|
Week of 1/25/99 Friday, July 05, 2002 08:12 A (mostly) daily journal of the trials, tribulations, and random observations of Robert Bruce Thompson, a writer of computer books. |
|
Monday |
Monday, January 25, 1999 No book of the week this week. I've gotten a few mail messages recently that ask why I don't publish the email addresses of people who send me mail. I'd not been doing it in the past simply as a privacy measure. However, I notice that others who publish reader mail publish the addresses of the people whose messages are posted. As a couple of people have pointed out to me, it comes in handy to have the email address if you want to respond to that person privately. So, I'm going to start publishing the addresses with the messages. If you don't want your email address published with your message, please tell me so at the top of the message. I'm changing policy with no notice, so if anyone is upset that his mail address appears, please let me know. I'll go back and remove it immediately. * * * * * The following from Tom Bledsaw [tbledsaw@worldnet.att.net]
Thanks for the kind words. I agree with you that USB is not quite ready for prime time, although I think that's changing fast. The problem with the USB keyboard not being recognized occurs because you are supporting USB via Win98 drivers. Depending on your BIOS, you may be able to avoid this problem. Some BIOSs have a "Enable Boot-time USB Keyboard Support" option. If your BIOS has that option and you enable it, the BIOS recognizes USB keyboards natively. As far as USB sound, I've not experimented with it yet, but your experiences are similar to those of others I've talked to. Again, not quite here yet... * * * * * And the following on CD-R from Tom Syroid [tsyroid@home.com]:
All good points, and I'm still thinking about it. I'll compare prices for ATAPI and SCSI CD-R drives. If the SCSI is within reason, I may go for it. I won't hold the cost of the SCSI card that I'll also need to buy against the CD-R drive, although I probably should. I don't think I'll have any problem deciding about the interface of my next hard drive, though. It'll be ATA no matter what kind of CD-R drive I get. * * * * * This from Chuck Waggoner [waggoner@gis.net]:
That is interesting. Conventional wisdom says that Linux is a developing threat to Windows NT in server space, but does not threaten Microsoft's hegemony in client space. I'm not sure the latter is true, particularly for the many clients that are not used for general purpose office productivity applications. It'll be interesting to see what happens in the next year or so. Around then, Microsoft will be releasing W2K, and Linux should have evolved quite a bit towards being a reasonable client platform. OSS is the biggest threat that Microsoft has ever faced. It may make their troubles with the Justice Department pale into insignificance. * * * * * This from Jerry Mah [jerry-mah@home.com]:
Thanks. Actually, there are some common threads developing in the CD-R discussion: (a) anything recent seems to work, (b) SCSI is better than IDE, but IDE works if you're careful, (c) use the Adaptec software, and (d) all blanks are not created equal. And you're right that SCSI is still widely used. My point was not so much that SCSI is dead as that it's a dead-end technology for clients. * * * * * And now I'd better get to work. I have bills to pay, a chapter to send off to O'Reilly, another chapter to get started, and my handyman friend showing up to replace the ballasts in the fluorescents in the basement kitchen. * * * * * On Friday, Gary M. Berg mailed me to ask if there were problems with the Diskeeper defragger and Windows NT SP4. I installed Diskeeper Lite on the only machine I have running SP4, and, sure enough, it wouldn't work. I mailed my primary contact at Executive Software, and he asked me to submit a detailed report of the situation to one of their tech support folks. I sent that late on Friday, and received the following response this morning from Lance Jensen of Executive Software Tech Support:
So, I did as Lance suggested, and reinstalled DKLite. It's defragging the drive as I write this. I'm not sure what happened. When I installed DKLite the first time, I was logged on (as usual) to my main account, which has full administrative privileges. That's a horrible practice, and one I caution others to avoid, but truth in journalism means I have to admit to doing it myself. I started to re-install DKLite without removing the original installation. DKLite Setup blew up with a file copy error, saying that the program Control.exe could't be deleted. I stopped the install and used NT Explorer to delete the entire D:\Program Files\DKLite folder, where I'd originally installed the program. Explorer refused to delete the folder, giving me an access denied error. I went in and found Control.exe was the file causing the problem. I took ownership of it and tried to delete it again. No joy. I finally fired up Task Manager. Sure enough, Control.exe was on the processes list. I killed that process and was then able to delete Control.exe and the folder itself. I then re-installed DKLite, and it is now defragging the drive normally. It seems that it was just a coincidence that I had this problem on an SP4 machine immediately after getting mail from Gary that asked if Diskeeper had problems with SP4. At any rate, there is no problem with Diskeeper Lite and NT SP4. Just to make sure, I just installed and ran Diskeeper 4.0 (the full commercial version) on bastet. It also works fine. * * * * * Late Afternoon: I've spent the day working on the new chapter, and have gotten a fair amount done. * * * * * This from Randy Moreau [rmoreau@4dnet.com]:
I've been using modems heavily since the early 1980's, and I use external modems exclusively. I've had a few internals over the years, but I've always regretted buying them. In theory, an internal modem may be a bit faster (because it has a direct bus connection rather than being limited to the 115.2 kbps of a standard serial port) but in practice there's no discernable difference in speed. The controller/controllerless issue arises because some internal modems are designed to save a few bucks by using the PC processor rather than having a dedicated processor of their own. Steer clear of these, which are usually described as a "Winmodem" or something similar. The externals cost a bit more than internals because of the enclosure and power brick, but the difference is worth it. The lights on the external are useful. Even more important is that you can reset the power to an external modem if it locks up rather than having to restart the PC. I'd recommend that you buy a U.S. Robotics V.90 external modem. They're available in various models, including ones with Fax or Fax/Voice capabilities. USRs make and hold a connection when others can't. I've been using them for twenty years now, and have never gotten a bad one. Back in the days when Hayes was the standard, USR got their foot in the door by focusing on BBS sysops, who had a true appreciation for a good modem. Nowadays, Hayes is bankrupt and USR has been the standard for many years. USR used to have two main models. The Courier was their high-end modem. It was quite expensive, but favored by BBS sysops and corporations for its rock-solid dependability and compatibility with other modems. The Sportster was their lower-end, "consumer grade" modem, but it still beat the pants off competing models. As of late, it appears that USR may be eliminating the Courier/Sportster differentiation, although I noticed on the USR web site just now that they still sell the Courier V.Everything as a corporate solution. * * * * * I finally got around to downloading Opera, the web browser that isn't free. I've spent half an hour or so playing with it, and I begin to see a glimmer of why its fans are so enthusiastic. As one example, when I installed it, I completely forgot that I was using a proxy server and so I just fired it up and entered a URL. Opera displayed the URL normally. It wasn't until later when I was playing around in Preferences that I noticed that Opera had configured itself automatically to use the proxy server. I'd guess it went out and looked at either the IE or Navigator configurations (or both) to do this. No matter how it did it, this shows that Opera is concerned with usability. If it hadn't automatically configured itself for the proxy server, that would have only slowed me down for a few seconds. But a novice user might have been stymied, and it's nice to see that Opera cares enough to try to avoid common problems like that. * * * * * Early Evening: I didn't get the stuff from this afternoon published, because I was having problems with the Internet. I first noticed them trying to POP my mail. Then I noticed that web sites were coming up very slowly, if at all, in Opera. I thought maybe there was a problem with my connection, so I dropped it and dialed in again. Same slowness. So I tried dialing another ISP account that uses a completely different route to the Internet. Same deal. I'm beginning to think that the Internet as a whole is running very slowly today. I'll try publishing this in a few minutes, although something tells me that FrontPage will time out on the transfer. * * * * * And Chuck Waggoner [waggoner@gis.net] has this to say about the problem I am having with Internet Explorer links spontaneously changing between Small Icon mode and Large Icon mode:
Are you certain that it's always the first one on the left? In my case, whichever one I open first during a session displays with small icons for the rest of that session, but any one I open subsequently displays with large icons. It doesn't matter which one I open first. It's always the first one opened that uses small icons forever after during that session. Exiting and then re-opening IE (or opening another instance) resets it so that again the first folder I open uses small icons. I'd love to find out the answer as well.
|
|
Tuesday |
Tuesday, January 26, 1999 I was reading Bo Leuf's site the other day when I came upon his Links page. One of those links was to Edmund Scientific. I remember buying things from Edmund Scientific when I was a teenager thirty years ago. Back then, they carried everything you needed--from mirror blanks and grit in bulk and pitch to tubes and spiders to equatorial mount components--to build your own reflector telescope. Anything from 3" to 12" as I recall. Geez, they even sold the domes you needed to build your own small-scale observatory. They also sold lab equipment in small quantities for home users. A kid could (and I did) equip his own basement chemistry lab from the offerings of Edmund, everything from beakers, flasks and graduated cylinders to reasonably priced microscopes and centrifuges. And just about anything else a kid with a scientific bent might need. Lenses, radio components, power supplies, test equipment, etc. Some new, some used, much of it military and industrial surplus. Nowadays, the stuff they sell is all commercialized, much like what I'd expect to find at a toy store. It makes me wonder where our next generation of scientists and engineers is going to come from. When I was a kid, we did things. We built ham radio equipment from old salvaged tube radios and TVs. We worked on automobile engines and experimented with our chemistry sets. We built our own telescopes, from grinding the mirror to assembling the equatorial mount. We built darkrooms and learned about photography (and some chemistry along the way). We built and launched rockets and built and flew RC airplanes from individual relays, receivers, and other components. Nowadays, everything is packaged. Most hams buy stuff off the shelf instead of building their own. You can't work on car engines anymore unless you're willing to risk being busted by the air quality police. Today's chemistry sets are a joke. The manufacturers are so afraid of being sued that their so-called chemistry sets are pathetic. I found only a couple of places on the entire Internet that still sell mirror blanks, grit, and the other stuff needed to build a simple Newtonian reflector. Doing darkroom work is a dying hobby. Rocketry clubs probably can't afford the liability insurance nowadays, and RC equipment probably has to be FCC type-accepted. When the first electronic calculators came out, I remember teachers objecting to them on the basis that people would become dependent on them rather than learning math. In retrospect, I think they were right. How many kids today can extract a square root manually or factor a quadratic equation? So technology does have a down side. That, the huge increase in litigation, and the general dumbing-down of society have all conspired to make it tough for kids with a scientific bent to do what comes naturally. They're probably off playing with their computers. Computers are great, but they aren't everything. * * * * * This from Tom Syroid [tsyroid@home.com] who is considering bringing up a local web server on his cable modem connection:
