|
Week of 2/22/99 Sunday, February 28, 1999 12:39 A (mostly) daily journal of the trials, tribulations, and random observations of Robert Bruce Thompson, a writer of computer books. |
|
Monday |
[Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] Monday, February 22, 1999 Back to work on the book. As I open a new week's Daynotes file I'm wondering if I should stick with the weekly format or go to a daily format. The drawback to the weekly format is that the files get very large. The FrontPage status bar gives the download time for this file right now as about 5 seconds. That assumes a 28.8 dialup, I think. Once filled out, the files take a great deal longer to download. I noticed when I finished last week's file, for example, that the download time was showing as 93 seconds. Someone who reads this every day ends up reading the Monday notes on Monday, but downloading those Monday notes six more times throughout the week. But I haven't thought through all the implications of doing a daily file. Certainly, I could put a "Yesterday" and a "Tomorrow" link at the top and/or bottom of each page. That'd allow people to scroll back and forth easily through the days. But I really don't have time to do a lot of manual organizing and link creation, so I'm not sure how I'd index the thing. I'll keep thinking about it. If anyone feels strongly one way or the other, please let me know. * * * * * This from Jeff Powell [powelljh@ci.richmond.va.us] about the future of modems:
Modems have no future. They've reached their ultimate pinnacle with V.90. The limiting factor is telephone lines themselves. The telcos restrict bandwidth on a standard phone line to sidebands of 3 kilohertz. Subtracting guard bands and so forth leaves about 2,600 Hertz usable. The Shannon Limit says that a channel this size can transfer, as I recall, a maximum of about 3,600 baud. The best currently available encoding methods allow you to encode 16 bits per baud, yielding 57,200 bits/second. I may have these numbers a bit scrambled, but the essence of it is that we're never going to be able to transfer data faster across a standard telephone line than we do now. And that's a hard limitation of physics rather than a technological limitation. The wires themselves, you understand, can transfer data much faster by using an alternative technology like ISDN or xDSL.. It's the POTS circuit that can't go any faster. And the days of circuit-switching for data connections are rapidly drawing to a close. Back when I ran a bulletin board system, I had lots of users. Each of them typically called not just my BBS, but perhaps a dozen or so other BBSs. So the connections were serial many-to-many. Nowadays, most people call just one number with their modems, that of an ISP. There's no longer any need for circuit-switched connections. A point-to-point connection serves the purpose more efficiently, and allows much higher data throughput, as with xDSL or a cable modem. As xDSL and cable modems become ubiquitous, I suspect that the phone company is going to have a lot of people cancelling extra phone lines. If the standards bodies ever get around to developing standards for IP-based faxing, I suspect it'll largely replace standard faxing almost overnight. The phone companies make a lot of money on fax lines and fax long-distance calls, so I suspect they're doing everything they can to prevent IP faxing from becoming standardized. It's probably a bigger real threat to them right now than voice over IP. * * * * * UPS just showed up with an Intel RC440BX motherboard. Like the EPoX EP-BXT, the Intel RC440BX is an integrated motherboard. But where the EP-BXT includes Intel740 video and Yamaha sound, the RC440BX includes integrated nVidia RIVA 128ZX video and integrated Creative Labs SoundBlasterPCI 64V audio. The RC440BX is available in two form factors: the Micro-ATX version has three PCI slots and one shared PCI/ISA slot. The full-size ATX version has three PCI slots, three ISA slots, and one shared PCI/ISA slot. The two versions of the RC440BX are otherwise identical. I've worked quite a bit with the EP-BXT already, and have a very high opinion of it. Although I haven't had a chance to work with the RC440BX yet, I suspect it will also end up being one of my favorite boards. The EP-BXT is a bit less expensive (street price is about ~$150US for the EP-BXT versus ~$200US for the RC440BX), but some will prefer the RC440BX nVidia graphics to the EP-BXT Intel740 graphics, or the RC440BX SoundBlaster audio to the EP-BXT Yamaha audio. Unfortunately, both of these motherboards have poor distribution. I checked PriceWatch and found a few hits for the EP-BXT and only one for the RC440BX. But it's worth spending some extra time tracking down an EP-BXT, and I suspect the same will prove true for the RC440BX. I'll have detailed reports on the RC440BX once I get some free time to work with it. * * * * * The following from Gary M. Berg [Gary_Berg@ibm.net]:
Okay, thanks. That makes sense. I should have realized this myself, because I have exactly the same problem on Pournelle's site. Tracking down all the new stuff there seems like an Easter egg hunt at times, and I'm always left with the vague feeling that I may have missed something. Actually, I was kind of hoping that people would prefer the way I'm doing things now, because doing daily pages would be more work for me. Not so much in creating the pages themselves as maintaining the links. Your message is the first I've received about this. It'll be interesting to see if anyone prefers the idea of daily pages.
