TTG Home Robert Home Daynotes Journal Home Journal for Week of 30 July 2001

photo-rbt.jpg (2942 bytes)Daynotes Journal

Week of 30 July 2001

Latest Update: Friday, 05 July 2002 09:16
 

Search Site [tips]


Click Here to Subscribe

Visit Barbara's Journal Page

Monday, 30 July 2001

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Messageboard]  [HardwareGuys.com Messageboard]

Things to do this morning. More this afternoon or this evening...

I've been hacking on the Game Controllers chapter all day long, but it's finally starting to take shape. Part of the problem is that none of this stuff works the way it's supposed to, particularly USB. I remember the first time I heard the phrase "Plug-'N-Play". I immediately turned to my friend and said, "Prug-'N-Pray, more like." And I've not seen much to change that opinion. Oh, PnP usually works pretty well, but the promise of USB as a PnP interface has not been fulfilled. USB is better than it was in the early days, but there are still way too many conflicts and incompatibilities. I wish Intel and Microsoft had never pushed USB. Instead, they should have pushed SCSI hard. The cost of silicon always drops precipitously when high-volume production starts, and I suspect SCSI would have been no exception. Granted, it's parallel rather than serial, with all the cost disadvantages that implies, but at least SCSI always works like it's supposed to.

The weather liars are now saying that we can expect clear skies from 18:00 on, so Barbara and I may head up to Bullington this evening and do a little observing. Luna is up and big, so it'll wash out a lot of the deep sky stuff, but I could look at Luna for years and not get bored. Thirty years ago, I was foolish enough to hope that the landings were just the first step and that there would likely be a Lunar Colony by now. I'd have made it there somehow, too. Alas, that was not to be.

I think I'm going to become sporadic in my postings, just like Pournelle and others. For three years or so, I've posted nearly every morning, and usually about the same time. I did that to force myself to write something every day, but I think there is no longer a need for that as a motivator. I keep FrontPage up on my Roadrunner box, so it's easy enough to publish at any time. So, from now on, I'll write and post updates whenever I feel like doing so. That means I may go a day or two without posting an update, particularly on weekends, but it also means I may post two or three updates in a day.

It'll be interesting to see what effect, if any, this has on my web access statistics. When I run mine and Pournelle's, I see that his hits are more evenly distributed throughout the day, presumably because he posts more or less at random so people check his site more frequently. My own tend to peak around 10:00 to 11:00 local time, which is suspiciously close to the typical 09:30 time that I used to post my updates. I'm betting my traffic will jump by 25% to 50%, not that I really care one way or the other. When I started out, I was lucky to get 100 page reads a day, and felt that I was writing mostly for myself. That number gradually climbed to where it generally sits now, which is something on the close order of 2,000 page reads a day.

 

Click here to read or post responses to this week's journal entries

Click here to read or post responses to the Linux Chronicles Forum

[Top]

Tuesday, 31 July 2001

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Messageboard]  [HardwareGuys.com Messageboard]

It was indeed mostly clear yesterday evening, so we headed for Bullington about 20:00, arriving just as Bonnie Richardson did. We got the scopes set up and then sat around and waited for them to cool down and for darkness to arrive. As we waited, the clouds moved in. By the time it was dark enough to see anything but the moon, we had 9/10 cloud cover. And even Luna spent most of its time peeking through a cloud. So we sat around talking until 23:00 and then drove home.

Our social life is very strange these days. We drive 25 miles one-way to sit in the dark on a decaying concrete pad in the middle of a tobacco field, chatting with friends. When the women need to use the bathroom, they go behind the barn (literally). Oh, well. Tonight is supposed to be clear.

Brian Bilbrey sends this link with his comment that it for some reason made him think of me. Read the brief remark and then scroll down to the picture, which appears to show a new hi-tech weapons system being tested on an old F4 Phantom. I've forwarded the link to Pournelle, who's probably known about it for years...