DNS is a network service that resolves host names into the
corresponding IP addresses. For example, when you enter www.ttgnet.com
into your browser, the DNS resolver stub in the browser queries the local
DNS Server (in your case, one located at Shaw) to determine that the IP
address that maps to www.ttgnet.com is 165.90.46.122. Grossly
oversimplified, the DNS server at Shaw queries the .com DNS server to
determine the address of the ttgnet.com DNS server. Once it finds that
address, it queries the ttgnet.com DNS server to find the IP address of
the host named "www". In order for this to work, there has to be
a "home" DNS server for each domain name. Actually, two DNS
servers are required by InterNIC to make sure that if one is down the
other can still respond to DNS queries. * * * * * This from Shawn Wallbridge [swallbridge@home.com]:
Yes, the consensus seems to be that IDE works, but SCSI works better. I'm running a Seattle-based homebrew with a Pentium II/300 (soon to be 450) and 128 MB of RAM (soon to be 256 MB) under Windows NT, which should be enough system to burn CDs reliably even with an ATAPI burner. I'm sorry to hear that your government caved in to the music industry, but I'm not surprised. I suspect they spend a lot of money on lobbyists, a lesson Bill Gates may yet learn. Most of my sources here sell name brand blanks for anything from $1 to $2 each, and it's easy to find rebate deals and specials that drop that well under $1. How long can it be before we end up paying a music tax here, too? As far as Exchange and SMTP queueing, I ran Exchange briefly and remember vaguely dealing with SMTP queueing, but not enough to do you any good. You might talk to my friend and future co-author Paul Robichaux. He recently wrote an Exchange 5.5 book. I don't know how much he knows about SMTP queueing, but he might be able to at least point you in the right direction. * * * * * And a follow-up from Tom Syroid [mailto:tsyroid@home.com]:
Ultimately, the IANA (Internet Assigned Numbers Authority)
is responsible for assigning IP network addresses. Or at least they used
to be. There have been some changes in that area recently that I haven't
been tracking too closely. Anyone can apply for a network number. For
example, Triad Technology Group applied to InterNIC for a Class C network
number, and was granted 204.238.30. That means that I own the range of IP
addresses from 204.238.30.1 through 204.238.30.254, which I can assign to
hosts as I please. However, in order for this to be meaningful, I have to
have a full-time Internet connection, which I no longer have (although you
do). But, the upstream provider has to be willing to route packets to and
from your network address, which @Home most definitely will not. * * * * * This from Chuck Waggoner [waggoner@gis.net] about the Internet Explorer problem:
Okay. I guess it's official. This is a real bug, one that's reproducible by different people running different environments. I wonder if IE5 still has the problem. * * * * * This from Jerry Mah [jerry-mah@home.com]:
Thanks. I'll keep the software you mention in mind. * * * * * This from Tom Syroid [tsyroid@home.com]:
Thanks. I went over to look at it briefly, but I couldn't figure out how to do a real search. There's a text box there that you can enter your search string into, but it seems to limit you to the width of the box, which isn't wide enough to put a useful string into. When I tried a couple of simple searches within the width limits of the box, e.g. "Internet Explorer", it returned no hits. Perhaps I couldn't make it work properly because I was using Navigator rather than IE at the time. If it works, this is indeed a very nice resource. * * * * * The following from Frank J. Albert [frankalbert@coastalnet.com]:
No, I think I'll hold onto my old Pentax, but thanks for the offer. If you're interested in buying an old, fully manual camera, try one of the New York camera stores like B&H or 47th Street Photo. Many of them sell used equipment, including some pretty old stuff like the original Honeywell Pentax Spotmatic, the Nikon F, and so on. You might be unpleasantly surpised by the prices, however. These things have now achieved classic/collector status, and most are priced accordingly. Another thing you might consider is new equipment. I haven't followed the photography scene for years now, but it used to be that many Chinese manufacturers (like Seagull) made fully manual knock-offs of classic manual cameras like the Rolleiflex. Many of them were of decent quality, if not up to the fit and finish of the German and Japanese manufacturers. I suspect that such cameras are still widely available. If so, a quick look at the ad pages of one of the popular photography magazines should turn up quite a few candidates. * * * * * Barbara was in the midst of cooking her lunch when the microwave oven died as it was baking her potato. And I mean dead. Even the clock and internal light wouldn't come on. I went downstairs to check the breakers, and found that none had tripped. I went back upstairs to plug the can opener into the same receptacle. It ran fine. That didn't look good. I slid the oven out of cabinet so I could get to the back panel, hoping for a breaker on the back, or at least a fuse. The back panel was bare. I called my appliance repair guy. He's honest, a pearl beyond price. When it's not worth fixing something, he says so. He told me that there is an internal fuze or breaker, but if it blew it probably did so for a good reason. He doesn't work on microwaves and doesn't recommend repairing them once they're out of warranty. He says that those who do repair microwaves typically charge a $50 minimum and that you can buy a decent 1000 watt unit at Wal-Mart for $125. At a guess, this oven is at least four or five years old, so it's probably not worth repairing. Barbara and I have established a new tradition. Wednesday afternoons we visit the library and then go out to dinner. I guess I'll check Consumer Reports for microwave recommendations. I don't think much of Consumer Reports. Their Frequency of Repair statistics are about the only thing I pay any attention to. So then I guess we'll head for Wal-Mart and buy a new microwave.
