Email
Robert |
Week of 26 June 2000 Friday, 05 July 2002 08:04 A (mostly) daily journal of the trials, tribulations, and random observations of Robert Bruce Thompson, a writer of computer books. |
|
wpoison
Search [tips]
|
Monday, 26 June 2000 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] As usual, I have more to do this week than time to do it. I have three chapters in progress for PC Hardware: The Definitive Guide, a web site to get up for PC Hardware in a Nutshell, and a system to build around the Intel D810EMO motherboard (for which I still haven't found a case--a FlexATX case with no front slot for an FDD but a front I/O panel for USB ports and an audio connector.) I'll probably build it on my new-fangled plywood chassis until I can come up with a case to hold it. If all else fails, I'll stick it in a regular ATX case, but that seems a pity. Actually the new web site is in progress, although it's not even fully stubbed out yet. If you want to see it under construction, visit www.hardwareguys.com (but please don't tell anyone else yet). If you get a 404 error, that's because the DNS information hasn't propagated to your nameservers yet. It may be a couple more days before that happens for many of my readers. My target date for the web site to go live is 31 July, which'll be a stretch. But at least the site is responding and I can get mail addressed to the hardwareguys.com domain. I don't usually spend much time reading legal documents, but The Register has posted the text of Hitachi's response to the suit filed against it by Rambus. You can safely ignore the prefatory material and instead search for "affirmative" to locate the defenses raised by Hitachi. If all or even most of this is true, Rambus has behaved abominably and deserves to be slapped down. Unfortunately, Hitachi settled with Rambus, apparently because Hitachi is in the process of selling their memory business to NEC and didn't want complicating factors to delay the sale. Although Toshiba also settled, that doesn't mean that Rambus will go unchallenged. Micron and other memory manufacturers have not settled, so it's likely that this case will eventually go to court. * * * * *
Yep. Very elderly drivers are a big concern here in the US. On the one hand, they need to be able to drive for practical reasons, because in the US for most people few things are within walking distance of their residences. Just getting groceries or visiting the drugstore requires a car in most US areas, let alone visiting the doctor or the library. On the other hand, elderly drivers are a hazard to everyone else. In general, they have poor vision and poor reaction time, which means they tend to get in a lot of accidents. Fortunately, most of those are at low speed, which brings up the other problem. They tend to *cause* a lot of accidents. Not long ago, I followed a car driven by a very elderly man with his equally elderly wife as passenger. He was doing literally 12 or 15 MPH in a 45 MPH zone, with traffic backed up behind him. It was a two-lane road, with no passing allowed. In my younger days, I would have said the hell with it and passed him anyway. Nowadays, I have (a little) more patience. I've often said that the only requirement for a driver's license should be a simple road-test. All you'd need would be a radar gun, a set of giant calipers to measure the maximum width of the car being used for the test, and two adjustable vertical steel beams. Set the beams so that their minimum separation is 6" (150 mm) wider than the maximum width of the car being used for the road test. To pass the test for one's driver's license, all one must do is pass through those vertical beams at a minimum speed of 60 MPH (say 97 KPH). If you make it through without contacting the beams, you get your license. If you don't (or aren't willing to try), you don't get your license. They could even have practice areas with nerf-beams. When my father taught me to drive, he insisted that I learn absolutely where the passenger's side of the car was. I learned my lesson well. When I took my driver's license test at 16 years old, the examiner for the road test had me do all the usual stuff. As we were returning to the office, he pointed out a car parked parallel to the rear of the building and told me to maintain my speed and pass that car as closely as I felt comfortable doing. As we approached the car, he shouted "Jesus Christ" and stepped on the (non-existent) brake pedal on the passenger side. He was still shaking as he signed the paperwork that granted me my license. As he handed it to me, he said, "You know, that was my car, and you missed it by literally an inch." I was a smart-ass even then, so I replied that I could probably drop that to half an inch if he wanted to give me another shot at it. * * * * *
Thanks for the kind words. Sounds like you have your wife well trained. * * * * *
Yes, I got one of those obnoxious .cc mailings a couple of weeks ago. They hope to use the power of marketing to convince people that a sow's ear (a .cc domain name) is just as valuable as a silk purse (a .com domain name). In reality, of course, there's nothing special about the .cc top-level domain. It's no more prestigious than any other ISO-3166 geographic domain name, which is to say not very. And, while many geographic TLDs are cheap or free, this outfit is charging a huge premium for .cc names. I can register a .com for less than $15/year at Joker.com. They want $100 for two years (and presumably $50/year thereafter) for a .cc domain name. They can keep their .cc domain names. I hope no one is stupid enough to pay their price, but I'm afraid there are plenty of morons around who will. Reminds me of years ago when Ford ran an ad campaign to try to convince people that the Ford Granada was just as good as a Mercedes. Yeah, right. People just laughed at them. It was a bigger joke than the Edsel. Perhaps the same thing will happen with this. * * * * *
Yes, certainly. I was aware of the domains (you can find a complete list here) but what I don't understand is how some of these places qualify as countries. I also don't understand why anyone would pay the inflated rates that most of these places charge for domain names when .com, .net, and .org domains are available so cheaply, as are geographic domain names in many countries. Actually, I do, because these dinky little places have basically eliminated the hierarchical nature of geographic domain names. That is, if IBM wants to register a .us domain, they might have to register ibm.armonk.ny.us, while they could instead register simply ibm.cc, for example. Actually, the biggest rip-off I've seen is run by the registry for the Republic of Moldova (.md), which tries to sell domain names at very high prices to physicians. * * * * *
Well, that may have been true before, but $45 is no longer cheap. I can register a .com, .net, or .org domain at joker.com for about $11/year now, which is about half of what an .nu domain costs. You should probably register your domain as a .com domain and drop the .nu domain. The potential danger with .nu is that they charge a "nominal" $10 to change the technical contact or the nameserver information, which is a practice I'd never heard of. You'll find that leaves you wide open to unexpected charges in some situations. Most web hosting companies, for example, require that you list them as the technical contact on your domain record. If you change web hosting companies, which some people do frequently until they settle on a good one, you get charged at least $10 each time. Also, the reason that most web hosts insist on being technical contact is so that they can change nameserver information themselves any time they want to. Most don't do that frequently, but some do it often enough that your .nu domain name could end up costing quite a bit more than you thought it would.