Another nominee for the Darwin Awards. Emmit Scott, 60, of Roanoke Rapids was growing marijuana in his back yard, apparently for pain relief. Someone was coming into his yard at night and stealing his marijuana plants. He told the guy to stop and threatened him with trouble otherwise. The guy didn't stop, so Mr. Scott called the police to report the man for stealing his marijuana plants. Mr. Scott was surprised when the police arrested him for growing marijuana. In a statement, Mr. Scott said he didn't think it was fair for the police to arrest him because he'd showed them his marijuana plants voluntarily. Duh.

I got a call yesterday afternoon from a good friend of mine who's an IT manager at a mid-sized local company. He wanted to talk about Linux, and the direction the conversation took gives me concern for the future of Linux in corporations. John announced that he needed to replace RedHat Linux and asked which distribution I'd recommend. The core problem was that Linux doesn't work properly with his Adaptec SCSI host adapter. John contacted the guy that maintains the Adaptec SCSI drivers for Linux, and that guy told him that there's a problem with the Linux kernel itself. Apparently, the workaround, such as it is, is to use the multiprocessor kernel. That apparently works, but provides pathetic performance.

At that point, I was confused. I asked John if the problem was with the kernel itself and he said it was. I asked if the problem was in the 2.4 kernel, and he said it was. So at that point, I couldn't figure out what the point was to changing from Red Hat to some other distribution. After all, if it's a kernel problem, changing distributions isn't going to solve anything. Well, it turns out that John's reason for wanting to change distributions had nothing to do with the SCSI problem, except indirectly.

In trying to solve the problem, he'd attempted to run Red Hat's automatic update, which downloads all files and patches that are more recent than what is currently installed. He'd used that service a month or so ago with no problem, but this time it informed him that he had to sign up for a paid maintenance agreement in order to access the service. Red Hat wanted $15 per month for that service, and John was irate. He allowed that he'd be willing to pay $15 per year, but he thought $15 per month was simply outrageous, given the low cost of buying the Red Hat distribution itself. His exact words were, "That's like buying a car for $20,000 and then finding out that the maintenance contract costs $60,000 per year."

Now, understand that John is used to spending money on software. His company runs PeopleSoft with hundreds of users, so John is used to signing off on purchase orders of a million dollars or more for software, and tens to hundreds of thousands of dollars for consultants. His company runs UNIX on heavy-iron HP servers, so he's used to signing off on maintenance contracts that cost tens to hundreds of thousands. Nor is John a pointy-haired manager. He's a techie, with a degree in computer science. He's been running Linux at home for several years, and is as comfortable in Linux as he is in Windows. He's forgotten more about configuring Cisco routers and building internetworks than I'll ever learn. He's a competent C programmer. If he wanted to take the time, I'm sure John has the skills necessary to write his own Adaptec SCSI driver for Linux. If anyone should be willing to support Open Source with his company's dollars, John is that person.

And yet John was outraged because Red Hat wanted to charge him $15/month for Linux updates. If John's attitude is representative--and the financial performance of Linux companies makes me think it may be--this can't be good news for Open Source in general and Linux in particular. 

I'm not having a good week. A mistake in PCHIAN, and now I somehow misidentify an F-16 as an F-4. Thanks to Jon Barrett, who says:

It's an F-16. Among other things, they weren't still making Phantoms (alas) in '84.

Phantoms Phorever!

Yeah, you're right. I didn't even look at the vertical stabilizer, which would have given it away just by its shape, let alone the "84" on there. I just saw that drooping tailplane behind the guy and assumed it was a Phantom. I should have noticed the flaperon, also. No wonder it makes Barbara nervous when I crank up the ZSU-23/4.