|
|
Wednesday |
Wednesday, January 27, 1999 Barbara tells me that the microwave is closer to ten years old than five. And, as usual, an incident like this emphasizes the difference in the way that men and women remember things. Like most guys, I think to myself, "well we've had this for a while. Longer than a couple of years? Yeah, probably. As much as ten years? No, probably not. I guess it's about 5 years old." Barbara, on the other hand, like most women, goes about it entirely differently, putting everything into context with other events in our lives. Something like, "we moved into this house in October 1987. My parents gave us the microwave as a Christmas present, but I don't think it was the first year that we were living here. As a first approximation, then, it might have been Christmas 1988. Alison was here for Christmas dinner the year we got it. She moved to Mississippi in 1992 or 1993, and there were two or three years before she moved that she didn't come here for Christmas dinner. She must have moved in 1992, because she'd already moved to Mississippi when I moved from the Reynolda Manor branch library to the Rural Hall branch in ..." How does she remember that Alison was here for Christmas dinner the year we got the thing? How does she remember what month and year we moved here, come to that? Or what year Alison moved to Mississippi? I was proud of myself for remembering that her parents gave it to us and that Alison had moved to Mississippi. And, being a good and wise husband, I showed the preceding to Barbara before I posted it. She laughed, and explained that the reason she remembers that Alison was here is that Alison gave her microwave dishes. Okay, now the only thing I don't understand is how she remembers that Alison gave her microwave dishes. Of course, my father once literally drove away from a highway rest stop leaving my mother in the restroom. He got quite a few miles down the road before he remembered to turn around and pick her up. That's a hard one to explain, let alone live down. Actually, it was worse than I thought. Just now, I shouted downstairs to ask mom how far dad had gotten before he remembered he'd left her. She responded, "Which time?" Which time? Apparently, he did it again a few years later, driving away to pick me up at high school and leaving my mother at the grocery store, standing there with her arms filled with groceries. I wonder what she was thinking as she watched him retrieve the car from the parking lot and then cruise past her oblivious. But now that I've been reminded of that incident, I do remember the look on his face when I asked, "where's mom?" I know what must have happened. Sometimes I get to concentrating so hard on something that I literally forget to eat. So perhaps it's genetic. Either that, or men's and women's memories operate the same way, but we each of us only remember things that are important to us at the time, and focusing our attention tightly on something can make us forget other things... * * * * * This from Chuck Waggoner [waggoner@gis.net]:
Excellent points, and I must admit that I hadn't really thought that aspect through. But I don't use a safe deposit box to store backups and similar material, and for a much more mundane reason. I need to be able to get to them whenever I want, not just when the bank happens to be open. * * * * * More on the issue of using Diskeeper 4 with Windows NT Service Pack 4, this from Gary M. Berg [Gary_Berg@ibm.net]:
* * * * * And more from Gary M. Berg [Gary_Berg@ibm.net], this time about manual cameras:
Thanks. I'm sure that'll be of help to Frank. I haven't done much more than shoot family snapshots for many years, so I've lost track of the business. Auto-anything is pretty foreign to me. I got started in serious photography more than 30 years ago with a 11X14 Ansco cherrywood field camera (with Red Dot Artar lens) that my school was about to discard! So anything that doesn't use a focusing cloth is the ultimate in automation for me. The first time I used a 35mm SLR (a Nikon F), it took me quite a while to determine that I couldn't figure out the camera movements because there weren't any. * * * * * And a followup from Bo Leuf [bleuf@algonet.se] about Edmund Scientific:
Exactly. Looking back, I can see that some of the stuff we did back then would have given our parents heart attacks if they'd known about it. Hell, it almost gives me a heart attack thinking about it now. We made nitroglycerine in the basement, and not in tiny quantities, either. Although we never completed it, we were far along toward building a small rocket engine that we planned to fuel with 30% hydrogen peroxide and methanol/hydrazine (what the Germans called T-Stoff and C-Stoff, respectively, and used to fuel their ME-163 Komet rocket fighter, which frequently blew up catastrophically). We messed around with obsolete photographic processes that used nasty stuff like potassium cyanide. We made daguerreotypes, which required fuming a silver-plated copper sheet with mercury. We did any number of things that could easily have had fatal results. And yet, we're all still here, and we learned a lot. I guess you could say that we had no sense, but at least we were careful. * * * * * More from Gary M. Berg [Gary_Berg@ibm.net] about photography:
I actually got started in photography when I was 11. I'm not sure why I decided I wanted to learn about photography, but I headed for the library to find some books about it. At that time, the ground floor of the library was the adult section, and the children's area was the second floor. I headed for the stacks in the children's area and found a group of books, all by Ansel Adams. The series name was Basic Photography, with individual books on The Camera, The Negative, The Print, etc. They were thin gray books, and I remember thinking they were small enough that they should be pretty easy to get through. When I got home and started reading them, I was glad I'd picked up the series on Basic Photography instead of holding out for one on Advanced Photography. So, my start in photography was learning about camera movements, the Zone System, compounding one's own developers, and so on. I didn't actually start with the large-format stuff. That came later, when I was about 14. I started with a collection of box and folding cameras that used roll film and were mostly 2.25" square or 2.25 X 3.25". Shortly after that, someone gave me an old 4X5, which was my intro to sheet film. I never used an SLR until I got to high school. But you're right about automation. Barbara and I have a couple of point-and-shoot auto-everything 35's around. But we mostly use our old manual SLRs. And I do miss autofocus sometimes when I'm using them, particularly when I'm using Barbara's main camera, which she usually leaves the plain matte focusing screen in. No microprisms, no rangefinder, just plain ground glass. Ugh. Somehow, it was different looking at a big piece of ground glass under a focusing cloth.