|
|
wpoison
Search [tips]
|
Tuesday, 27 June 2000 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] Dumb headlines I see in the morning newspaper: "High Prices Barring Poor from Home Ownership" Well, duh. High prices are barring me from Rolls-Royce and Learjet ownership, too. It's just so unfair. RoadRunner sucks dead rodents through a tubular suction-oriented drinking device. Yesterday mid-afternoon, their DNS server died. The IP connectivity was still fine, as I proved by hitting my site by IP number. Then, half an hour or so later, IP connectivity died, too. I called RoadRunner tech support, who told me that Triad RoadRunner was having problems. Duh. Yes, I know. No idea what the problem was. No idea when it'd be fixed. As it turned out, it came back to life mid-evening, after being out for about six hours. This is not an impressive record. I've had RoadRunner for about ten days, and they've had three major outages during that time. These folks need to study what it means (or should mean) to be a utility. To me, at least, something classified as a utility should be always-on. Natural gas is a utility, as is water, sewer service, power, and telephones. All of those are always-on, barring unusual occurrences such as major storms, someone digging a trench where they shouldn't, and so on. Cable television service (and now cable modem service) is an order of magnitude less reliable. These folks need to concentrate on delivering uninterrupted service if they expect to be taken seriously. I have a bit of experience with networks, so I can say with some assurance that the problem is likely money. It costs money, a lot of it, to provide reliability and redundancy. It costs money for redundant, high-availability servers, redundant backbone links, hot spares, on-site technical staff, and so on. Most of which, I suspect, has not been implemented because of cost. But perhaps I'm being unfair to the cable companies. Perhaps they've done formal cost-benefit studies and found that people really would rather pay $40/month for unreliable service than, say, $50/month for reliable service. But I doubt it. So at this point I'm very seriously considering dropping RoadRunner and going back to using dial-up. For all the problems I had with BellSouth immediately before RoadRunner was installed, BellSouth provided reliable service for months on end between major glitches. When I asked the very nice lady at RoadRunner tech support if what I'd experienced in my first ten days' service was typical of their service level, she was horrified. "No," she said, "we've been having quite a few problems with Greensboro/Triad RoadRunner over the last couple of weeks, but I don't usually get many problem calls from people in your area." Perhaps she was telling the truth, so I guess I'll give it a bit longer. I've used only Crucial memory for years, but Pournelle has always been a big fan of Kingston memory. I built a few systems with Kingston memory years ago, but for some reason I hadn't tried any Kingston memory for many years. So Pournelle convinced me, and I asked Kingston to send me a few samples. I must say that I was impressed with how nicely they're packaged.
These are from Kingston's ValueRAM line, which is designed to provide a good alternative to "white box" commodity memory, on the theory that you pay very little more for ValueRAM, but get immensely better quality. Kingston direct-sources the chips themselves from top-name manufacturers (including Micron/Crucial), assembles the modules and tests each of them fully. Certainly a superior alternative to commodity memory, which often uses brokered/gray market chips slapped together with little or no testing. I have a couple of older no-name motherboards that are extremely picky about what memory they'll use, probably because they deliver marginal voltage to the memory sockets, or perhaps their BIOSs were written by a gang of drunken sailors. Over the years, I've tested many different memory modules with these motherboards. I've never had any problem with Crucial DIMMs, but commodity DIMMs frequently fail entirely or generate sporadic memory errors. I plan to install these Kingston DIMMs in those boards to see how stable the Kingston DIMMs are. If they pass, which I suspect they will, the Kingston DIMMs go into several long-term project systems I'm building for the book Pournelle and I are doing. Stay tuned. * * * * *
Thanks for the kind words. As far as testing the ability of applicants to parallel park, what a great idea. No, on second thought, if they required parallel parking to pass the test, no woman would ever pass. Wait, I'd better not say that... * * * * *
That is indeed interesting, if it's legitimate. My first thought upon seeing something like this is to wonder if it's a clever front for spammers looking to get valid email addresses. That's probably not the case, but having been burned in the past, I'm more cautious now. I confess that years ago when I got my first spam that had removal instructions at the bottom I foolishly sent email asking them to remove my name. It never occurred to me that someone would be slimy enough to use removal messages to compile a list of valid email addresses. Nowadays, of course, nearly everyone knows that it's a sucker bet to respond to removal instructions, so the spammers are getting more clever about validating email addresses. For example, I'm getting an increasing number of spam messages that have "return receipt requested" in the headers, directed to some doubtless temporary mailbox, which is why I've set Outlook not to honor receipt requests. * * * * *
Well, yes, but the domain you choose for your web site has nothing whatever to do with how you show up in the web logs when you visit my site. For example, my domain is a dot-com, but until recently I was accessing your site from my BellSouth.net dialup connection. I showed up in your web stats as a hit from a dot-net rather than a dot-com, because I was accessing you on a dot-net connection. Now I'm using a connection from triad-rr.com, so I'll show up in your stats as a dot-com visitor. I suspect there are very few people accessing my site with a dot-it, dot-fr, or whatever connection who are not actually located in Italy or France or wherever. * * * * *