 

Click here to read or post responses to this week's journal entries

Click here to read or post responses to the Linux Chronicles Forum

[Top]

Wednesday, 1 August 2001

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Messageboard]  [HardwareGuys.com Messageboard]

Okay, this is fascinating. I may have done the weather-liars a disservice. In the past, I'd just looked up the weather on weather.com and weatherunderground.com using our zip code (27106) on the assumption that there wouldn't be any finer granularity in forecasts available. Who's going to forecast weather for Pilot Mountain, which is after all only about 20 miles NW of us as the crow flies? But just for the heck of it I decided to try both. As of yesterday afternoon, here's what the Weather Channel forecast for 27106 last night:

Tonight, Jul 31
Hour by Hour Forecast   more details...
3 PM 6 PM 9 PM 12 AM 3 AM 6 AM

81 F

78 F

73 F

69 F

67 F

68 F

And their forecast for 27041 (Pilot Mountain) last night:

Tonight, Jul 31
Hour by Hour Forecast   more details...
3 PM 6 PM 9 PM 12 AM 3 AM 6 AM

81 F

79 F

74 F

70 F

69 F

68 F

Not even close. 

We've noticed similar real disparities in weather several times. We'd leave Winston-Salem with clear skies and half an hour later arrive in Pilot Mountain with clouds moving in. We just assumed all this time that the clouds were also moving in in Winston-Salem in our absence. Apparently, that's not the case. Several times, we'd bag it up at Pilot Mountain because the cloud cover was 9/10 or 10/10 and drive back to Winston-Salem. When we arrived there, the cloud cover might be 2/10 or 3/10. We, of course, assumed that the skies had been clearing in WS during our drive and that if we'd stuck it out at Pilot Mountain the skies would have cleared up there as well.

I'm probably the last one to figure this out, but I'll post this just in case anyone else was making the same wrong assumption I was. Not that the weather can vary so dramatically in locations less than 20 miles apart. I knew that already. But I was assuming that forecasts were available only for larger towns, and that's obviously not the case. Now, granted, the forecast is actually for Mt. Airy rather than Pilot Mountain, but as anyone who's seen The Andy Griffith Show knows, Mayberry (Mt. Airy) is just a stone's throw from Mount Pilot (Pilot Mountain).

Late yesterday afternoon, they changed the forecast to rain all evening. We never got the rain, but it was indeed cloudy, so we stayed indoors and read.

The Code Red worm, which generated all kinds of media hysteria about the end of the world coming at 8:00 p.m., pretty much came to nothing. At worst, some people noticed a minor slow-down in net access. Cod Red even made the local newscast, which distributed some amazingly bad advice. They were obviously confused about the difference between a server running IIS and a desktop PC running Windows, so they probably scared a lot of people unnecessarily. 

Their advice on dealing with the worm was so bad it was laughable. They suggested, get this, that if you found yourself infected with the Code Red worm you could eradicate it by turning off your server to clear the worm from memory and then turning your server back on again! Wow. Why didn't we so-called experts think of that solution? As a result of that story, there are probably a lot of people now who think that they can protect their systems from viruses and worms simply by turning their computers off and then back on periodically. There are probably also more than a few sysadmins upset because users powered down servers that they had no business touching.

Steve Gibson, as usual, was frothing at the mouth. He's pretty much lost all credibility among knowledgeable security folks. His warnings generally have a small kernel of truth surrounded by massive amounts of hype. It's no wonder that few people take him seriously any more.

I mentioned not long ago that the music industry had started using a MacroVision kludge to prevent people from ripping CDs. The Register reports that this kludge has now been bypassed. I'm kind of surprised that it took this long. Hmmm. I wonder if talking about this makes me liable to prosecution under the DMCA.

As it's the first of the month, I ran web access reports this morning for my sites and Pournelle's. I used to do them weekly, but that got to be too much work for too little purpose. As I suspected, posting my updates at less predictable times (and posting multiple updates during the day) resulted in increased page reads. On a normal Monday, this site might get something like 2,800 page reads. Monday 30 July it got nearly 4,000. On a normal Tuesday, the site might average 2,300 page reads. Tuesday 31 July, it got about 3,000. For the week, our sites combined (including hardwareguys.com and Barbara's) did close to 30,000 page reads, which is decent. For the year, we're averaging just under 4,000 page reads per day for all sites combined, so traffic is increasing gradually.