|
|
Tuesday |
Tuesday, February 23, 1999 The following exchange with Bruce Denman mentions Tom's Hardware and got me to thinking. Actually, I'm seriously considering opening a new site devoted to hardware reviews, advice, and so on. It would have absolutely no advertising, and would be completely user-supported. A lot of it would be open to the public, but I'd also offer memberships on a subscription basis. Members would have access to the members-only areas, would have first access to stuff before it was posted a week or a month later in the public areas, would be able to post (versus just read) to the forums, etc. But in order to do something like that, I'd have to have some assurance that it would pay its own way. Realistically, it would have to generate a minimum gross of $100,000/year to start in order to pay for the staffing and other costs involved. With something like this, a business school concept called Price Elasticity of Demand is a major factor. In other words, setting the annual subscription fee is key. If I could get X number of subscribers at $30/year, I'd have to get 3X at $10/year to break even. Actually, I'd need a bit more than 3X to account for the additional administrative load in handling three times the number of subscribers. So the question becomes, can I get 10,000 subscribers at $10/year, or would I be better off aiming for 3,333 subscribers at $30/year, or even 2,000 subscribers at $50/year? What do you think? Would you be willing to pay anything at all for member-level access to such a site. What would you want to see on that site? What could I do to differentiate the site and make it useful, and worth paying for? If you're willing to take a few minutes to tell me what you think, I'd appreciate it. * * * * * This follow-up on cookies from Bruce Denman [bdenman@FTC-I.NET]:
I kind of figured that a clean install might clear up the problems you were having with Cookie Crusher. That's a pretty well-known product, and I'd be surprised if it had bugs that would cause problems as serious as those you described earlier. Of course, with Windows one never knows what interactions are going on between products. As far as Tom's Hardware, it is an advertising supported site, so I'm not surprised that it's cookie heaven there. The man gets a half million page reads a day, and I've never been able to figure out why. Surely I can't be the only one who sees at least the appearance of a conflict of interest in a site that purports to offer unbiased reviews of products and yet accepts advertising. And we're not talking about trivial amounts of money, either. Mr. Pabst apparently pays employees from that advertising revenue, and I believe someone told me that he was opening a testing lab. I have the same problem with AnandTech. * * * * * And this follow-up from Bruce Denman:
I'm not surprised that you need to strip Win9x down to bare metal periodically. I've been running Windows NT almost exclusively since version 3.5 shipped. I keep a couple of Win9x boxes around the house for testing, doing screen shots for books, etc., but I can't imagine actually using it as my day-to-day operating system. As far as Tom's Hardware and ads, I always think of that fundamental precept of Roman law, cui bono (who benefits?). As regular readers of my Daynotes column know, I'm pretty free with my criticisms. I don't go out of my way to find bad products and trash them in print, but neither do I hesitate to point out bad things. I'd like to think that if I had advertisers I could remain completely objective, but I can't imagine that that would be the case. * * * * * I've gotten a bunch of replies about the page-a-day versus page-a-week question. The first is from Bruce Denman:
Thanks. I've done everything I can think of to reduce download times, including pretty much eliminating graphics, making each day a separate table, etc. But the reality is that by the end of the week, my Daynotes page is going to take a minute and a half or so to download, just because there's a lot of text in it. * * * * * And this from Alberto S. Lopez:
Thanks. Your opinion about the weekly format seems to be shared by all. Good luck with the NT box. * * * * * And this from Wesley Moore:
Thanks for the kind words and the comments. You're by no means alone in finding my site from Dr. Pournelle's. I'd guess fully 10% of my regular readers found my site by following a link from Jerry's. I do like Buffy. It's well-written, well-acted, and surprisingly complex for a TV show. I mean, there are Good Good Guys, Bad Good Guys, Good Bad Guys, and Bad Bad Guys. I like women who stand up for themselves, and that's certainly the case there. My wife, who is a librarian, likes the show for obvious reasons. How many other TV shows have a librarian who goes out and kicks butt? And you're right about changing email applications. If there's one thing that I don't change lightly, it's my email. I'm quite satisfied with Outlook 98. It's not the most powerful mailer I've used (that honor goes to Pegasus Mail in spades), but it does what I need to do and integrates PIM functionality. I've also used Eudora Pro, and I can't figure out why it's as popular as it is. It's neither as powerful as Pegasus nor as convenient as Outlook 98. I've been on the Internet since about 1987, and I've probably tried at least a dozen mailers during that time. For me, at least, Outlook 98 is Good Enough. * * * * * And this from Bo Leuf:
Yep, that would all make sense if I went to a page-per-day format. But I don't think I'm going to. I've gotten a dozen or so responses already, and going to a page-per-day format has gotten zero votes so far. Thanks for the suggestions, though. * * * * * This from Chuck Waggoner [waggoner@gis.net]:
I think Microsoft blew it in naming subscriptions. Until I installed IE5, I thought subscriptions were related to channels and push technology. I had no interest in that, so I didn't pay any attention to subscriptions. I suspect many other people assumed the same thing, because Microsoft no longer calls it "subscriptions" in IE5. It's now "off-line browsing," which got my attention immediately. I'm not surprised that digital is coming to the fore in traditional movie theatres now. Like television, movies use relatively low picture quality compared to still photography. Stringing together frames at 24 or 30 per second overcomes minor defects in the individual images, so I guess digital is good enough for movies, although it's still not ready for prime-time in still photography for other than very specialized purposes like news photography. You're not likely, for example, to see Advertising or Industrial photographers move to digital any time soon. Although movie theatres may cheer the arrival of electronic distribution, it's really another nail in their coffins. Movie theatres, like book and music stores, have traditionally used distribution friction to ensure their profitability. Anything that reduces that distribution friction is not good news for them in the long run. Although I seldom agree with John Dvorak, he has a very good column this week that explains why MP3 is going to kill the traditional music industry. Electronic distribution will ultimately kill the book and movie industries as well. Not the content development portions of them, but the distribution mechanisms. * * * * * And this from Tom Syroid Tom Syroid [tsyroid@home.com]:
Well, the week-per-page format is certainly less work for me. Everyone seems to prefer it, so I'm sticking with it. Everyone wins. * * * * * This from Robert Webb [webbr2@nationwide.com]:
Thanks. Good Enough is good enough. I think the saying "Perfect is the enemy of good enough" is now attributed to Soviet Admiral Gorshkov, but I'm pretty sure it dates back much further than that. Something tells me that someone will track that down and tell me.
|
|
Wednesday |
Wednesday, February 24, 1999 Thanks to everyone who emailed me to say that this page was scrolling horizontally. I tried to fix the problem yesterday and publish the changes, but I wasn't able to publish to my server. At first, I thought the problem was at the server, but I was having all kinds of problems getting to anything on the Internet yesterday afternoon, so it appears that either my ISP or the Internet in general was having some problems. So I'm covered up in mail. I'd thought about staring a separate mail page, but that would give me two pages to keep updated and you two pages to read each day instead of one. The overwhelming concensus seems to be that people want everything in one place. So I'll keep putting reader mail here unless it just become ridiculous. I finished up one chapter yesterday and got a good start on another. Barbara started feeling ill last night. I'm afraid she may have what I had. When she starts feeling better, I'll have her go over the chapter I just finished in detail before I submit it. And now to see if I can publish this morning... * * * * * A followup from Bruce Denman [bdenman@FTC-I.NET]:
Thanks. Yes, you really should try NT. As far as IE5 beta, I'm not surprised they pulled it. They probably don't want people downloading and installing the beta if the release version is going to ship shortly. I expect they'll post the release version for download before long. In the meantime, IE5 beta is working fine for me. And this follow-up from Bruce:
Okay. Thanks. * * * * * This from Gary M. Berg [Gary_Berg@ibm.net]:
Hmmm. This doesn't happen to me too often, but when it does it's almost always deeply nested blockquotes that are causing the problem. Ordinarily, I can just call up the page in IE, do a View-Source, scroll Notepad's display all the way over, and then start scrolling down. The line causing the problem normally stands out because it's extremley long in Notepad. I just did that, and can't find an extremely long line. I noticed a couple--most those from people's URLs in their sigs and the embedded references to NY Times URLS, etc.--that were a bit long, but nothing extraordinary. At any rate, I've gone back and put the long URLs on their own lines, so that may fix the problem. * * * * * This from Cam Voelker from Waterloo Ontario Canada [cvoelker@ionline.net]:
Well, I guess it's a matter of philosophy. I believe that people should get what they pay for and pay for what they get. Medical care should, like anything else, be rationed by market forces. I'd like to drive a Rolls-Royce, but I can't afford one so I make do with my Isuza Trooper. I don't begrudge those who can afford a Rolls-Royce their right to buy one. And the argument that health care is somehow essential and so should be subsidized misses the point. The problem with subsidizing something like medical care is that the demand, although finite in an absolute sense, is effectively unlimited, while resources are very much limited. All that subsidized "universal" health care does is make resource allocation decisions that twist resource delivery in favor of the poor at the expense of the middle- and upper classes, who are ultimately the ones paying for it. In theory, the idea of universal health insurance coverage is attractive to many people. But in practice, it never works out. Rather than raising the poor and uninsured to the level of the rich and insured, it simply reduces the quality of care available to everyone to a least-common-denominator level. We see that happening now in the US, where women who have just given birth are being sent home early and against their doctors' advice because insurance no longer covers a longer stay. These are women that have medical insurance, and would formerly have stayed in hospital for two or three days, but are now being forced out in one day. There's a good reason why people from all over the world, including Canada, come to the US for health care if they can afford it. The level of care available to those who can pay for it is still by far the highest in the world. But the evil "if everyone can't have it, no one can have it" mindset damages that. Equipment or drugs that are too expensive for general use will no longer be developed because there will be no one to pay for them. In the past, this stuff trickled down. The wealthy and insured paid the development costs. As those costs were amortized, the technology became cheaper and more widely available. But the way things are going now, companies won't risk the venture funding needed to make breakthroughs because they won't be able to earn back those costs. That damages the structure of the medical care system, and it's all due to socialized medicine. As far as grouping people for insurance, that is the entire heart of the matter. If you don't group people into reasonable risk-groups, a process called adverse selection occurs. Assume, for example, that life insurance was sold without consideration for sex, age, or health. In order to make money, the insurance company would have to set premiums at a level that covered their losses and administrative costs. But under your system, a 25 year old woman in good health would pay exactly the same premium as an elderly man with severe health problems. What happens, of course, is that the elderly man buys as much insurance as he can afford at a rate that's being subsidized by the young woman. But she's no fool. She understands that she's being grossly overcharged, and so either goes without insurance entirely or else buys insurance from another company that does recognize that she should pay a lower premium. My objection to their charging me a higher rate for smoking a pipe, grouping me in with cigarette smokers, is that pipe smokers demonstrably (by every study ever done) live longer than even non-smokers and have fewer and less severe health problems than do non-smokers. You are correct that this is probably because people who smoke pipes are classic laid-back B-type personalities, but the fact remains that overall pipe smokers should get a better rate than non-smokers. And, yes, pipe smokers do have a higher incidence of mouth cancer than non-smokers, but pipe smoker's lower incidence of such things as heart attacks, strokes, etc. more than makes up for that. * * * * * This from Dave Farquhar [farquhar@lcms.org]:
I'd like to think there's a market, too, but I'm not sure there is. At least not on a subscription basis. As the US government discovered a long time ago with withholding, it's a lot easier to get money if you make the process essentially invisible than if people actually have to cough it up. Advertising-based sites work because their funding is invisible to the user. We all pay a hidden "advertising tax" every time we buy something. Then we get "free" access to stuff, whether it's Tom's Hardware or network television shows. So, in essence, companies make us pay for the privilege of watching their commercials. But with that comes biased material. Unless and until one realizes that the "content" is just the cheese in the trap, one might believe that the content is the purpose and the commercials are just a distraction. The reality, of course, is that the commercials are the purpose and the content is simply a means to an end. And that's why I believe you're wrong about firewalling advertising from editorial. Oh, I know that many organizations make an effort to do that. Jerry frequently mentions BYTE and the way they completely isolated him from advertiser pressure. And I'm sure he's right in a direct tactical sense. If Jerry had something bad to say about the latest Frammus, I'm sure that he could say it without worrying about what would happen when BYTE got a letter of protest from Frammus, Inc. But in a larger, strategic sense, no advertising-supported endeavor can long survive if it does not place the needs of advertisers first. The needs of consumers must take a far-distant second, if that. Perhaps that's why BYTE folded. Maybe they made the mistake of thinking that they could place the needs of consumers above those of advertisers. If I start up such a web site, I want to be accountable to my readers, and only to my readers. As anyone familiar with public television will tell you, very few people voluntarily support something if there's no good reason to do so. So if I am going to make a subscription site work, it'd have to be by giving people a strong incentive to become paying members. If micromoney ever becomes ubiquitous, that'd work, too. Most people wouldn't mind in the slightest having five cents a day or a penny per page or whatever deducted from their wallets invisibly. Actually, micromoney would be the best method, because one would get immediate feedback. With annual subscriptions, if you're doing something wrong the first real sign is when people don't renew their subscriptions in droves. With micromoney you get that feedback on a daily basis. * * * * * This from Tim Werth [timothy.werth@eds.com]:
Thanks. I suspect the reason that these sites concentrate on 3D and gaming is because that's what a lot of people want to read about. As Willie Sutton never said, "I rob banks because that's where the money is." But the Good Enough issue is something I hadn't though about. * * * * * The following from Warrick M. Locke [warlocke@mail.wf.net]:
Thanks. It's not just people running 800X600, either. I run 1024X768, and I can't see past about Tuesday or Wednesday. I've put a horizontal menu at the very top that should solve the problem. * * * * * This from Tom Syroid:
You're right that such a site would be very labor intensive. I couldn't do it all myself, which is one of the reasons why I mentioned that it'd need to generate at least $100K/year gross to start. I have four books currently under contract with O'Reilly, including the one I'm working on now, the book with Pournelle, and two W2K books. * * * * * The following from Dennis Loretz:
Yes, that sounds fine. Windows NT Server 4 will run happily on your main system. In fact, with 64 MB, it'll probably "feel" a little faster than Win95. If you don't care about NTFS security and so on, you can just install Windows NT to your existing FAT partitions, assuming they're FAT16 rather than FAT32. The 486/66 is marginal for NT, at least as an interactive system. But it'll do fine as a BDC, and could also be converted to a network "resource server" which you could use to share hard disk space, share Internet access via a proxy server like WinGate, etc. With 100BaseT getting cheap now, I'd be inclined to go with 100BaseT stuff. Make sure the cables you get are Category 5 (rated for 100 Mbps use) whether or not you start with 10BaseT or 100BaseT. Small 100BaseT hubs aren't that much more expensive than 10BaseT hubs nowadays. You can find four- and five-port 100BaseT hubs for $50 or $75 and eight-port ones for $150. Make sure to get enough ports. Four- and five-port versions are normally very close in price, and that extra port is worthwhile. Consider going to an eight-port version if there's any chance you'll ever want more than four or five ports. If you decide to use your existing 10BaseT adapters with a Fast Ethernet hub, make sure it's a "10/100" hub rather than a "100 only" hub. The latter work only with 100BaseT cards, while the former automatically detect 10 versus 100 and run the port at the proper speed, allowing you to mix and match 10 and 100 cards. * * * * * Late Morning: Well, it took more than 15 minutes to publish this morning, but it did eventually publish. I wasn't sure if I'd fixed the horizontal scrolling problem which wasn't showing up on my 1024X768 monitor, so I tried downloading the page on a system that runs 800X600. It was still scrolling. I finally tracked the problem down to a long NY Times URL. I hate to edit reader mail in even minor ways, but I decided it was justified to change that message slightly by embedding the URL into the description rather than leaving it visible as plain text. And that solved the problem. The URL was simply too long in text form to display properly on an 800X600 screen. And Internet Explorer 5.0 is about to drive me mad. The first bad thing I noticed versus IE4 is that the auto-complete doesn't work the same way. For example, with IE4 when I wanted to visit Pournelle's web site, I could just type "www.jer" in the address field. After a brief pause, IE4 would fill in http://www.jerrypournelle.com and I could just hit enter and go to the site. With IE5, it doesn't work this way. When I type in "www.jer", IE5 displays a drop-down list of sites that match the fragment I entered. I then have to move the mouse to the one I want and click on it. This is significantly less convenient than the IE4 method. In fact, it's useless. Microsoft still calls it "auto-complete" but what it really is is "auto-list". But that's not what's driving me mad about IE5. The problem is this: I open a site, say www.microsoft.com in the first instance of IE5. I navigate that site and find a page I want to keep up for the time being. Then I decide to visit another web site, so I open another instance of IE5 and enter the address for that new site. But IE5 displays that new site in the first instance of the program, rather than the new instance. The new instance always displays my start page, which happens to be the home page of this site on the local copy. And it doesn't matter how I open the new instance. Whether I double click the IE5 icon on the desktop or choose File - New - Window from within IE5, it always opens the new site in the old window. That's just about enough to make the product unusable for me, and I can't find anywhere in Tools - Internet Options or elsewhere to change this behavior. Well, now it's time to try publishing this. It's about 11:45 a.m. EST. If I had a tough time getting it published earlier this morning, it's liable not to work at all now. But I'll give it a shot.