Thanks. No, at this point I'm still stubbing-out the structure. There are still lots of pages missing. Heck, there are entire sections missing. I'm using Include Pages for stuff like the right- and left-column menus, so fixing them is easy enough. All I have to do is fix the one include page, and that fixes all occurrences. I haven't gotten around to viewing the pages in Navigator but I will. Thanks for the suggestions. * * * * *
Thanks. I haven't looked at the patents, but I may do if I get a spare moment. I don't think the patents themselves are the root issue, but rather the source of the ideas upon which those patents are based. My understanding is that this whole mess started at a JEDEC conference at which Rambus was a participant. In theory, a bunch of memory manufacturers got together to share ideas, with the intent of developing a standard that would incorporate those shared ideas and be freely usable by all. Supposedly, Rambus turned around and patented stuff based on information shared by others at that conference. I don't know whether or not that's true. For all I know, Rambus is in the right here and everyone else is lying. But I suspect not. * * * * *
Thanks. Do me a favor and tell Svenson that, will you?
|
|
wpoison
Search [tips]
|
Wednesday, 28 June 2000 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] The Register reports on the most egregious case of domain-name squatting I've heard about. A Belgian doctor has apparently registered 100,000 domain names since last August, spending $2,000,000 to do so. He's used his own money and that of investors in his little scheme, and now expects to sell the domain names for $1,500,000,000, or 750 times what he's paid for them. Since NSI has a policy against cyber-squatting in effect, and has had for the entire time this person has been registering these domains, it seems to me that there's an obvious solution here. Take the domain names away from him, all 100,000, and release them to be re-registered by other people. Doing that would turn these folks' expected 75,000% profit into a 100% loss, and send a strong message to other would-be cybersquatters. I don't object to people having multiple domain names, you understand. I have half a dozen myself and I know quite a few people who have registered multiple domain names. But there's a huge difference between registering multiple domain names for your own use and registering domain names with no intention of using them yourself, but simply squatting on them until you can sell them at a profit to someone who actually intends to use them. The latter practice needs to be wiped out, and the best way to do that is to see to it that people who attempt to do that simply lose the domain name(s) without any recompense. Bo Leuf reports that he's tested the mailexpire.com service and it appears to work as advertised. He created an account with a 12-hour expiration, and it expired as advertised. He also likes the control panel, which allows you to change settings on the fly so, for example, if you start getting swamped with spam to one of your aliases, you can go in and kill it instantly. Obviously, we have no idea whether mailexpire.com honors your privacy as they claim, but it's beginning to look as though this service works as claimed. Tom Syroid sends an article that reports that Microsoft Internet Explorer now has more than 86% of the browser market, with Netscape Navigator at 13.9% and dropping, and all other browsers having a 0.02% share. Those numbers sound realistic to me. My own web logs show that about 75% of my hits come from IE. Given that the mix of people who visit my site is somewhat skewed toward people who are more likely than average to use Navigator (people running Linux and so on), an 86% market share sounds reasonable. If Microsoft had any sense, they'd port IE to Linux. It wouldn't hurt Windows sales noticeably, and it would give Linux users a decent browser to substitute for the obsolete piece of garbage they're stuck with now. * * * * *
Okay, change "no woman" to "nearly no women", but I must say that you've let down the side here. In case you haven't read the news about the Human Genome Mapping Project, they located the "parallel parking" gene, which was (as expected) on the Y chromosome. In fairness, I should also say that they located the "asking for directions when lost" gene, and found that it was on all X chromosomes, but only on a tiny percentage of Y chromosomes. That gene is recessive, requiring reinforcement from the other chromosome, so we men now have a solid genetic excuse for refusing to ask for directions when lost. * * * * *
Funny you should mention that. I almost bought a used Rolls-Royce twenty years ago or more. It was about 20 years old then, and the asking price was something like $15,000. I could have afforded that, just about, but one of my friends pointed out that insurance and repairs would kill me. As I recall, his exact words were, "So are you going to get a bank loan when you need to change the spark plugs or oil filter?" So I decided not to do it. After I got your message, I spent some time on the net looking for used Rolls-Royces. I found a couple of very nice 15-year old models in the $35,000 to $50,000 range, but nothing anywhere near the prices you mention. Still, I might be inclined to go for a 15-year old Rolls-Royce the next time I'm considering plunking down $35,000 for a new 4X4. * * * * *
Ah, but they don't write songs about drunken marines, soldiers, or aircrew. Just search the Internet for the strings "drunken sailor" and "early in the morning" and you'll see what I mean. As far as driving, I think things have changed. Twenty and thirty years ago, women really were safer drivers than men. When I got my driver's license at 16, I drove like all 16-year old boys did, which is to say fast and aggressively. At that time, 30 years ago, 16-year old girls almost without exception drove sedately, doing just as they'd been taught. It was the boys who were the real hazards. The same thing was generally true of adult drivers. Very few women drove fast or aggressively. Things have changed. Nowadays, I regularly see young girls driving like maniacs, and I've been cut off at least as often by women drivers as by men. I'm not sure what's happened to change things, but something has. * * * * *
Thanks for taking the time to write such a lucid explanation of the situation. Your final paragraph reminds me of Mae West's famous retort while she was testifying at her obscenity trial. Judge: "Miss West, are you trying to show contempt for this court?" Mae West:: "On the contrary, your honor, I was doing my best to conceal it."
