|
Week of 6 September 1999 Sunday, 12 September 1999 10:52 A (mostly) daily journal of the trials, tribulations, and random observations of Robert Bruce Thompson, a writer of computer books. |
|
|
Monday, 6 September 1999 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] It's Labor Day in the US, and we're treating today as a kind of holiday, although we'll still do some work. I'll go ahead and post this this morning, but I'll be back to posting in the evening starting tomorrow. * * * * * That maniac Anand (when does he sleep?) has posted an article that compares every available AMD K6-*, Celeron, Pentium II, Pentium III, and Athlon processor running various benchmarks under Windows 98 and Windows NT. Anand has a strong anti-Pentium III / pro-Athlon bias, as his casual discarding of SSE-optimized benchmark results makes clear (Quote: "If you ignore the Pentium III's performance, the benchmark grows in usefulness."), but the comparison is interesting nonetheless. Although it's impossible to draw any overall conclusions about which is the "best" processor from this article, it's pretty clear that the Celeron is still the bang-for-the-buck champion by a large margin. But the recent price cuts on the slower Pentium III CPUs (and the fact that they'll be dropping even further in the next couple of weeks) makes the 450 MHz Pentium III a real alternative for those who will be running a lot of SSE-optimized applications. For this reason, we're still recommending that those who are considering buying a new motherboard go with Slot 1. That allows you to use an inexpensive Socket 370 Celeron in a Slocket for now, while leaving open the possibility of upgrading later to a Slot 1 Pentium III. * * * * * This from Paul S R Chisholm psrchisholm@yahoo.com:
Thanks. Although I haven't done a lot of testing yet, Start Office 5.1 appears to open and save Word 2000 and Excel 2000 files just fine. As far as addresses, I apologize if I used the incorrect one. I get a *lot* of mail, and when I'm processing it on the run I sometimes miss requests for me to use a different address. I usually pull the address off the header, so unless you have a sentence at the top that specifically requests me to use a different address, I may miss it. The other thing, of course, is that you'll have to explicitly request that I use a different address on a message-by-message basis. I've been taken to task by a couple of readers for using their real address on a message that didn't request otherwise because they'd previously requested that I not publish their addresses. Unfortunately, there's no way I can remember preferences for specific readers, so if you want to be anonymous or have me use a different address, please remind me with each message you send. * * * * * This from Tom Syroid [tsyroid@home.com]:
Well, you have more experience with StarOffice than I do, but it's good to hear that you haven't run into any show-stoppers yet. I haven't played with Styles in SO, but I did import the styles from Office as you suggested. It appears that they can be applied from the drop-down list in the upper left corner. If so, that's all I need. As far as the SCSI hard drive prices, I just went over to KillerApp, listed the top 100 hard drives, and looked for Seagate models. It's true that I just used the lowest price listed for each drive (where if I were really buying a drive I'd check prices from vendors I know and trust), but I figured that was a reasonable comparison. I think the Barracuda was listed for $293 and change. * * * * * This from Fred Mora [fmora at ibm dot net]:
Well, yes and no. No doubt I could have gotten started after eight hours, but I have years of experience with Windows NT. Getting to the same level with Linux would take me more years, and getting to what I consider minimum acceptable competence would certainly take months at least. I hope you're correct about StarOffice continuing to be available as a "fat client". I'd hate to make the effort to switch only to find that it had been mutated into a thin client. * * * * * This from J.H. Ricketson [culam@neteze.com]:
You are correct that I'm not covering MO in the Nutshell hardware book, although Pournelle and I will probably cover it in our big hardware book. For Nutshell, that's a matter of available space more so than any bias I have against MO. MO is and always has been a niche technology, although I grant that there are applications where it makes sense. But CD-R(W) has hurt MO, and DVD-RAM/+RW are going to hurt it worse. I don't necessarily think that MO will go away entirely, but I think large writable CD/DVD optical and products like the Castlewood ORB are going to turn traditional MO into even more of a niche product. * * * * * This from Brian Bilbrey [bilbrey@pacbell.net]:
Thanks. My current account at pair does not permit running custom scripts, although Jerry's does. I've been using the Thunderstone search engine, which works fine for my site. When we attempted to index Pournelle's site, the Thunderstone engine returned an error saying that the site was too large, so Pournelle definitely needs an alternative. Paul Robichaux sent me mail a week or two back that mentioned the atomz.com site, which I may check out for my own site. But your scripts may well be an alternative for Jerry.