|
|
Thursday |
Thursday, January 28, 1999 Barbara had some errands to run yesterday, so she stopped by Best Buy to check microwave prices. I was surprised how low they were. Name brand 1.5 to 1.9 cubic foot units with a kilowatt or more, large carousels, and all the bells and whistles were in the $130 to $175 range. Barbara and I stopped by the library yesterday afternoon to check Consumer Reports for a microwave review. They did one in the December, 1998 issue, so I read what they had to say. No much worth reading, unfortunately. I don't think much of Consumer Reports. Their liberal/left/environmentalist leanings color everything they write about. When they've reviewed things that I know a lot about, I've often disagreed strongly with what they've had to say. So, I see no reason to think their advice is any better when they report about things I know little about. In the past, the only thing I've considered worthwhile is their Frequency of Repair ratings, but I'm starting to wonder about those. The Frequency of Repair ratings chart for microwaves listed eight or ten manufacturers. The most reliable was Sanyo, at 4%. All of the others were between 4% and 8%, except Quasar, at 10%. The Panasonic was next to last, at 8%. Our old microwave, which we used daily for ten years, was a Panasonic. The chart noted that differences of four percentage points or less were not significant, so from the information given it was impossible to judge whether the 8% Panasonic was more or less reliable than the 4% Sanyo. About the only useful information that chart provided was that the Quasar was less reliable than any other model they rated. But I wonder if the chart provided any useful information at all. First, what does "repair" mean? Who repairs a broken microwave, anyway? The things cost anything from $89 to maybe $250, and I've been told that a typical repair might cost anything from $100 to $200. Who is going to pay that much to repair a broken unit when new units are available so cheaply? So what is the real size of the universe they used in evalutating frequency of repair? Then, as they do point out if not in much detail, they haven't normalized the data. They send out a survey and simply accept whatever responses come back. That's not a valid sampling method. And they are simply grouping all models from a particular manufacturer and assigning one value for reliability to that manufacturer. Again, not a valid method. So it seems that, although they are providing what amounts to casual anecdotal data, they present it in a format that makes it appear to have been collected and analyzed in a scientifically and statistically rigorous manner. So I decided the heck with it. Barbara and I left the library and headed for a restaurant. After dinner, we cruised out to Wal-Mart and checked their offerings. They had a Panasonic 1.9 cubic foot kilowatt unit on sale for $130. It looked to be about the same size as our old Panasonic unit, and had similar features. We figured we'd be better off with another Panasonic so that we wouldn't have to re-learn all the functions. We bought it, carried it home, unpacked it, plugged it in, and it works fine. Then I started thinking about the Consumer Reports Frequency of Repair data. If they really want to provide useful information, they could do so simply by asking their readers what they thought of their microwaves, without making any attempt to put a pseudo-scientific gloss on the resulting data. Simply asking "Are you satisfied with the reliability of your Panasonic microwave?" would generate useful data. Although it's not a well-known phenomenon, the central tendency of collective judgement can give amazingly good results. If you ask a large group of people to guess about something about which they know nothing, the average of the responses is often surprisingly close to the real figure. For example, if I picked a country at random, say Ivory Coast, and asked my readers to send in their best guesses of its surface area, the individual responses would vary greatly, but the average of those responses would likely be very close to the correct figure. As a matter of fact, let's do it. Email me your best guess of how many square miles (or square kilometers if you think in metric--I'll do the conversions) the boundaries of Ivory Coast encompass. Please don't look it up. Just give me your best guess off the top of your head (and tell me whether your guess is in square miles or square kilometers, because that won't be obvious from context). And don't worry about looking foolish. I won't publish anyone's individual guess. I'll simply post the results as the highest guess, lowest guess, the arithmetic mean of the guesses, and the actual figure. I'll let this run for several days to accumulate responses and then publish the findings, results be damned. And if you ask people to give their opinions about something they do know something about the results are better still. That's why PC Magazine's surveys provide good information. They ask how likely a responder is to buy the same product again. That's probably the most valuable piece of information they gather, because it provides a gestalt of how people feel about the equipment being surveyed. If you ask people whether they consider their Panasonic microwaves to be reliable, the results will vary all over the map, both because individual units vary in reliability and because people's expectations differ greatly. Some people will have had one fail after only a couple of years, but will rate it reliable because it worked well for them and they figure they got their money's worth out of it. Others will have one fail after fifteen year's use, and rate it unreliable simply because it died. So there will be "unfair" responses on both ends of the spectrum because of unrealistic expectations, but those responses will tend to cancel each other, and the central tendency will provide a reasonably good number. * * * * * The following is from Bo Leuf [bleuf@algonet.se], who obviously has forgotten more about memory than I've ever learned:
* * * * * And now I need to get back to work on the chapter. * * * * * Hmm. My web server and POP server, both located at bigbiz.com, went down just as I was trying to update my web page. Apparently, some files were updated and others weren't. Both servers are working again, so I'll repost this. * * * * * Well, that worked, but the publishing process took 15 or 20 minutes rather than the minute or two that it usually takes. Apparently, my server is overloaded at the moment. At least I got a "published" prompt at the end, so it was indeed completely published. In the past, I've sometimes gotten a time-out error during the "processing web updates" phase. Each time that's happened, I checked the web site and found that the publishing had in fact succeeded, at least in the sense that all updated pages were available on the server. If our cable and/or telephone company ever get around to offering cable modem and/or ADSL service with a static IP address, I think I'll start running a local web server, mail server, DNS server, etc. It'd be nice to have direct control over this whole process. * * * * * Mid-Afternoon: I'm still having problems with my mail and web servers, but have gotten several responses to my Ivory Coast question already, including the following from Lee Mandell [Lee_Mandell@iceinc.com]:
Thanks. As it turns out, the actual value is <removed> square miles or <removed> square kilometers. As it happens, my own guess was also <removed> square miles, or <removed> square kilometers. It'll be interesting to see how the results turn out. I'm not going to process the numbers as I receive them, just to make sure I don't unconsciously skew the results. I'll wait a few days, or until I've gotten 50 or so responses, or until the response rate slows to a trickle, whichever comes first, and then I'll plug the numbers into Excel and see how they turn out. As far as mixing objective and subjective data, you're correct in one sense. But the point we're trying to reach is an answer to the semi-quantifiable question, "does Panasonic make reliable microwave ovens?" Obviously, words like "reliable" constitute value judgements, but the cumulative responses to such a question can be considered as a pseudo-objective data point in the sense that responses from a reasonably large statisical universe represent a consensus of what "reliable" means. If you ask one person's opinion of the movies Bringing Up Baby and Plan 9 from Outer Space, you may receive an anomalous response. That particular person may simply dislike Cary Grant or screwball comedies, and think that the first movie was a poor one because of that prejudice. Conversely, that person may love campy 50's sci-fi movies and rate Plan 9 highly on that basis. But if you get responses from 100 or 1,000 people, I'll bet you'll find that the consensus is that Bringing Up Baby is one of the best movies ever made, and that Plan 9 from Outer Space is one of the worst. Although a movie can't be good or bad in any true objective sense unless you rule out personal preferences entirely, the consensus provides a reasonable yardstick when you're deciding which movie to rent.
|
|
Friday |
Friday, January 29, 1999 My web hosting service, Bigbiz.com, is about to drive me nuts. I didn't get a thing done yesterday, between problems with my mail and web servers and dealing with the Blue Cross/Blue Shield bureaucracy. At this point, connectivity with BigBiz is so bad that I probably won't be able to update my web site successfully. I'm going to try, just so that I'll at least have a notice posted about the problems. If you try to hit my web site and can't reach it, please try again later. I'm doing everything I can to get the problems resolved, but it may take a while. Also, I've been having problems getting mail. Some messages have been delayed for long periods, and others have apparently been lost entirely. I always respond quickly to mail messages, even if with only a short confirmation, so if you've mailed me and don't receive a reply within a day, please resend your message. Thanks, and sorry for the problems.
|
|
Saturday |
Saturday, January 30, 1999 The good news is that most of the problems with accessing this site appear to be fixed. The bad news is that I burnt most of yesterday trying to deal with them and looking into moving my site to another web hosting service. I still may do that. Robert Morgan recommended pair Networks, which was the one I spent the most time looking at. I checked several of the web hosting rating services, including Budgetweb, Host Find, Top Hosts, and The Ultimate Web Host List. The consensus seemed to be that pair Networks was a good one. They're cheaper than the web hosting service I'm with now, they have fewer disk and throughput constraints, and it appears they have better connectivity. I've already sent them a message with several questions that weren't clear from the materials posted on their web site. For now, at least, most of my web and mail server access problems seem to have gone away, but I'm still thinking seriously about moving to pair Networks. * * * * * Responses continue to come in to my query about the surface area of Ivory Coast. If you haven't responded yet, please do. If you responded, but didn't receive an acknowledgement note from me, please send your guess again. * * * * * I did my standard weekly full backup yesterday afternoon. Full in the sense that I back everything up that needs backed up, but not in the sense that I back up everything. I have entire partitions that I don't back up, for example those that contain copies of distribution CDs. But Backup Exec this time told me that I had about 6.1 GB of data to be backed up. In theory, that's well within the capability of my Seagate TapeStor Travan TR4 drive, a so-called "8 GB" drive. In fact, as with all tape drives, manufacturers assume 2:1 compression, and so double the native capacity of the drive. But I've never gotten anything like 2:1 compression with this or any other tape drive. Now it may be that my mix of data types is consistently less compressible than what the manufacturers are assuming, but over many years with many tapes drives on many systems using many backup packages, I've gotten 1.5:1 pretty consistently. That means that my 4 GB Travan TR4 can really store about 6 GB of compressed data, and the amount I needed to backup yesterday was perilously close to that figure. As it happened, I had 6,045,403,730 bytes to back up. When the verify pass