|
|
wpoison
Search [tips]
|
Thursday, 29 June 2000 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] I've been a busy hardwareguy lately. There's quite a lot to report today, and not much time to do it. Kingston ValueRAM memory works flawlessly. For testing purposes, I keep a couple of old motherboards around that are very picky about what memory they'll use. I run them with marginal power supplies, which makes an even tougher test for memory, although unfortunately those motherboards and power supplies are representative of a lot of systems out there in the real world. With top-notch memory, the systems run without error. With nearly all "white box" commodity memory I've tried with those boards, I get system lockups, memory errors, and so on. I tested the Kingston ValueRAM memory fairly extensively, both by simply booting the system repeatedly (which is when errors attributable to memory often show up) and by running various memory diagnostics. The Kingston ValueRAM memory passed with flying colors. Highly recommended. The Intel D810EMO Desktop Board--also called the MO810E--looks like a winner. The D810EMO is designed to fill a very specific niche--entry-level, ease-of-use systems--and it does so perfectly. This isn't a motherboard that the power user or the serious gamer will choose, but it appears ideal for systems targeted at casual and home PC users. If we were building a system for Barbara's sister, this is the board we'd use. At an expected street price of $160 or so, this board costs a bit more than some, but it is a true fully-integrated single-board solution. It incorporates Intel 810E video (which is more than good enough for nearly any casual user), Creative Labs SoundBlaster Audio PCI/128 sound, and an on-board Intel 82559 10/100 Ethernet adapter for connecting to a home network or to an ADSL or cable-modem Internet link. As you might expect with a motherboard that uses the tiny FlexATX form factor, expansion capabilities are limited. The board provides just one DIMM socket, which accepts a standard non-ECC, unbuffered, PC100 SDRAM DIMM of 32 MB to 256 MB. A second DIMM socket would have been nice, but a single DIMM socket is not a serious drawback for the market for which this board is intended. Similarly, the board provides just one PCI slot. Intel says this slot is primarily intended for an internal modem, but it might equally well be used to install a PCI video card to replace the on-board video. Again, a second PCI slot would have been nice, but the high degree of integration of the D810EMO and the intended market of FlexATX systems mean that its absence is not a serious drawback. Limited upgradability is a reasonable compromise for this class of system. Probably not one in 10,000 FlexATX systems will ever be upgraded. The remaining 99.99% will simply be purchased and used until they drop. This board is the first in what will probably become a flood of "legacy-reduced" motherboards. In this case, legacy-reduced means there's no floppy disk controller, no parallel port, no PS/2 keyboard or mouse port, no game port, and only one serial port. This board connects to external peripherals using USB, and it provides four USB ports to do so: the now-traditional two on the back panel and two more on a header designed to be extended to the front panel. Because multiple devices can be connected to each USB port, adding external devices should not be a problem. Internally, the board includes the standard two ATA interfaces, but with a twist. The Primary IDE Interface uses the standard 40-pin ATA connector. The Secondary IDE Interface is a bit different. It uses the so-called "mobile" IDE connector, which is a 50-pin superset of the standard 40-pin connector. The 50-pin connector, which is actually physically smaller than a standard IDE connector, uses a high-density pinout that includes the standard 40 IDE signal pins, but also incorporates power and audio pins. The only devices that use IDE-50 connectors are those intended for notebook computers and other mobile applications. Intel supplies an IDE-50 adapter cable that has an IDE-50 connector on one end and two IDE-40 connectors for connecting standard drives. Although that cable works well enough, the IDE-50 connector is a very loose slip-fit to the motherboard connector, and it appears that it could easily come loose. For testing purposes, we built a system around the D810EMO motherboard that included the following components:
Like all FlexATX motherboards, the D810EMO is designed to use an SFX power supply rather than a standard ATX. Fortunately, Intel includes an SFX-to-ATX adapter cable in the box. The system came up first time, with no problems. As a matter of fact, it came up better than expected, because Windows 98 SE was already installed on the hard disk. Ordinarily, we strip disks down to bare metal before putting them back on the shelf, but obviously we forgot to do that this time. We haven't had time to do any detailed testing on the system yet, but we expect its performance to be indistinguishable from the Intel CA810EAL board which we use in another of our test-bed systems. Stay tuned. Although it's certainly not what Intel designed this motherboard for, we can think of another application for which it may be ideal. Building micro-servers. We work surrounded by more than a dozen traditional systems which perform various functions. They're big, consume lots of power, generate lots of heat, and their fans produce an incredible amount of noise. FlexATX systems are small, consume little power, generate little heat, and use SFX power supplies, whose fans are nearly inaudible. A D810EMO might be the ideal foundation for a dedicated micro-server, although its dependence on USB rules out NT4. But that leaves Windows 2000 and eventually Linux. Pop in a Celeron or low-end Pentium III processor, a 128 MB DIMM, a 20 GB IDE hard disk, an ATAPI CD-ROM