good a time as any to pop the lid, pull the drives, and clean things up. Sherlock is a Dell XPSM200s mini-tower box. Incidentally, did you know that "tower" was once claimed as a trademark? I believe it was NCR. Whoever it was produced a machine they called "The Tower." This has to be about fifteen years ago. Back then, if you wanted a vertical PC, you stood your desktop on its side. But then someone started making vertical cases, and PC manufacturers started advertising "tower" and "mini-tower" systems. Whoever it was that claimed the trademark tried to stop PC vendors from using the term. I'm not sure if the company couldn't enforce its trademark, if it was deemed invalid, or what, but obviously everyone now uses the term generically. Reminds me of the company that claims a trade-mark on the term "ni-cad" and periodically sends letters to magazines to tell them that they're misusing a trademark as a generic term. I'm pretty sure my 1910 edition of Vinal's book on storage batteries uses that term, but I may be misremembering. I can tell I'm tired when I start rambling like that. I installed NT at least a dozen times yesterday, tore cases down, replaced drives repeatedly, etc. etc. That from about 9:00 Saturday morning through 8:00 Saturday night. I think I'll take the rest of this week off. * * * * * This from David Karlins [dkarlins@aol.com], a FrontPage expert and a fellow member of the Computer Book Publishing mailing list:
Several people have suggested using an editor that does multi-file search and replace, which seems a reasonable solution.
That'll do it. Thanks. You're the first person to mention that. I couldn't believe I hadn't seen that option until I went and looked. It's greyed out on my local copy of the web. I loaded another root web that happens to be local, and FP fired up its local stub server. Sure enough, with that web loaded and the stub server running, I was able to assign permissions.
That does it, too, thanks. I had actually looked at that, but avoided using it because I was under the impression that it meant FP would be using style sheets, which I had the impression were a Bad Thing. Is that true?
Will do. Thanks. * * * * * And this from Clas Kristiansson [Attle@mbox300.swipnet.se], which I received as an email message with a Word attachment. Mr. Kristiansson's letter follows the introductory paragraph. Despite his disclaimer, he writes far better English than do many I know who grew up speaking that language:
Dear Sir Through the pages of Bo Leuf, I came in contact with your site and became aware of the various problems that you reported on Friday, February 12. Though English is not my native language, I hope that I can shed some light upon your problems and that if my "abuse of language" makes it totally incomprehensible, I hope you will write back and ask me to explain further.
If you are curious how to layout on the web well Dreamweaver has a trial version. (at www.macromedia.com). It is not in everyones taste but it suits my purposes of doing readable layout. (Anyone can do fancy layout – I want the layout to improve the readability). Just a quote (in translation) from one of my favorite books upon Typography (Bokstaven, ordet, texten by Christer Hellmark) "The typographic constrains of Internet is not dictated by the care of the reader. Only a technology nerd could be satisfied with seven different sizes - not more, that the text reaches across the full screen and that it is set with an undefined typeface, not controlled a designer or someone else with knowledge in readability, but by the receiver himself." Tough words, but IMHO quite true. Hope this was of some help Best wishes Clas Kristiansson Thanks for taking the time to respond in detail, and, yes, it was helpful. When it comes to web page layout, I'm a novice. Worse still, I'm a novice who does not have the time or the inclination to learn the intricacies of the process. When I started this web site, I decided to focus first on content, second on readability, and third on fast download speed. The first means I don't have much time to spend worrying about much else. The second means I decided to stick pretty much with standard fonts (99% of what I write uses the "(default font)"). The third means I avoid graphics and fancy formatting as much as possible. I'm not sure that I agree with the quote you provided. Without minimizing the importance of typography and layout to readability, that quote sounds rather arrogant. Seven sizes should be more than enough for anyone, particularly because those sizes are relative rather than fixed. Regarding line widths, any web page designer who cares to can certainly produce pages in multi-column formats, or indeed simply limit the width of the text column, as Bo Leuf does. As far as the undefined typeface issue, surely the reader is ultimately the best judge of readability from his own perspective? So perhaps I am a "technology nerd", although as it happens I do have some experience in typesetting going back to the true Linotype days. But I agree that HTML is a poor tool, and you may well be right that FrontPage is also a poor tool. I use FrontPage not because I like it but because it is what I know. When I started working on my web site, I asked the opinion of several people who were much more experienced than I. Without exception, they suggested that I use FrontPage. Many of them did not use FrontPage themselves, so what they were suggesting was what they perceived as the best tool for me given my own circumstances, rather than the "best" tool in any absolute sense. But for all its deficiencies, FrontPage does allow me to get pages created and content published to my web site without spending much time thinking about the underlying technologies.
|