finished, it showed that it had compared 3,891,703,049 bytes, which translates to 1.55:1 compression. That came closer to the actual 4 GB (actually four billion byte) capacity of the tape than I wanted to come. I really don't want to have to go to multiple tapes. In addition to the inconvenience of having to change tapes during a backup, the tapes themselves aren't cheap. So it appears that I need to look at a larger capacity tape drive. The new Travan NS20 drives provide 10 GB native and 20 GB compressed, so I'll probably be looking at one of those. * * * * * And while the verify pass was running, I decided to download the latest version of the PC Magazine benchmark tests, WinBench 99. It's about a 9 MB download, and it finished just about the time the verify pass finished. I installed the benchmark tests on kerby, my main workstation, just to play around with them. They installed with no problems. Then I made the mistake of firing up WinBench just to see what it looked like. I picked the "Run all tests" option. WinBench displayed an information box to inform me that it was going to defrag my drive and then reboot the computer. Not so much as a by-your-leave. As it happened, I had eight or ten instances of Internet Explorer running, each with a web site that I was doing research on. I hadn't yet gotten around to bookmarking several of those sites yet, so I figured I'd better kill WinBench, save the bookmarks, and then restart it. No dice. I right clicked the Task Bar to get Task Manager up, but WinBench kept clearing the menu before I could click on Task Manager. I watched every one of those IE instances disappear at the rate of about one per second, one after the other. There wasn't a thing I could do. After all the programs were closed, WinBench did its defrag, which lasted all of five seconds (I have Diskeeper running every night to keep the disk defragged.) The system then rebooted. It examined my system and found several things it didn't like. The show-stopper was that I had the AutoHide and Always on top check boxes marked for my Task Bar. For some reason, WinBench 99 doesn't allow anything to be configured for Always on top. I exited WinBench, fixed the Task Bar Properties, and restarted it. The test suite started at 8:15 and completed about a hour and a half later. This program provides some good information, but if you install it remember that it's a very aggressive program. * * * * * And another data point concerning the intelligence of Border Collies. Our younger dog, Duncan, loves to go roaring out the back door to the deck. He has done since he was a tiny puppy. He knows that squirrels live out there. Our gas grill also lives out there, and he long ago learned the word "grill." If anyone mentions that word, he scrambles for the back door and stands there whining desperately, with whines sometimes turning into howls. But yesterday we found that he'd carried the association one step further. I asked Barbara what we were going to have for dinner. She said, and I quote, "I think I'll do some steaks." We heard the mad scramble and found Duncan standing at the back door, whimpering. He made the connection by himself, that "steaks" means "grill" and that "grill" means the back door is soon to be opened. Geez. * * * * * And I'd better sit down and make a list of the stuff I need to pick up at CSO. I have several system boards on the way in that I want to test, and I don't have an unused ATX chassis in the house. I want to build a test-bench system to swap stuff in and out of, so I think I'll get an ATX desktop case. I also need to pick up all the incidentals--mouse, keyboard, network card, memory, video card, etc. And I'd better get some Y-splitters for power supply connectors, because I might want to put a bunch of drives in it. Come to think of it, I'd better get a keyboard adapter for AT/mini-DIN, too. * * * * * And while Robert Morgan and I were exchanging mail about web hosting services, I mentioned that I thought I'd need pair Network's "Advanced" account, because I needed FrontPage extensions. One of the reasons I mentioned for needing the FP extensions was that I had already implemented the FrontPage Search capability, which requires them. Robert sent the following response:
Hmm. Good point. I'm in the midst of creating a "Reports Home Page" right now, which I'll add as a link on the left column of my Daynotes page. While I'm at it, I'll also add a Search link. As far as your search, 99 pages does indeed seem excessive. My guess is that you entered the string robert morgan and that Search found any page that included "robert" or "morgan". Obviously, my first name appears on many of my pages, so that may have caused all the hits. I think you can enter "robert morgan" instead to find just the pages with your name. If the quotes don't work, you might also try: robert and morgan I was going to check that myself, but I can't get my damn search page to load.... And I've already started to modify my Daynotes page. Starting next week, I'll add links for "Special Reports" and "Search TTG" in the left column. Now, if only I could get the search function to work. The servers at my web hosting company are apparently too busy. When I try a search, it always times out. * * * * * And regarding access problems, this from Frank McPherson [frank_mcpherson@email.msn.com]:
Yes, I suspect you're right. I've seen several days since the first of the year when the Internet itself was pathetically slow in general. But the problems I was having were specific to BigBiz, or perhaps to MAE-West. I did frequent tracerts while I was having the problem. They'd get all the way to MAE-West and then die. BigBiz is co-located at Hurricane Electric (he.net), which is just one hop from MAE-West. BigBiz has been responsive, in the sense that they replied to email messages and called back when I paged them. They even offered to move me from server1 to server5, which they said was less heavily loaded. And that may have helped out with the web server response time problem on searches and so on. But the root problem was connectivity, and I'm not sure what if anything they can do about that. Nothing, if the problem is at MAE-West. * * * * * And now I'd better get to work on that shopping list for Computer & Software Outlet.