drive, and perhaps a tape drive, and you have yourself a small, inexpensive, quiet dedicated server that uses little more power than a light bulb. The only drawback we can see to the D810EMO motherboard isn't really the board's fault at all. FlexATX cases are still very thin on the ground, and may remain so for some time. For now at least, case manufacturers obviously consider FlexATX cases to be an OEM product rather than a consumer product. We think that'll change. When it does, the D810EMO desktop board will be an excellent choice for building an "appliance PC". I finally took real advantage of the cable modem's speed this morning. With the 31.6 dialup connection, I put a more-or-less arbitrary limit of 100 MB on the size of files I would attempt to download. I know it sounds ridiculous that I'd download files that large over a 31.6 dialup line, but I've frequently done so without problems. The last really big file I downloaded over my dial-up connection was Ziff-Davis's 3D WinBench 2000, at 92,476,508 bytes. These downloads took place in the background, and the only noticeable impact was that it was a bit slower to check mail or load a web page. I often started the download immediately before going to bed, and would find it complete in the morning. Really huge downloads would still be in progress the following morning, but that didn't happen often enough to worry about. Actually, I wouldn't have put any limit on file size at all, except that BellSouth has a 12-hour timeout on connections. It doesn't matter what you're doing, when that timer ticks over to 12:00:00 call duration, it drops the connection. Depending on conditions, I get anything from 9 MB to 12 MB/hour throughput on the dialup, which means I could download somewhere between 108 MB and 144 MB during a 12-hour call. For safety's sake, I therefore decided to put a 100 MB limit on downloads Actually, I'd occasionally been successful at spanning a large download over two calls, but that doesn't always work and it's too depressing to get 100 MB into a big download and find that you have to start over. I wanted to download the full Ziff-Davis WinBench suite, which is 121,432,113 bytes. Expecting that my cable modem would make that easy, I hit the ZD site a couple of days ago and started the download. Unfortunately, the throughput was very low, about 3.5 KB/s, which is to say about the same as dial-up throughput. I killed that one and checked a couple other sites. They were working normally, so the problem was obviously the ZD server or its connection. I tried again later and got the same results. This time, I decided to let the download continue, with the hope that the slow throughput was temporary and would pick up as the download progressed. That wasn't to be. When I came back later, IE had a message up on the screen telling me the download was complete. Trouble was, it showed the total file size as about 30 MB rather than 120 MB. I tried several more times over the next couple of days, with similar results. IE always showed the download as complete, but it was always only 20 to 30 MB in size. Finally, at 0630 this morning, I decided to try using my FTP client to grab the file. The counter started clicking away at a satisfyingly high rate, eventually stabilizing around 80 KB/s. Twenty minutes or so later, the download completed, and this time it was the full 121,432,113 bytes. I should hasten to add that IE wasn't the problem. The ZD server was simply slow when I was hitting it with IE. I guess the moral is that you have to pick your times carefully if you expect to do huge downloads successfully. Maybe I'll try downloading the ~200 MB service pack for Windows 2000 when it's out of beta. 11:40: Boy, here's a lesson learned. I'm working on the CD-R(W) chapter at the moment. I needed to install Adaptec Easy CD Creator on thor (the Windows 98SE test box), so I fished out the CD and started the install. When the CD autoran, it displayed the Plextor software setup menu, offering to install Easy CD, DirectCD, or the Plextor utilities. As it happens, I have a Smart & Friendly CD-RW drive in this system, but I figured software is software so I started the install. Big mistake, as it turned out. Setup proceeded apparently normally. I restarted the system and fired up Easy CD. The system completely locked up with an hourglass showing. I restarted the system and played around with it a bit. Everything worked normally, so I decided that perhaps the first crash was a coincidence. I fired up Easy CD again. Bigger mistake. This time the system blue screened immediately. Oh, well, I thought, I just need to install the S&F version of Easy CD. Either that, or the commercial copy of Easy CD Creator Deluxe that's floating around somewhere here. So I restarted the system, but this time it blue screened at boot. Not good. I eventually concluded that Windows 98 was entirely trashed, and restored an image file to get me back to where I started. The moral here is that one shouldn't attempt to install an OEM copy of a software product with hardware of another brand. I've actually been taught that lesson before, but for some reason it never sinks in. And in all fairness to Plextor, S&F, and Adaptec, this problem may not be due to mixing and matching software and hardware. This system also has a Hitachi DVD-RAM drive on it, and I've installed the WriteDVD! software. That software includes a UDF driver to allow the DVD-RAM disc to appear to the system as just another disk volume, and it's quite possible that there was a conflict between the UDF driver and Adaptec Easy-CD.
|
|
wpoison
Search [tips]
|
Friday, 30 June 2000 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] Barbara was watching the dogs play last night and came in to grab the digital camera to shoot some pictures. Here's Malcolm (aka Demon Puppy or Damien) playing with one of his rope-balls, which are among his favorite toys.