|
|
Thursday |
Thursday, February 25, 1999 I was able to publish normally yesterday afternoon. It took a bit longer than it usually does when I publish in the morning, but not excessively so. I'm beginning to suspect that the time-out problems I sometimes experience while publishing have more to do with the current activity level on the Internet than anything else. There doesn't appear to be any way to modify time-out settings in the local copy of FrontPage. There may be such a setting on the Server Extensions, but I haven't checked the docs for them to find out. Even if there is, I'm not sure I could talk my web service provider into changing them. And there's been some discussion about whether and how much reader reviews on Amazon.com affect book sales. Common sense says that a lot of good reviews should help a book sell better, but I'm not sure how much of a factor, if any, they really play. Some books have a lot of mediocre and bad reader reviews and still appear to sell well, but a lot of good reviews can't hurt. At least I suppose that's the case. There are currently four reviews, all five stars, posted for my latest book, Windows NT TCP/IP Network Administration, but the more the better. If you've read it and would be willing to take a few minutes to write a reader review of it on Amazon.com, please click here. Please note that I'm not trying to stuff the ballot box. If you haven't read the book, please don't post a review as a favor to me or anything. * * * * * This follow-up from Dennis Loretz:
I'd think twice before I built a low-end Pentium system, unless you'll be building it exclusively from scrounged parts. If you have to buy stuff, for not much more money, you can put together a much more capable system around a Celeron/300A, which you can overclock to 450 MHz if you're ambitious. As far as dual-booting, you're going to have a problem with FAT32. Installing WinNT as dual-boot on an existing Win9x system running FAT16 is a no-brainer. You just install WinNT, and it configures dual-boot automatically. With FAT32, the problem is that WinNT can't see the partition and so can't install dual-boot. I dislike FAT32. It's noticeably slower than FAT16, and the only two problems it solves aren't really very big problems anyway. FAT32 lets you use larger partitions and avoids wasting disk space with large clusters that result in a lot of slack space. But for most people, using multiple 2 GB partitions is a better idea than using one huge partition, and with disk space at a couple cents a megabyte, slack space is a minor consideration. * * * * * This from Wesley Moore:
I looked at the specs on the FD-81 and FD-91 models, and I would indeed like to get evaluation units of them. But Sony is so large that even finding a contact person there is nearly impossible. I actually spent an hour trying to do that, sending emails to various addresses from which I never heard back, and bouncing around in their automated attendant system. If anyone knows of a good contact person there, I'd appreciate the information, but I simply don't have the time to spend on tracking one down. With digital cameras, naturally enough I suppose, people tend to fixate on the inherent resolution. That's important, certainly, but just as important is the optical quality of the lens itself. On the FD-71, Sony uses a decent glass zoom lens that provides about 10X optical zoom. I've seen many other cameras that use cheap lenses more suitable for an inexpensive point-and-shoot camera, and attempt to make up for the lack of decent optics with "digital zoom". The problem with digital zoom is that it's not really zoom at all. All they're doing is taking the existing image and making it larger, filling in pixels for which they have no information with interpolated pixels. Back when I was building telescopes, we called this "empty magnification". In other words, making the image larger, but without any additional information. In short, I'd rather have a Sony that offers 1024X768 inherent resolution and is equipped with a good lens that has a wide optical zoom range than a camera with somewhat higher inherent resolution that's equipped with an inferior lens. * * * * * I just spent fifteen minutes of my time to correct a 45 cent billing error on my cell phone bill. I use GTE Wireless, and they seem to have a problem billing me for the correct amount. It may be coincidental, but the mistake is always in their favor, and always by a small amount. The deal I have bundles in some unrelated services, including the Mr. Rescue roadside assistance program. A couple of months ago, they increased the price of that service from $2.50 to $2.95 per phone. They billed me the $2.95 on each of the two lines that's included in the deal, but credited me only $2.50 per line, leading to a bill that was $0.90 too high. I called them last month and they said they'd fixed it for that and all subsequent bills. When I got this month's bill, I noticed that it was for $44.87 rather than the correct $44.43. That sounded suspiciously like they'd failed to issue a $0.45 credit on one of the lines (with the minor difference being due to taxes charged and credited), so I looked at the bill. Sure enough, they'd credited me the full $2.95 on my first phone, but only $2.50 on the second. I called them up and they fixed it for this bill, again claiming that it'd be fixed for good. We'll see. But what I wonder is whether this is really unintentional or whether they do it on purpose. Overcharging by $0.45 or $0.90 may not seem like much, but I wonder how many people just pay the damned bill rather than wasting their time to fix an error of a few cents. But if that error is on 100,000 people's bills, that means that GTE is overcharging by at least $45,000 per month or $540,000 a year, which goes straight to their bottom line. If it's on a million bills, that adds $5.4 million to their profits. I wonder.