|
|
Sunday |
Sunday, January 31, 1999 The results are in on the Ivory Coast survey. I arbitrarily cut it off at 100 responses. As it turns out, the actual value is 124,503 square miles (sm) or 322,464 square kilometers (sk). The lowest guess was 125 sm (324 sk). The highest was 3,000,000 sm (7,769,995 sk). The arithmetic mean of all guesses was 151,814 sm (393,199 sk), which is roughly 20% high. However, when doing something like this, it's common practice to eliminate one or a few values from both top and bottom of the range, just to prevent guesses that are widely off from skewing the results. The 3,000,000 sm guess was by far the highest (the next highest was 500,000 sm). Knocking out it and the lowest guess gives an arithmetic mean of 124,299 sm (321,934 sk), which is 0.998+ of the correct value. * * * * * We picked up my friend Steve Tucker yesterday and headed for Computer & Software Outlet to pick up the stuff I need to build a test bed ATX box. I already have the system board, processor, keyboards, and a usable video card, so we got the following items:
Total damage, with tax, $303.47. Perhaps a bit more than it would have cost me mail-order, but there're no shipping charges, so it's nearly a wash. Some people hold sales taxes against local vendors, but that's not really reasonable. Even if I buy something mail-order, I have to tell the NC Department of Revenue about it and pay a "use tax" that's the same as the sales tax. Also, if something doesn't work, it's a lot less hassle to take it back to CSO than to have to ship stuff back and forth. I was also going to buy another keyboard/video/mouse switch and the necessary cables, but it's easy enough to switch the monitor between systems on my test bench, and I already have plenty of keyboards. Actually, I have plenty of unused mice, too, including probably half a dozen of the old "dove bar" mice. Just in case, I bought an AT to Mini-DIN keyboard adapter. I probably have both kinds lying around, but better safe than sorry. That's actually one of those items that I buy compulsively whether I need it or not. Yesterday was my first visit to a computer store in a long time that I didn't buy an IDE cable. I always buy IDE cables. I probably have literally fifty spare IDE cables in this place. But I always buy one just in case. I also picked up two CPU retention mechanisms for the Intel Seattle board. The eval unit that my Intel contact sent me was a used unit, and was missing the CPU retention bracket (the thing that stands vertically and provides supporting vertical slots to slide the CPU into). I was hoping that CSO had a dead Seattle board and that I could sweet-talk them into letting me salvage the CPU retention mechanism. As it turned out, they had a couple of spares in the back from another brand of system board that supplies the mechanism separately. I picked up a set for the Pentium II and another for the Celeron, although the Seattle-1 doesn't support the Celeron. Steve also picked up a set for his Pentium II, because the Seattle-1 he bought on an auction site didn't include the CPU retention bracket either. I looked at the HP SureStore 8100i IDE CD-ROM burner while I was there, but CSO wanted $379 for it, which seemed a bit much. I notice that those are selling for about $325 including shipping from mail-order places. Steve wasn't planning to buy anything before we arrived at CSO, but he too picked up a few compulsive cables before we left. We headed back for Steve's house to drop him off. We were planning to go over there around 5:00 p.m. anyway to have dinner and spend the evening. But then Steve and I got started down in his office/computer room. Eventually, Barbara offered to go back to our house, take care of the dogs, make dinner for my mother, and come back later on. So Steve and ended up working all afternoon on his computers. Around 6:00 p.m. we took a short break and headed for the West Town restaurant for dinner. When we got back, Steve and I headed downstairs again. We got to talking about my recent web site problems and the fact that I was considering changing web service providers. Steve ran the Wakeolda BBS (bulletin board system) for many years. He actually only shut it down for good a few months ago. He's been thinking about registering his own domain name and starting up a web site, and one thing led to another. Two of the web service providers on my short list are pair Networks and Burlee Networks. Both of them are well-regarded by users and get good reviews. After looking at both, Steve decided (with some nudging from me) to go with the pair Networks basic account. He registered wakeolda.com through them and signed up for service. pair Networks has a lot of good things going for them, not least of which is their superb connectivity and their sheer size (they run 160+ UNIX web servers). Their only downside for me (and it wasn't a major issue for Steve) is that they really want to focus on web serving versus handling email. They include only one mailbox in their standard price, but they do allow unlimited autoforwards to external mailboxes. You can buy additional mailboxes for $1/month. That's a fairly standard policy, but what's not standard is that they apply it to virtual mailboxes within your domain. For example, BigBiz includes one true POP mailbox, but I can set up as many virtual mailboxes within the ttgnet.com domain as I want and POP directly from each of them. The pair Networks mailbox policy may be a problem for me. If they allowed autoforwards to internal mailboxes, that'd be no problem. If, for example, my main mailbox could be thompson@ttgnet.com and I could setup autoforwards for webmaster@ttgnet.com, info@ttgnet.com, etc., that'd be fine. But it appears that I can't setup those autoforwards to thompson@ttgnet.com because it's an internal rather than an external mailbox. Well, I could set them up, but it looks like I have to pay a buck a month each for them. In addition, pair Networks places severe constraints on the size of those mailboxes. Various places on their web site, they list 0.5 MB, 1 MB, and 3 MB as the size limit for mailboxes. They also mention charging a $1 per week fee for going over. That's excessive. I don't tend to keep mail in my mailboxes very long. I POP every few minutes throughout the day. But sometimes I have quite a bit of mail accumulate overnight, and I do occasionally get very large files (e.g. chapters zipped up with figures) that would exceed the allowances. So, although pair Networks looks like a pretty good deal for Steve, I don't think pair Networks is for me. Once we got Steve signed up, we installed FrontPage 98 and I tried to show him some of the tricks and traps of using it. In a couple of days, Steve will be heading for Daytona for the race (he works for R. J. Reynolds), so we probably won't get his site up and running until after he gets back. * * * * * And I think my mail may be screwed up again. I checked my mail when we got home around midnight. I only had four new messages when I logged on this morning. Usually, I'd expect to have somewhere between 15 and 25 during that period. Again, if you send me mail and don't hear back from me within a day at most, please resend it. I probably didn't get it.
|
|
|
Coming Soon (I hope) Here are some things that are currently on my to-do list. I may start some of them this coming month. It may be a while before I start on some of the others, either because I don't yet have everything I need, because interdependencies make it necessary to do other things first, or simply because other work takes priority. But I'll get to all of them eventually.
|