Malcolm has a hard time playing nowadays. He loves to tug, but Duncan won't let him tug with us, probably because he regards it as too aggressive. But Duncan does play, too. The next picture shows some of the anti-puppy measures we've had to put into effect. All of the cushions on the sofa and love seat are old ones that we saved for just such an occasion. And luckily so. Malcolm is fully pillow trained. He steals pillows off the sofa and carries them around the house. As far as he's concerned, they're just big soft-toys. On occasion, he's actually removed the sofa cushions themselves and started batting them around. Also, as he's now reached puberty he's become interested in girls. Unfortunately, he frequently mistakes pillows for girls and attempts to have his way with them. Just the other day, I was taking a nap back in the bedroom when I was rudely awakened by Malcolm fanging onto my pillow, yanking it out from under my head, and trying to carry it off for nefarious purposes. He is one seriously confused puppy.
The rug on the floor is one that Barbara picked up for $40 or so at WalMart. It's there for our safety, because when Malcolm went on a tear in pre-rug days, those Ottomans went sliding literally all over the room as he leapt from sofa to Ottoman to love seat to sofa to Ottoman, and so forth. And, let me tell you, one of those Ottomans sliding across the floor at a high rate of speed was no joke. Knowing what would happen, Barbara bought a cheap rug. That was a wise choice, because Malcolm has already eaten a fair amount of it, as evidenced by the disgusting blue stringy results during his daily constitutionals. The storage areas under the end tables are blocked because Malcolm, like all Border Collie pups we've had, is a bookworm. The Coke cans are another anti-puppy measure. Ordinarily, I treat 12 ounce cans of Cokes with the contempt they deserve. I prefer the convenient single-serving 3-litre size. But I must admit that the cans are useful. Barbara puts a few pennies in each one and we use them as booby-traps. Some puppies, including Duncan, have been afraid of the noise the cans make when they fall. Not Malcolm. But at least the cans give us an audible signal that he's into something he shouldn't be. Last night, I was working in my office and Barbara in hers when we heard the cans falling over. I went running out to find Malcolm standing with his front paws on the arm of the sofa, pillaging my end table. It must have been thirsty work, because he'd also drunk about half the glass of water I'd left sitting there. It's been nine months or so since I last built myself a new main system, and the first item on the agenda for today is designing a new personal system. I know nine months doesn't sound like long, but when one writes PC hardware books, one must keep up. Although I'm an advocate of dual processors and SCSI, I recognize that most of the world uses single-processor systems and ATA, so I've decided to build the best single-processor IDE-based system I can. More Monday on what I've chosen, but I'll say now that there'll be some interesting stuff. I'll be busy this weekend, so the posts, if any, will be short. Thanks to everyone who sent email to tell me about the slashdot review of Windows NT TCP/IP Network Administration. I visited the page yesterday just to see what they'd said. It was a favorable review. Several people commented to me that the review might help the book climb in the Amazon.com rankings, but that doesn't appear to be the case. It was sitting at about 3,000 when I looked this morning. Its usual place lately is in the 2,000 to 5,000 range, with occasional peaks into the 1,XXX range or a bit higher and occasional drops into the 6,XXX to 8,XXX range. I suspect the slashdot review won't have much benefit, because many slashdot readers wouldn't be caught dead reading a Windows NT book. * * * * *
Ah, but you don't have to picture it. I've done that for you. There are pictures of my office over on my images page. Granted, those are from last fall sometime, but it looks much the same now. * * * * *
No, registering variants of your own name or your business name doesn't count as domain-name squatting. For example, Barbara (whose maiden name is Fritchman) has registered fritchman.com, fritchman.net, and fritchman.org. Similarly, Bo Leuf has leuf.com, leuf.net, and leuf.org. As far as how you discriminate, I'd say there are several easy ways. First, one can assume that anyone who registers 100,000 domain names has not done so intending to use all of them. Second, registering multiple domain names that are not related to your own name or your business name and then not using them indicates that you are hoarding domain names. Third, offering domain names for sale clearly indicates that the intention was to profit from reselling the domain name. As far as I'm concerned, the simple action of offering a domain name for sale should be grounds for immediate cancellation of that domain name. If someone comes to Barbara or Bo and offers to buy one of their domain names, fine. Let the two parties come to an agreement. But if Barbara or Bo advertises their domain names for sale, take them away from them. As far as IE versus Nav on Linux, anyone who would choose Nav rather than IE is clearly doing so for irrational reasons, assuming that IE were as well implemented on Linux as on Windows. IE is now at least two, and probably three, major revisions ahead of Nav. Nav is a dead product, and their so-called version 6 looks to me to be about equivalent to IE4 in features, performance, and functionality. If Nav 6 is the best they can do, Nav is in sad shape. Good point about using the D810EMO for business purposes, particularly for stuff like heads-down data entry, office productivity applications, and so on. I suspect we'll soon see a flood of FlexATX based systems for such purposes. They'll be relatively inexpensive, small, and quiet. * * * * *