|
|
Friday |
Friday, February 26, 1999 I sent another chapter off to O'Reilly first thing this morning, and will dive into the next chapter as soon as I finish publishing this. I'm doing some internal reorganization of the site, so if you notice any broken bookmarks, that's probably why. Right now, I have probably 200 files or more in my main folder, including close to fifty Daynotes weekly files, many images, etc. I'm gradually creating a folder structure and moving files to folders appropriate to their purpose. I say gradually, because I don't want to have to republish my entire web all at one time. Publishing is iffy enough when there are only a few files to be updated. But I'll leave the commonly accessed files--the home page, thisweek.html, lastweek.html, etc. in the home folder, so most bookmarks should continue to work normally. And in the course of doing this I've found yet another bug in FrontPage. The Timestamp bot has two settings. One causes FrontPage to display the date the file was last edited manually, but ignores automatic edits such as link updates when files are moved. The second causes FrontPage to display the date the file was last changed in any manner, manually or automatically. I have the timestamp bots on all my pages set to the first option. But it doesn't work properly. Some, perhaps all, of my pages are displaying an updated date when all that's happened is that they've been updated automatically to reflect a new folder location. Oh well. And I have my weekly network backup and other administrative stuff to get done. As the penguin says, it's back to work for me...
|
| TTG
Home
Saturday |
Saturday, February 27, 1999 Well, it finally happened. I tend to put off going to the barber, particularly when I'm busy, which I usually am these days. My barber doesn't take appointments, and I sometimes have to sit for an hour waiting. So I seldom get my hair cut more than about once every three months, and it's sometimes four. My hair is curly, so it doesn't get in my eyes or anything. It just gets thicker and thicker. When my hair starts to get very long, Barbara frequently urges me to visit the barber by telling me that if I don't she'll cut it herself. The other night, I told her to go ahead and do it. She did, and she did a pretty good job of it, I think. We went to dinner last night at a little Chinese place in a strip shopping center not far from our home. When we arrived, there was a parking place right out front. We'd pulled partially into it before we noticed a couple of ducks (Mallards, I think) nestled up against the concrete log that blocked the parking place from the sidewalk. They were a male and a female, obviously a mating pair, and we at first thought that one of them had been injured. We were going to call the Wildlife Rescue folks, but noticed the restaurant staff watching us and the ducks. We went in, and they told us that the ducks were regular visitors. They come by every evening for a handout of fried rice. I woke up early this morning, so I decided to write a brief update here and see what happens when I try to publish before dawn on a Saturday morning. I'll give it a try now. * * * * * 8:00 a.m. - well, that didn't work very well. FrontPage appeared to be publishing normally, but after it uploaded the pages it sat for several minutes and then timed out, telling me that publishing had not succeeded. I also relearned a FrontPage lesson: never answer Yes to all when FrontPage wants to overwrite or delete files on the server. Actually, I verified each file individually, but I answered Yes to each prompt. That was a mistake. FrontPage told me that I had deleted several images from the Images folder and asked if I wanted to delete them on the server. I didn't think I'd deleted them, but I had been moving stuff around, so I told it to go ahead and delete them from the server. That was a mistake, as I verified later by looking at the pages that contained the images. I copied and pasted them from my backup, so they should be back where they belong when I publish this. * * * * * Barbara was terrified last night by the way our younger Border Collie, Duncan, was behaving. He was making hacking sounds and shaking his head. He'd just been outside, and she was afraid he'd gotten a twig lodged in his throat or something. I thought he just had a severe case of hiccoughs, because he was breathing normally, bringing the ball to play, and otherwise seemed normal. Because of Barbara's concern, I even tried the Doglich Maneuver to expel anything he might have caught in his throat. I finally got Barbara to call our friend, Sue Stephens, who is also our vet. Barbara hates to impose on her, but I have no such qualms when the health of my dog is in question. As it turns out, Duncan (who has allergies) was sneezing. Sue explained that dogs sneeze two ways. First, in a way that pretty much resembles a person sneezing. That we are used to seeing. But Dr. Sue says that dogs also sneeze internally when they're trying to clear their snouts. We'd never seen that before, and that's what Duncan was doing.