|
|
|
Tuesday, 7 September 1999 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] Watching the U.S. Open yesterday, the same question came to mind as does every year. What happened to the score "40-all"? For those not familiar with tennis scoring, the point score in a game proceeds as follows: No points = Love (probably from the French l'oeuf, or "the egg"); the first point is 15; the second 30, the third 40; and the fourth wins the game. Unless, that is, the opponent has also scored points. In the case of a tied point score, the scores used to be called 15-all, 30-all, and 40-all. Once 40-all was reached, the opponent who won the next point was referred to as having the advantage. If the player with the advantage lost the following point, the score was called "deuce." But somewhere along the line, the score 40-all stopped being used. Everyone, including the official scoring system, now refers to a game tied at three points each as "deuce." A deuce can only follow an ad, so that isn't right, unless they've officially changed the scoring nomenclature while I wasn't looking. Granted, I last played serious tennis in the late 60's and early 70's, but surely I would have noticed such an earth-shaking change. I notice that the official rules of paddle tennis still list 40-all. So what happened to 40-all in tennis? * * * * * A couple of readers commented that the picture I posted earlier of Gypsie, the 10-week old Border Collie pup we're temporarily hosting, resembles a full-grown dog. For those readers, here is a (slightly blurred) picture of Duncan and Gypsie together. Duncan is our 4.5 year old male Border Collie. Although he's technically medium-size (it says so on his papers), he's definitely on the large end of medium. He weighs about 65 pounds, and towers over all the other Border Collies at Flyball practice. Gypsie will probably grow up to weigh about 35 pounds, and will probably be about 4 inches (10 cm) shorter than Duncan. * * * * * This from Tom Syroid [tsyroid@home.com]:
You probably already have what you need. Many SCSI host adapters have firmware utility suites that include a low-level format program. Run that program, and select detailed surface scan if the option is offered. If there is no format option in firmware, check the host adapter manufacturer's web site for a standalone format/maintenance utility. Alternatively, check the Quantum web site. Every hard disk maker provides low-level maintenance utilities, which are often a custom verion of Disk Manager. Again, run the utility, do a low-level format with detailed surface scan, and enable error remapping if that's an option. Be aware that it's extremely unusual for modern hard drives to start developing bad sectors unless they're very near failing, so any repair you do may be a temporary measure at best. However, the good news is that if you originally low-level formatted the drive under one brand of host adapter and didn't subsequently reformat it when you moved it to the current host adapter, the drive itself may be fine and the problems due simply to the change from one host adapter to another. * * * * * This from J.H. Ricketson [culam@neteze.com]:
Actually, DVD-ROM is dead standard at this point. The standards problem you mention arises because of competing writable DVD standards, and there may be fewer competitors than you think. DVD-R and DVD-RW are niche products, and are likely to remain so. The drives are currently selling in the $5,000 to $15,000 range, and are unlikely to get much cheaper. DVD-R disks (write-once) cost $40 to $50 per disc, and again that's not likely to drop all that much any time soon. DVD-R(W) is primarily used for DVD mastering and is unlikely to become a mainstream technology. NEC's MMVF has some advantages for video, but does not have widespread support in the industry. The real competition is between DVD-RAM and +RW. The market will almost certainly choose a winner from those two standards, and DVD-RAM has some very significant advantages. DVD-RAM is an official standard of the DVD Forum, which +RW is not. DVD-RAM has actually been shipping for more than a year now. You can actually go out and buy a DVD-RAM drive today, and it uses $18 to $30 rewritable cartridges. +RW is allegedly sampling now, although I don't see any evidence of that. I have a DVD-RAM drive sitting on my desk, but I can't get a +RW drive for love or money. That one-year advantage out of the starting gate may not be as much as an advantage as it first appears, though, because DVD-RAM has not sold a lot of units. But it does mean that DVD-RAM has had a year to debug and improve, so first-generation +RW drives will be competing against second- and third-generation DVD-RAM drives. The one major advantage that the +RW camp claims over DVD-RAM is higher capacity, but that won't last long. Although I'm under NDA, I think I can safely say that you can expect to see a new generation of DVD-RAM drives this fall that will boost single-side capacities from about 2.6 GB to about 4.7 GB. One DVD compatibility issue that is seldom mentioned is DVD-ROM drive read compatibility for writable DVD media. Most current DVD-ROM drives won't read any sort of writable DVD media. But note that the DVD-RAM camp (Hitachi, Matsushita/Panasonic, and Toshiba) manufactures the vast majority of DVD-ROM drives sold. At least one of those has already added support for reading DVD-RAM media in their latest DVD-ROM drives, and the others are likely to do the same. And, although the +RW camp (HP and Sony) say that only "minor changes" are needed to allow DVD-ROM drives to read +RW media, the chances are good that the DVD-RAM camp will not add +RW read compatibility to their drives.. So, although the market will determine the winner between DVD-RAM and +RW, and although my guess is that that winner won't be determined until perhaps mid- to late-2000, I'd bet my money on DVD-RAM. * * * * * This from Paul Robichaux [paul@robichaux.net] regarding the recent to-do about possible back doors in Microsoft encryption:
I'll bow to your superior knowledge of the topic. But I'm surprised that open source isn't the standard for encryption products. I thought that the standard method used to ensure the robustness of an encryption algorithm or program was to make the source publicly available. That would also obviously make it impossible to conceal a back door in the product. * * * * * This from Fred Mora [fmora at ibm dot net]:
Okay, I didn't realize that there was any guarantee that Sun would continue to make a non-networked version available. As far as your book idea, I think not. It's a good idea to focus a book on a specific market, but that's a bit too specific. However, I think a book on implementing Linux and Linux applications as a replacement for Windows and Office might do quite well. * * * * * This from Paul S R Chisholm psrchisholm@yahoo.com:
Thanks. I probably won't mess with scripts, simply because the www.atomz.com site that Paul Robichaux mentioned appears to do everything I need to do as far as searching, and does it simply. Right now, I'm still using the similar Thunderstone search service, but I'll probably convert over to the Atomz service in the next couple of weeks. * * * * * This from Paul Robichaux [paul@robichaux.net]:
Okay. That makes sense. Thanks. * * * * * Another from Paul Robichaux [paul@robichaux.net]:
Well, you obviously know a whole lot more about this subject than I do, so I'll just shut up now. But that does explain my question about why open-source encryption isn't the way things are done... * * * * * Still another from Paul Robichaux [paul@robichaux.net]:
He might think he'd love One Big Rack until he prices components for it. He can find rack mountable PC chassis (chasses?) at just about any vendor who specializes in network gear, but the prices are outrageous. The last time I looked, a high-quality unpopulated chassis with power supply ran upwards of $500, and if you want redundant power it's easy to spend a couple of grand without having anything but a chassis and power supplies. I had thought that PC Power & Cooling sold RM stuff, but I didn't see any when I did a quick check of their web site. * * * * * This from J.H. Ricketson [culam@neteze.com]:
DVD-ROM drives provide CD-ROM-like throughput. A 1X DVD corresponds roughly to a 9X CD-ROM. That might make it seem as though a 6X DVD would run CD-ROMs at 54X, but in reality DVD-ROM drives top out at about 36X or 40X max when reading CDs. That's certainly adequate throughput--up to about 6 MB/s. The problem, as with CD-ROM drives, is average access time. New generation DVD-ROM drives typically provide ~150 ms average access when reading CDs, or about a tenth the speed of a hard disk. That's usable for running an application, particularly if the app is loaded and cached, but it isn't going to be very fast for apps that frequently hit the source drive. MO has an advantage in this respect. As far as packaging, DVD-RAM media comes in two types, single-sided (2.6 GB) and double-sided (5.2 GB). There are also two types of cartridge, one in which the media is fixed and one in which it is removable. The DS media come only in fixed cartridges. The SS media come in either type of cartridge. The advantage to the removable media is that you can record a DVD-RAM disc and then remove it from its cartridge to play it in a standard DVD player (if that player can read DVD-RAM discs, which most can't). The disadvantage is that once you have removed the media from the cartridge, it probably won't ever be writable-to again. Incidentally, the +RW camp actually presents the fact that +RW media do not use a cartridge as an advantage. They point out that bare media will be cheaper, and that they will be able to fit DVD players and DVD drives in notebook computers. Still, on balance, as you say, the advantages of cartridge-based media outweighs the few advantages of bare media. And, as the DVD-RAM camp points out, CDs are the only high-capacity media that do not use a cartridge, and there's probably good reason for that. As far as buying a DVD-RAM drive, my advice would be to wait for at least a couple of months. The main DVD-RAM players are likely to be releasing a new generation of DVD-RAM drives that will have some very nice additional features. I can't say more right now because of NDAs, but I'd wait if I were you.
|
|
|
Wednesday, 8 September 1999 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] Early update today. Barbara is cleaning house, and I can't get much real work done anyway until she finishes. I'll go back to the evening update schedule tomorrow. * * * * * I sporadically receive comments from readers who are unable to retrieve the most current version of my journal page even when they Reload or Refresh. This doesn't happen often (or at least I don't hear about it often), but I've seen it happen myself, although not on my own web site. I've established to my satisfaction that the problem is not with pair networks or my server, so the only thing I can conclude is that some ISPs cache content and do not honor Reload or Refresh as a request to retrieve the document from the actual server. When I was looking at Ars Technica the other day, I noticed that their main page refreshed automatically, which it didn't used to do. Scanning the sourse for their page, I noticed a new line:
The purpose of that was pretty obvious. In their case, I suspect the reason for adding it is to force page reads to increase their ad revenues. But it's an equally good solution for my problem. Thinking about how my readers likely use my site, I concluded that the typical reader calls up my page and then clicks Refresh. In looking at my web stats, I found that only about 15% of my page reads showed the status "Not modified since last retrieval" (as compared, for example, to Pournelle's site, which shows that status for about 40% of page reads) So, unless I'm missing something, there's really not much downside to forcing a refresh. I've added that code to my index page and daily journal page. I'm not even sure if the "no-cache" tag will be honored by ISP caches, or if its effect is limited to local browser caches. Please let me know if there's something I haven't considered or if this change causes a problem. * * * * * Barbara is preparing this week to leave on a bus tour with her sister and parents on Saturday. That means I can have wild women and parties for a week or so. Either that or just work all the time. We talked about photos, and she's decided to take a couple of 35mm bodies with her and just shoot conventional color negative film. When she gets prints back, we'll scan them. At 1,200 dpi resolution and 36-bit color depth, that'll generate a 150 MB uncompressed data file from each 4X6" (10X15 cm) print. It'll be interesting to see what final file sizes PNG or one of the other lossless compression methods will yield. * * * * * This from tech_refresh@frontiercorp.com:
Thanks. I wasn't aware of the problem. It's not Netscape causing the problem. I just checked the last several weeks in IE4 on a system that runs 800X600. Some had the horizontal scrolling problem, and some didn't. When I've seen that problem in the past, it was invariably due to a long URL line or a bunch of nested blockquotes. Unfortunately, my pages are very long and I don't know of an easy way to find the offending line(s). When this has happened in the past, I've used Notepad to view the source, scrolled over to the right and looked for "long" lines. But I use a lot of embedded links, so many of the lines that appear "long" really aren't. Finding, fixing, and testing the fixes for the offending line(s) can easily take several minutes per page, and I simply don't have time to do that. Perhaps someone who knows more about HTML and FrontPage can suggest a fix. I seem to remember that I could force a wrap by making my page tables three cells wide (adding a one pixel wide cell to my table structure on the right). But I think I tried doing that once and gave up on it for some reason I don't remember. Perhaps one of my readers will have a good fix for this. * * * * * This from Paul Robichaux [paul@robichaux.net]:
Hmm. Your wife is more forgiving than mine. I showed Barbara a picture of a 19" rack one time, and she said "no way am I having that in my house." Actually, she was more emphatic than that. As far as re-wiring, you might want to think about using Category 6 or Category 7 cable rather than Category 5. I haven't priced them lately, so I don't know how much of premium they sell for over Category 5, but you might as well get ready for Gigabit Ethernet. * * * * * This from Bo Leuf [bo@leuf.com]:
No, I don't think I'm mis-remembering. In the late 60's, we had a very pedantic tennis coach. We players, of course, used barbaric abbreviations when calling the score during practices and matches ("five-all" instead of "fifteen-all", "ad-in" and "ad-out", etc.) One of things that drove the coach nuts was the fact that we called "deuce" when the score reached thirty-all. He carried around a little paperback pamphlet whose title was something like "USLTA Official Rules of Lawn Tennis". More than once, he showed me the section on scoring a game, and pointed out that "deuce" should not be called until an advantage had followed forty-all. For all of that, he was a pretty good coach. He was the one that finally convinced me to stop trying swinging forehand topspin volleys from mid-court when my opponent hit a floater... * * * * * This followup from Bo Leuf [bo@leuf.com]:
Yes, the question is not whether a tie score after eight or more points have been played is called "deuce." It is. The question is, what is the proper name for a score tied at three points all. As far as the rules changing, I believe that they hadn't changed for centuries as of 1968. But they've certainly changed a great deal since 1968. Consider: sudden death (sets used to have to be won by two games, and I recall watching many matches with set scores of 18-16 or higher--it was a better game before VASSS. Nowadays, one can lose not just a set, but an entire match without once having one's serve broken); the rule on foot faults (formerly, your feet could not cross the plane of the baseline before you struck the serve. Now, one is permitted to leap into the court, so long as one's feet do not touch the surface of the court before the serve is struck); "Play shall be continuous" is no longer honored (vis-a-vis that recent hideous innovation, the "injury time-out"); all men's matches were formerly best of five sets, and now the vast majority are only best of three sets. And so on. Things were better under the old rules, when tennis was played on grass--as it was intended to be--when everyone wore white (with only accents of maroon and navy permitted), when the balls were white, and when rackets were made of wood. I was probably born too late. The greatest player who ever picked up a racket, Big Bill Tilden, died on 6 June 1953, the same day I was born. I don't like what tennis has become. * * * * * This from Chuck Waggoner [waggoner at gis dot net]:
I hadn't thought about XyWrite in years, although I used it for a couple of weeks at one point, probably a dozen years ago or more. I remember when most serious writers used XyWrite for its speed and power. I was using WordPerfect 4.1 or 4.2 then, and I was so used to it that I decided it wasn't worth making the change to XyWrite. I'm sorry to hear that Y2K will render it obsolete. * * * * * This from S.S.Ghosal [alpana@cal.vsnl.net.in]:
Congratulations on your new site. I visited it this morning (I note that I was visitor number 95) and it appears that you've made a very good start. I'll check in periodically to watch it as it grows. Incidentally, although I am not a lawyer, I suggest that you post a stronger notice to protect your content, something like "Copyright © 1999 by Soumya Shankar Ghosal. All Rights Reserved."
|
|
|
Thursday, 9 September 1999 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] Tomorrow is our wedding anniversary, which I've never forgotten thanks to PCs and calendaring software. We will have been married 16 years as of tomorrow, and have had surprisingly few problems during that time. Especially after I learned not to introduce Barbara to people as my "first wife". I thought people's reactions to that were interesting. She was not amused, however. Come to think of it, I tried to convince her to get married early the following year instead of in September. I thought my argument made good sense, because the "marriage penalty" at that time meant that getting married in September ended up costing us something like $4,000 in additional income taxes. But Barbara wasn't swayed by that argument. Interestingly, every man I subsequently mentioned that idea to agreed with me, and every woman agreed with Barbara. But that's something I've never seen listed as a secondary sex characteristic. * * * * * And speaking of the Battle of the Sexes, I took a short break this afternoon and noticed that USA Network was televising the Mixed Doubles final. Mixed Doubles is interesting because the ability of the woman player almost always determines which side wins. Most people who don't play tennis seriously don't realize just how big the differences are between male and female players, and both the television commentators and the tennis establishment try to minimize those differences, as with the bogus speed-gun service speeds. Consider that Richard Krajicek served 48 aces during his match last night, and only one of those was faster than Venus Williams's supposed fastest. Yeah, right. In fact, the speed of Krajicek's serve (or that of any other hard-serving man) is at least 50% faster than the fastest of the women servers. Back when the women were paid much less--as they should have been, since they played one less round and didn't draw nearly the paying