The short answer is that my regular FTP client does everything I need to do, including resuming downloads. The longer answer is that GetRight is one of those products that incorporates what Steve Gibson calls Spyware. To verify that, I went over to the Aureate/Radiate web site and sure enough, both GetRight and My Getright are listed as products that incorporate the Aureate/Radiate spyware. To me, it is entirely unacceptable for software to incorporate such spyware functions, particularly if it is done surreptitiously, and still the more when uninstalling the original software leaves the spyware on your hard disk and still working. If you haven't read what Mr. Gibson has to say on the subject, I'd strongly encourage you to do so. Even if you believe that Aureate/Radiate are being completely up-front about what their product does, which I think is a risky assumption to make, having their DLL installed on your system opens some frightening security vulnerabilities. In short, I'd rate GetRight as unacceptable no matter what its features because it incorporates spyware. Even if the registered version does not include the Aureate/Radiate software, I'd still rate the software unacceptable, because I won't deal with companies that engage in such questionable practices. I'd encourage you to download the OptOut program from Gibson Research and run it to see what's lurking on your hard disk. * * * * *
You're asking the wrong guy. I saw a Mac once, but that's about it. You're correct that the CTT-8000 drive was made by Conner, which was subsequently purchased by Seagate. In fact, I have a CTT-8000 tape drive, and it's a very good drive. I spent a few minutes checking the Seagate web site, but couldn't find any documents that'd do you any good. At this point, I'd suggest that you send email to both Seagate and Dantz tech support describing your problem and asking for suggestions. There may be an easy solution. If so, they're the ones that should know about it. Also, I have more than a few readers who use Macs, so perhaps one of them will have a suggestion. * * * * *
I suspect you're referring to the MTU setting. That setting is indeed important to Windows 95/98, but the Windows NT TCP/IP stack is considerably more sophisticated than that of W9X, and needs little or no tweaking. I'm surprised you saw any change with Windows 2000. Although I haven't used W2K much, I'd have thought that it would be similar to Windows NT 4 in that respect. As far as speed, the few times I've checked, I typically get between 75 KB/s and 150 KB/s when downloading from ftp sites, and the various net "speedometers" say that my connection is anything from 100 KB/s to 250 KB/s.
|
|
wpoison
Search [tips]
|
Saturday, 1 July 2000 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] Another new month. I was just roughing out my schedule, and came up against a seemingly intractable problem. Assuming a 12 hour/day schedule, I need to allocate about half a day a week to running this site, four days per week to writing the big hardware book, another four days per week to building the hardwareguys.com web site, and two days per week to building systems, testing new hardware, and so on. That totals 126 hours per week, which leaves me only six hours per day for eating, sleeping, reading, and all other activities. This just isn't gonna work. Perhaps I should pull a Syroid and just disappear for a couple of months. Some time ago, I sent email to Network Solutions, Inc. asking them to remove my information from the list they sell to spammers. I sent that message to privacy@networksolutions.com, the address they gave for such requests. Here's the response I got, which doesn't need further comment:
* * * * *
I suspect you're using Windows 9X rather than NT. Perhaps you should upgrade to NT4 or even Windows 2000. Both are an order of magnitude more stable than Windows 9X. Navigator crashes constantly for me. Even if it didn't, it's pathetically slow and renders pages poorly. I crash IE5 once in a great while, but every time that's happened, it's been when I had a lot of IE5 windows open, usually 30 or more. Most people don't use a browser that way, and I suspect they'd find that IE5 very seldom crashes. Also, under NT4, when IE5 crashes, all that crashes is that one instance. All other programs, as well as all other instances of IE5, continue running normally. * * * * *
Glad you got it worked out. The Connor/Seagate CTT-8000 is a very nice Travan TR-4 drive. I'm glad you mentioned it, because I'd forgotten that I had one sitting here in an old Pentium/133 tower system. I may pull it and use it in something else. 4/8 GB is still a respectable tape drive. * * * * *
What concerns me even more than the spyware issue per se is that Gibson reports that having the Aureate/Radiate DLL running on your system opens a hideous security hole. As Gibson says, it hasn't been exploited yet as far as he knows, but who knows when that will change? * * * * *
There are actually a lot of reasonable people over on slashdot (witness some of the temperate responses in the thread on that book review), but they're outshouted by the idiots. I must say that I've never been able to figure out why anyone would treat the OS someone chooses to use as a religious issue. I run NT4 because it happens to have the best mix of features for what I do. If Linux were better (or became better), I'd change to it in a heartbeat. Nor do I treat Windows 9X users with contempt, as some NT users do. I recognize that Windows 9X is superior to NT for some purposes, notably games. I don't play games, so that's not a factor for me, but I recognize that people who do play games choose Windows 9X rather than NT4 for good reasons. * * * * *
That's terrible. You have my sympathy. I'm surprised that Coke has anything to do with kidney stones. I mean, it's just fizzy water with some flavoring. I could see it if they'd told you to knock off the bourbon or something, but Coke seems pretty innocuous. * * * * *
Your best bet is probably to buy a reconditioned unit directly from the manufacturer. I know that both Plextor and Smart & Friendly sell such units, often at a significantly lower price than new ones, sometimes less than $100. The reconditioned units are probably about as good as new units in terms of function and reliability, although they may have a relatively short warranty. Reconditioned units may have been returned under warranty, in which case the maker repairs them, brings them up to standard, and sells them at a deep discount. They may also be returned units which had no malfunction at all, such as units whose boxes were damaged in shipping, units that were purchased and then soon returned for one reason or another, and so on.