|
|
Sunday |
Sunday, February 28, 1999 As my long-time readers know, one of my concerns is the perversion of science to suit the goals of the Politically Correct and other know-nothings. Although this should go without saying, it doesn't nowadays: a scientist is an investigator, not an advocate. Terms like "environmentalist scientist" are oxymorons. A scientist has no predetermined notions about how an experiment will (or should) turn out. To a scientist, the purpose of an experiment is to go there and find out, not to "prove" a preconceived belief. A scientist gathers facts, examines them, and attempts to craft a hypothesis that fits all of the known facts. If a scientist uncovers new facts that do not fit the existing hypothesis, he discards or revises his hypothesis to accommodate the new data. A fact is. A fact speaks for itself. It requires no justification. A fact must be reproducible by others who perform the same experiment. To a scientist, there is no such thing as an inconvenient fact, or one that should be explained away or ignored. Science, then, is based upon reason and logic. But pseudo-science is becoming increasingly accepted in the collective consciousness. Worse still, the distinction between science and pseudo-science is becoming blurred. Pseudo-science has always been with us. Alchemy, astrology, numerology, and so on appeal to the credulous. In the past, intelligent, educated people have recognized pseudo-sciences for the frauds that they are. But our schools long ago ceased teaching our children even the rudiments of logic and reasoning, and many (perhaps most) of our newly-minted "scientists" are not scientists at all. Oh, they have degrees in chemistry or physics or biology, they wear white coats, and they work in laboratories, but they're not really scientists at all, any more than astrologers are. Their logic processes are flawed or non-existent. A scientist is a scientist because of the way he thinks, not because of his college degree or his job title. As an example, I corresponded not long ago with a young man who represented himself to be a scientist, and who was in fact employed by the U.S. government in such a role. During the exchange of emails, I mentioned the famous post hoc ergo propter hoc (after this, therefore because of this) logical fallacy. He had never heard the term. Worse still, he was unfamiliar with the concept itself, which is fundamental to science. I referred him to a web page that explains logical fallacies. Perhaps that will be enough to get him started in the right direction, but I somehow doubt it. In the current Politically Correct environment, science is under attack. That is so because science concerns itself with facts, and facts are often inconvenient things for those who are pursuing an agenda. Those who are the enemies of science would prefer that it just go away quietly, leaving them free to proliferate their ignorance. But science is strong, so they must instead take measures to kill it or co-opt it. A frontal assault is always costly and uncertain of success. Attack from within is slower, perhaps, but ultimately more certain. And so we see a two-pronged attack on science. First, attack science through the popular media, by attributing positions to science which no self-respecting scientist holds. See the Junk Science web site for many examples. Second, choke off the flow of new scientists by mandating Politically Correct curricula in science education. Science must ultimately die if there are no new scientists. If this attack on science succeeds, we're destined for a new Dark Age. And right now it looks as though the bad guys are winning. * * * * * Afternoon: Well, I wasn't able to publish this morning's update because my web server is non-responsive. Something tells me I won't have much luck this time, either, but I'll give it a try. I know it's the server itself causing the problem, because I'm able to ping it. The connectivity is fine. It's the server itself that is apparently running so slowly that requests time out. I've mailed BigBiz several times over the last few days, but have yet to hear anything from them. Supposedly, the move from Server01 to Server05 was to have solved the problem. If anything, it's gotten worse. I've about had it with BigBiz. I plan to move my web site to pair Networks in March, so there may be some disruptions later this month. I'm also having problems with mail being delivered very late or not at all. If you send me mail at my ttgnet.com domain and don't receive a response in a day at most, you can resend it to thompsrb@bellsouth.net. This is actually my wife's POP account, so please don't send stuff there unless it's urgent that I see it quickly.
|