crowds--no one made any bones about the vast differences between men and women on the court. But nowadays, Political Correctness and Marketing rule, and we're all supposed to pretend that there aren't any differences. In fact, a good high-school boy player hits the ball much harder and covers the court much better than the best woman who's ever played. Chris Evert at her peak, for example, practiced with the Fort Lauderdale high school boys tennis team and lost most of her matches against them. In a serious Mixed Doubles match, both men spend all their time pounding on the opposing woman, and whichever woman does a better job of standing up to that pounding wins. In an ideal Mixed Doubles match, neither man will ever have an opportunity to strike the ball except when he is serving or returning serve. The closer a Mixed Doubles team can come to enforcing that ideal on their opponents, the more likely they are to win the match. I used to play a lot of Mixed Doubles, and enjoyed it a great deal. The wonder was that the women we played with also seemed to enjoy it, although they spent most of their time trying to dodge cannonball serves, getting nailed at the net by hard, flat groundstrokes, or getting overheads shoved down their throats. I always figured that men enjoyed playing Mixed Doubles because it was the only venue in which it was not only acceptable but expected that they would behave aggressively toward women. Mad at your girlfriend? Take it out on the woman across the net. Just never, ever play a Mixed Doubles match with your wife or girlfriend on the other side of the net. You'll never live it down. I made that mistake just once, and ended up having to find a new girlfriend. * * * * * Barbara has been cleaning out filing cabinets downstairs, and has found quite a few interesting things, including many old photographs, some apparently dating from the 1880's, old documents, and so on. One of the things she found was a stock certificate from the Glass Casket Company (now there's a disturbing thought) of Altoona, Pennsylvania. Apparently, my grandfather bought $150 worth of shares in that company in 1919, which was serious money back then. She's researching the company, which will probably turn out to have disappeared many years ago. But we're hoping, of course, that it turned into Microsoft or something. She doesn't want to handle old original documents any more than necessary, so she asked if I could scan them for her. On the More Is Better Theory, I did a scan this morning of the cover letter at 1200 dpi and True Color. That took about half an hour to do and turned into a 267 MB PNG file. Better, I suppose, than its uncompressed size of 394 MB. I'm going to have to do some thinking about what resolution to scan her trip pictures at. I think I'll go with the highest optical resolution my scanner supports, True Color, and PNG compression. We'll see. * * * * * This from Bo Leuf [bo@leuf.net]:
Yes, tennis has become completely commercialized, more's the pity. That was brought home to me last night when we were watching Serena Williams beating Monica Seles. The commentators announced the teams who had made the mixed doubles final. I'd never heard of any of the four players, and I doubt any of them has ever even qualified for a singles main draw. In the old days, the mixed doubles final would have been something like Rod Laver/Margaret Court versus John Newcombe/Evonne Goolagong. Nowadays, very few of the big-name singles players even play doubles, let alone mixed doubles. There was never any money in doubles or mixed doubles, but the players recognized that playing those events was part of their obligation to the fans. Now they can't even get the big names to play Davis Cup. They say that they want more money. As it is, they're being paid $100,000 per match, win or lose. As with other sports, big money, television, agents, and the whole commercial bit has ruined tennis. * * * * * This from Paul Robichaux [paul@robichaux.net]:
Well, there's no doubt that a rack organizes things, but I see Barbara's point. It is kind of industrial looking. As far as the Category 5, forget I said anything. I saw the other day that 1000BaseT over Category 5 is now a formal standard, so your new wiring will take you to at least 1000 Mbps in the future. Actually, perhaps more than that. In the same way that you can run 100 Mbps over Category 3 cable by using all four pairs, I suspect you'll be able to do the same for whatever comes after 1000BaseT on all four pairs of a Category 5. * * * * * This from centralsoft@ukonline.co.uk:
Actually, my problems have more to do with Microsoft applications than Microsoft operating systems. I've had relatively few problems with Windows NT 4 over the years (although I wouldn't use Windows 9X on a bet). Actually, it's ultimately the fault of the government. If they'd just left Microsoft alone, perhaps Microsoft wouldn't have felt compelled to make their browser an inherent part of the operating system. I think most of the problems come from that. As far as OS/2, it's a dead product, so I can't see the wisdom of moving to it. Even IBM have abandoned OS/2. Infoworld says that IBM have announced that they plan to migrate all 250,000 of their systems to Windows 2000, which pretty much puts the final nail in the coffin of OS/2. Although I'd like to move to Linux simply because I don't like Microsoft, the reality is that I haven't time to learn Linux and anyway I have commitments to write two Windows 2000 books. So, I'm stuck with Microsoft operating systems for the foreseeable future, but that doesn't mean that I'm stuck with Microsoft applications. * * * * * This from Svenson Sjon [sjon@svenson.com]:
Thanks. I'm glad to hear something honors that tag. Immediately after I uploaded my page yesterday, I tried calling it up in IE5. It didn't force a refresh the first time, of course, since it had the older version of the page in cache. I refreshed to load the new version with the no-cache tag, exited IE5, and started it up again. It re-displayed the new version of the page without forcing a refresh. I have no idea what's going on. I'd think it was something about the way I have my copy of IE5 configured, but that same copy does a forced refresh every time on the Ars Technica page. Perhaps one of my readers can tell me what I'm doing wrong. * * * * * This from Robert Rudzki [rasterho@pacbell.net]:
I know nothing about Macs, but I agree that that claim seems a bit excessive. Tom's Hardware, for example, probably does much less than a tenth that amount of traffic, and it's running on about six Pentium III class servers at pair Networks. But then Apple is famous for making wildly exaggerated performance claims. Remember their ridiculous claims for the iMac? * * * * * This from Tom Syroid [tsyroid@home.com]:
I hate it when someone gets burned by a product I recommend. But then, in all fairness to Maxtor, it seems likely that you were just unlucky. I've been running several of those drives for months without any problems. If there were any widespread problems with them, I'm sure I would have heard about them by now. As far as the Atlas, it is a Quantum model. I'm not very familiar with Quantum drives, particularly their SCSI models, but I did a quick check of their web site and found two Atlas models--the Atlas IV and the Atlas 10K. Checking KillerApp.com turned up some other models, incuding an Atlas III, which I assume are recently discontinued. The Atlas IV page shows that it's a 7,200 RPM 6.9 ms drive with several SCSI interface options, including the new 160 MB/s extensions to U2W. Just from reading the specs, it appears that the Atlas IV should be noticeably faster than the ATA Maxtor. As I recall, your EPoX motherboard has U2W SCSI, which would probably be the best choice. * * * * * This from Scott McIntyre [smcintyr@mail.com]:
Thanks. It's clear to me from your and others' messages that Netscape and IE behave differently. I don't use Navigator for much other than checking page appearance occasionally, so I appreciate readers who use Nav letting me know about problems. * * * * * This from Jan Swijsen [qjsw@oce.nl]:
Yes, at first I thought that my problems with IE5 not doing the forced refresh had to do with its settings, but that appears not be the case. I can't say I'm surprised at Netscape not handling the HTML error. People ding IE for not being compliant with various HTML issues, but the fact remains that I prefer my browser to do something rational when it encounters an HTML problem rather than simply blowing up. * * * * * This from Paul Robichaux [paul@robichaux.net]:
I really can't see migrating to 1000BaseX anytime soon, either. I've just completed the migration to 100BaseT. The performance is noticeably higher than 10BaseT (perhaps twice as fast), but I think I'll be satisfied at 100 Mb/s for the next few years anyway. By that time, 1000BaseT will probably be selling at commodity prices. If I can run it over my existing Cat 5, I'll think about upgrading then. Otherwise, I'll just stick with 100BaseT. Incidentally, you do know that you can get an auto-sensing 10/100 hub that will allow you to run mixed 10 and 100 cards on one segment, right? Those hubs cost a bit more, because they're really two hubs in one case, with the 10 hub bridged to the 100 hub. But you'd mentioned having to buy a new hub, and I can't see buying a 10BaseT hub nowadays. Incidentally, I notice on The Register that Apple may be backing off their position that G3s can't be upgraded to G4s. I don't know much about Apple, but my understanding is that the 1.0 ROMs would have allowed the processor upgrade, but Apple sneaked through a hack on the 1.1 ROMs that prevents G3 owners from upgrading a a G4 CPU. Apparently, the huge outcry has caused Apple to reconsider, and they may be releasing a 1.2 ROM that re-enables the CPU upgrade. I don't know if that affects you or not. * * * * * This from Tom Syroid [tsyroid@home.com]:
Yes. About 99% of the people who think they understand what MTBF means, don't. If graphed, the MTBF curve on your Maxtor would take the form of a Poisson Distribution, with a few drives failing very young, the vast majority failing a few years out (tailing off to almost nothing), and a tiny number surviving for decades. I remember reading back in the 1980's about a lightbulb that had been turned on in 1918 and had been burning ever since. The Poisson curve grants the possibility of such anomalies. I've often wondered about Old Tom Parr, who supposedly lived to be 352 years old. I've never run the numbers on it, but it's theoretically possible. At a guess, though, he'd be something like one in a quadrillion or quintillion. There haven't been that many people who have ever lived, so it would definitely be an anomaly. * * * * * This followup from Tom Syroid [tsyroid@home.com]:
Well, I was thinking about installing RC1 this weekend, but I think I may just wait for RC2. Thanks for the tip. * * * * * Another followup from Tom Syroid [tsyroid@home.com]:
Sorry to hear that. I haven't tried running NT5 on a dual processor system with USB, but I'll let you know what happens when I build my dual processor system.
|
|
|
Friday, 10 September 1999 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] The Register reports that Intel is about to cut prices yet again. As of 12 September, Intel will cut the price of the Celeron/400 from $78 to $70; the Celeron/433 from $98 to $85; the Celeron/466 from $118 to $105; the Celeron/500 from $173 to $160; the Pentium III/550 from $470 to $410; and the Pentium III/600 from $670 to $610. The Pentium III/450 remains at $185, and the Pentium III/500 at $254. All of this in preparation for the 27 September introduction of the 133 MHz FSB Pentium III/533B, which will debut at $380, and the Pentium III/600B, which debuts at $660. AMD has so far held out against making significant price cuts on the Athlon, but that can't last much longer. At this point, the Athlon rollout is beginning to look like an utter failure. They're not shipping product in volume, few motherboards are available, and those that are are expensive and of very poor quality (some have already been recalled). Although Tom's Hardware, Anand, and the other enthusiast web sites are trumpeting the wonders of the Athlon, it appears to me that the Athlon is turning into an unmitigated disaster for AMD. Saying that the Athlon is faster than the Pentium III (which is dubious in itself) is all well and good, but the point is that AMD must deliver product in volume to make the Athlon a commercial success, and they're showing no signs of being able to do that. AMD has bet the company on the Athlon, and it's looking like a losing bet. Anyone who buys an Athlon now is choosing a risky course. And just in time to cancel out those Intel price cuts, the price of RAM is shooting up dramatically. Since the first of the month, the cost of memory has about doubled. * * * * * Barbara came home this afternoon from errands with two anniversary presents for me (and we agreed that we wouldn't get each other anything--never believe a woman when she says that). She got me a new pair of jeans and the latest Tess Gerritsen novel, Gravity. If you like medical thrillers, get to know Tess Gerritsen. This is her fourth book, and the first three were superb. There's a Stephen King blurb on the front of this one that says she's better than Palmer, Cook, and Crichton. He's right. She is. The best part was how much Barbara paid for the book. It lists at $25 and had a 30% off sticker on the front. Barbara is a member of some frequent buyer program at the bookstore where she bought it, which got her an additional 10% off. So her price was $15. She hadn't been in lately, so they sent her a $5 gift certificate, dropping the net price to $10. Not much more than remainder price. She is one crafty shopper. And Ms. Gerritsen gets her royalty of $3.75 or so, which she wouldn't have gotten on a remainder sale. * * * * * This from M. Praeger [athyrio@hotmail.com]:
You may be right, we may be thinking of different people, or the source I read this in many years ago may have contained a misprint. I think perhaps it was an old edition of the Guinness Book of World Records, although it was 25 years or more ago that I read about it. I don't usually forget numbers, though, and an age of 152 is easily within the realm of possibility, so that wouldn't have stuck in my mind. I did an Internet search last night, and found only one reference to Old Tom Parr, which gave his age at death as 135. * * * * * This from Shawn Wallbridge [swallbridge@home.com]:
I agree. I've never been able to figure out Netscape apologists who maintain that it is proper behavior for a browser that encounters an HTML error to crash and burn horribly. In what way is it better, unless one considers Nav to be an HTML error checker rather than a working browser? Surely most users would prefer the browser do something reasonable and attempt to render the HTML to the best of its ability rather than GPFing. * * * * * This from Jan Swijsen [qjsw@oce.nl]:
There's something strange going on with my own page as well. I'm not sure exactly what combination of things results in the behavior I've observed, but my page automatically refreshed yesterday when I called it up in my browser. Apparently, some combination of the no-cache meta tag, the browser cache settings, and perhaps the phase of the moon controls whether or not the page auto-refreshes. To recap, I took the following steps:
The only thing I can figure out is that the browser cache is comparing the page in cache with the page on the server, and doing an auto-refresh only if the page on the server is more recent. But it shouldn't be caching the page at all. I have no idea what's going on. * * * * * This from Ed Bott [ed@bott.com], one of the authors who participates on a private Windows 2000 mailing list oriented toward authors writing W2K books. The ongoing discussion had to do with ACPI and power management problems with W2K running on notebook systems.
Thanks. As of this afternoon, I couldn't find anything on the ASUS site about the problem. I've never used ASUS motherboards, although I know they're well-thought of by many. I'm sure that some of my readers use this board, however, so they'll no doubt appreciate the heads-up. * * * * * This from Scott, Terry [tscott@atinc.com]:
I've never used a TNT2 board under Windows NT, although I do have systems running NT with the Riva 128 and Riva TNT chipsets. I've never had any problem with those. The fact that you see the same results with two different models of TNT2 card is probably not significant. Those cards are probably identical for all practical purposes and could likely use the same drivers. The fact that you experience a hang even when loading the vanilla VGA driver tells me that the driver is unlikely to be the problem. More likely, there's some kind of hardware resource conflict that Windows 98 is resolving and Windows NT isn't. Windows 98, of course, is treating these boards as AGP video adapters. NT doesn't support AGP, so it recognizes them as simple PCI devices. My first thought would be to check for IRQ and/or Base I/O address conflicts. Other than that, I'm not sure what I'd try. I'll post this to my journal page tonight. Perhaps one of my readers will have some ideas.
|
|
|
Saturday, 11 September 1999 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] Barbara left last night to spend the night at her parents' house. They departed on their bus tour at about 7:00 a.m. this morning, leaving her Trooper in the K-Mart parking lot where the bus picked them up. My friend Steve Tucker ran me over there to pick up Barbara's truck so that it wouldn't sit there for a week. I remembered that much. Now I *must* remember to return to that parking lot next Sunday afternoon to pick up Barbara, her sister, and her parents. If I forget to do that, I'm dead meat. * * * * * ASUS has posted a page describing the problems with ACPI and Windows 2000 on P2B motherboards. * * * * * I continue to be amazed by this Tecmar Travan NS20 tape drive. I just used it to run a backup on Barbara's main machine, theodore, and the Windows 98 box odin. The backup data set totaled about 6.5 GB--3.7 GB on theodore (local to the tape drive), and 2.8 GB on odin (backed up via my 100BaseT network). Using BackupExec 7.3, the overall backup rate was about 82 MB/min (5 GB/hr), less than the 120 MB/min at which the drive is rated, but impressive nonetheless. For data on the local hard drive, the Tecmar maintained just under 100 MB/min (6 GB/hr). This is one fast tape drive. At a street price of $350 to $400 for the bare drive, it may be overkill for a typical individual PC, but it's just about ideal for a small network server. Highly recommended. * * * * * This from Brian Bilbrey [bilbrey@pacbell.net]:
Thanks. When I first started this site, I attempted to make sure that it was friendly to both Netscape and IE. I had to use FrontPage (or something similar) just because I didn't have the time or inclination to learn how to hand-code HTML. I decided to keep things as simple as possible--text and tables--and avoid anything that might cause problems in either browser. At first, I'd preview each page in both IE and Nav before posting it, but eventually I concluded that that wasn't worth the time it took. So nowadays I just type away in FrontPage Editor and post it. Other than the occasional "long line" problem I haven't had many complaints from either IE or Nav users. As far as hit rates, I normally download and process my web stats once a week, so I hadn't done them since I made the change. I just checked yesterday's stats, and they appear pretty normal. I got 791 page views yesterday, which is par for a Friday. What's really odd is that the percentages of status code 200 (OK) and 304 (Not modified since last retrieval) are about the same as always. Since I've had this no-cache meta tag in effect for a couple of days, I've noticed that it does indeed have an effect in my browser (IE5) at least. The page doesn't auto-refresh unless I've published a new version of the page since the last time I retrieved it. So, if the page is old, nothing happens. If the page is new, it gets retrieved automatically. I think that's probably a pretty good arrangement. * * * * * This follow-up from Brian Bilbrey [bilbrey@pacbell.net]:
I'm sure you're right, but I still don't understand some of the behavior I'm seeing. I do run a proxy server (WinGate), but I have caching turned off on the server. Just this morning, I hit the Ars Techica site, and it did not auto-refresh. At first I thought they'd eliminated the no-cache tag, but it's still there. So I don't understand what's going on. * * * * * This from tek1@inwave.com:
Thanks. I was familiar with the EXPIRES meta tag, and had actually considered using it at one time. I'm not sure why I decided not to, but there was a reason, or so I thought at the time. In fact, I seem to recall that there's an option to expire immediately, which would supposedly cause the page to be reloaded every time it was viewed. The way things are set up now appears to work properly, at least in my browser. This page auto-refreshes if and only if the page on the server has changed since the last time I retrieved it. * * * * * This from J.H. Ricketson [culam@neteze.com]:
Thanks. The only one of those I've used is MemTurbo (which I concluded was probably useful for Win9X, but of limited use for Windows NT). I may take a look at some of the others. * * * * * This from Tim Werth [twerth@kcnet.com]:
Good questions. I don't know the answers, although I suspect that everyone who learns tennis scoring immediately wonders the same things. Tennis is an ancient game. Henry VIII, for example, in his younger (and slimmer) days played a version of tennis called court tennis, which kind of resembled a cross between modern tennis, squash, and racquetball. That was nearly 500 years ago, and the game is more ancient still. As far as I know, all versions of tennis since the beginning have used the same names for point scores. When I read your question, my first thought was to convert to Roman numerals to see if that made sense. But somehow "XV", "XXX" and "XL" don't seem to make much more sense. Of course, forty was sometimes written as "XXXX", but that doesn't help much either. Perhaps one of my readers will have the time to track down a good answer, if in fact one exists. * * * * * This from Bo Leuf [bo@leuf.com]:
Thanks. Obviously, I either misremembered the 352 year old figure or there was a typo in the source I saw it in. Probably the latter, given that I seldom misremember numbers and had the final two correct. That and the fact that I explicitly recall thinking at the time that that age would make him more than a third of Methusalah's alleged age. I don't doubt an age of 152. Statistically, that would probably make him on the close order of one in a billion, so it's not unlikely that he did in fact reach that age. * * * * * This from J.H. Ricketson [culam@neteze.com]:
I don't think you need to worry about starting a flame war (probably a poor choice of phrase given the subject). No one, no matter what his political stripe, can fail to be disturbed by the recent revelations about what went on at Waco. I am no friend of religion of any sort, and most certainly not of "cults." In fact, twenty-three years ago I lost the woman I intended to marry to Moon's Unification Church. So, few people could have been expected to have less sympathy for the Branch Davidians as they were being presented on the evening news during the siege. And yet, even then I suspected that we were being lied to by the government and that those people in the compound were probably victims of the government rather than aggressors as they were represented to be. When I heard about the feds using tanks against civilians, I knew something evil was going on. I haven't changed that opinion, and little that has been revealed recently surprises me at all. I think that all of those involved on the government side, up to and including Mr. Clinton, should be prosecutable and prosecuted by the state of Texas on first-degree murder charges. I've read this article. Obviously, the gentleman who wrote it may have an agenda of his own. But I find what he has to say much more credible than I do what the government has said.
|
|
|
Sunday, 12 September 1999 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] Well, this is getting ridiculous. I just noticed the download time estimator at the bottom of FrontPage Editor, and it tells me that this page will take nearly 2 minutes to download. Obviously I need to do something. I could, of course, cut down on the number of messages from readers that I post, but I didn't want to do that. So, I considered two possibilities:
So, starting tomorrow, there will be two pages. I'll create links to the current Daynotes page and current Mail page on the home page, and continue to make the current pages bookmarkable. The journal page will continue to be thisweek.html. The current mail page will be newmail.html. I'm going to discontinue the bookmarkable page lastweek.html for two reasons: first, it gets very little traffic (only 69 page reads last week). Second, it's easy enough to view last week's Daynotes simply by clicking the "Last Week" link in the current Daynotes page. I've avoided taking this action in the past largely because of two irrational concerns: (a) that I won't be able to think of anything to write about and my journal page will be blank, and (b) that readers won't send me any mail to post. Neither of those seems likely. I hope this change meets with readers' approval. If nothing else, it will allow you to read one page while the other loads in the background. * * * * * This from Scott McIntyre [smcintyr@mail.com]:
Yes, that is interesting, and sounds like a possible explanation. The only reason I think it may not be true is that I think the 15-30-40 tennis scoring predates clock faces. But who knows. * * * * * This from J.H. Ricketson [culam@neteze.com]:
Well, it *is* an advantage, for three reasons:
All of that said, I think DVD-RAM will probably win the war, but I don't think anyone will know for sure which standard will triumph until mid- to late-2000. * * * * * This from Tarun Ghosal [alpana@cal.vsnl.net.in]:
Thanks. I'll post you message on my daily journal page in case any of my readers care to visit your site. Good luck.
[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] |