|
|
wpoison
Search [tips]
|
[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] I spent most of yesterday working on the HardwareGuys.com web site. It's starting to come together, but there's a long way to go. One problem I'm encountering is Netscape's poor rendering abilities. I'm using styles to generate thin rules to separate sections, and those rules don't display in Navigator. On this site, I've attempted to keep things simple enough that pages display readably regardless of which browser the reader is using. But I'm not sure that I'll worry about that on the new site. I'll try to make sure that the pages are readable with Navigator, yes, but I won't make any special efforts to accommodate the least common denominator rendering problems. Netscape is down to about 14% share right now, and seems to be losing share at about one percentage point a month. How long can it be until Navigator becomes a non-issue? The feedback I get from the readers of this site who use Netscape tells me that nearly all of them use Netscape rather than IE for one or both of two reasons: (a) they don't like Microsoft, (b) they're running older hardware or an operating system, like Linux or earlier versions of Windows, for which IE is either unavailable or not optimized. Almost no one tells me they use Netscape because they think it's technically superior to IE, and frankly I regard those who do as likely being insane under the McNaughten rules. Most Linux users would change to IE or Opera in a heartbeat. I'm told that Opera will soon be available for Linux, and I wouldn't be at all surprised to see Microsoft release a free version of IE for Linux, on the theory that it might do them a lot of good legally while not having any noticeable adverse impact on Windows sales. All that tells me that Navigator will soon disappear as a consideration in designing web pages, if it hasn't already. I'm sure I'll get messages telling me that I'm assisting the Borg in its quest for world domination, and they may even be right. But that's not why I'm doing it. I simply don't have time to spend accommodating the peculiarities of an aging and increasingly irrelevant browser. I remember years ago many webmasters spent a lot of time creating versions of their pages that would display properly in both IE and Nav, or even creating two parallel sets of pages to accommodate users who chose either browser. I wonder how many do that now. Not many, I suspect. Perhaps Netscape 6 renders pages that use styles correctly. I'll have to download it and install it on a scratch machine to find out. If so, I won't be abandoning Navigator itself, but only the creaking Nav 4, which was overdue for replacement years ago. * * * * *
I'm not sure what would be involved in optimizing IE for Windows 98 versus NT/2K or why it should be necessary. They're both Win32 platforms, right? Our experiences differ. Admittedly, I don't run Win9X much, but I've found IE much more stable on Win9X than is Navigator. To me, at least, Navigator seems just as slow and just as crash-prone on Win98 as on NT4. I haven't tried Opera since version 3.6X. I wasn't impressed with it then, although I periodically get messages from people telling me how wonderful it is. Every time I look at Opera, it seems to me that IE has it beat all hollow. I understand that version 4 of Opera is or soon will be released, so perhaps I'll look at it. * * * * *
Dunno. If so, that's the first I've heard of it. I've been over to their site a few times in the last couple of months, and saw the message you mention. I just assumed that it had to do with the fact that they did a complete overhaul of their web site recently. * * * * *
Apple and the Linux bigots remind me of our nine-month old male Border Collie pup. Malcolm feels threatened by the alpha males around the house--me and our 5-year old male BC, Duncan. Sometimes when Duncan or I approach Malcolm, even with kind intentions, he growls and snarls. He's afraid of us (senior males are a threat to junior males), so he puts on a threat display--"You can kill and eat me if you want to, but I'll go down fighting." Apple and the Linux bigots react in a lot the same way to Microsoft as the alpha male. And the way that Microsoft reacts to criticism in comparison to Apple and the Linux bigots is also instructive. Microsoft are above all professional in their relationship with the press. Say I write an article that appears on the front page of the New York Times under the headline, "Microsoft is the Antichrist" and in which I rip Microsoft up one side and down the other. The next day, I find that I need an evaluation copy of Windows 2000 Professional. I telephone WaggEd (Microsoft's main PR agency), who are obviously fully aware of the article I've just written. They treat me as though I'm Bill Gates' best friend, and FedEx me a copy of W2KP, with a nice hand-written note thanking me for my interest. Say the next day I write an article about Apple which appears on the front page of the New York Time under the headline, "Apple is God" which goes on for pages and pages about how wonderful their new model is. I conclude the article by saying, "This new model from Apple is just about perfect. The only thing I'd change is to make the case a little darker beige." That's enough to get me added to Apple's "Enemies List". I get flooded with email asking how I dare criticize the new Apple model and demanding that I publish a retraction. If I call up Apple to ask a technical question, there's no one available to talk to me. Okay, that's a bit of an exaggeration, but not much. What Apple and Linux radicals have never been able to learn is that they're not doing themselves any favors by attacking their friends. As far as networking Apple, Linux, and Microsoft, I'm sure there are solutions out there, but I'm not aware of any Linux-centric products. Perhaps my readers will know. I do know that Windows NT Server (and presumably Windows 2000 Server) supports Mac clients via Services for Macintosh, which permits Mac clients to log on to an NT network and share files and printers. I can't speak to the functionality of Services for Macintosh, because I've never set up a network with Mac clients. * * * * *
Yep, I agree with that. * * * * *
Ugh. Well, they *do* make caffeine-free Coke (shudder).
[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] |