Here are the TLDs that had 0.05% or more of the bytes transferred last week for this site:

 reqs: %bytes: domain
-----: ------: ------
65842: 50.55%: .com (Commercial)
47709: 20.28%: .net (Network)
32150: 15.90%: [unresolved numerical addresses]
 2559:  1.84%: .uk (United Kingdom)
 3748:  1.49%: .ca (Canada)
 3051:  1.43%: .au (Australia)
 3273:  1.38%: .edu (USA Educational)
  767:  1.20%: .jp (Japan)
 1975:  0.74%: .mil (USA Military)
  552:  0.61%: .fr (France)
 1656:  0.56%: .org (Non-Profit Making Organisations)
  865:  0.42%: .nl (Netherlands)
  961:  0.38%: .us (United States)
  651:  0.31%: .nz (New Zealand)
  469:  0.28%: .ch (Switzerland)
  537:  0.24%: .gov (USA Government)
  351:  0.23%: .de (Germany)
  390:  0.22%: .be (Belgium)
  650:  0.21%: .se (Sweden)
  373:  0.19%: .dk (Denmark)
  337:  0.15%: .it (Italy)
  281:  0.14%: .pt (Portugal)
  141:  0.13%: .fi (Finland)
  236:  0.11%: .mx (Mexico)
   66:  0.08%: .br (Brazil)
   88:  0.07%: .es (Spain)
   72:  0.05%: .ru (Russia)
  130:  0.05%: .ie (Ireland)   

There were 50 or so more TLDs with 0.04% or less of the traffic. Note that "reqs" is requests, rather than page reads. When you call up this page, the page itself is one request, the background image is a second request, and so on, so a large number of requests translate into a smaller number of page reads. I like to look at the domain report to see where people are visiting from.

The visitor count is up somewhat, too, to around 5,000 for the week and about 18,000 for the month on all our sites. Actually calling it "visitor count" isn't an accurate representation, because what my reporting software calls "Distinct hosts served" is determined by IP address. So if you have dial-up access and drop your connection and reconnect five times during the day and visit this site during each session, that might count as five distinct hosts (assuming you received a different IP address for each connection). Conversely, if you have broadband and have the same IP address all month long, you could visit this site ten times a day every day during the month, and you'd still show up as only one distinct host.

Pournelle smoked me, as usual, with about 67,000 page reads for the week, or nearly 10,000 a day.

Click here to read or post responses to this week's journal entries

Click here to read or post responses to the Linux Chronicles Forum

[Top]

Thursday, 2 August 2001

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Messageboard]  [HardwareGuys.com Messageboard]

I'm still getting copies of SirCam--three more overnight--but I got something new in my inbox overnight as well. A copy of SirCam that had no attachment. It had the normal text, but file attached to it. Reminds me of the dreaded Linux email virus, "After you read this message, please send copies of it to all your friends and delete some of your own files."

What's interesting about SirCam is that I've yet to get a copy of it from anyone whose name I recognized. I've gotten many copies from Asia, several from France, a few from South America, and onesies and twosies from Arab countries, India, and Australia. Not a one of them came from anyone I knew. And I'm getting multiple copies from several people. One guy in France has sent me six copies of it in the last three days. As I recall, each had a different attachment. Oh, well. They all go into the bit bucket.

As others have said, SirCam and future viruses/worms like it have the potential to distribute confidential information randomly (or not so randomly) across the Internet. I think if my company dealt in confidential information (and what company doesn't?) I'd be inclined to establish a policy that all documents and spreadsheets were to be secured by password. Word and Excel may not have the strongest encryption available, but assuming a password of reasonable length is used they're secure against anything but industrial strength encryption crackers. There are a lot of Word and Excel password crackers available on the Internet, but as far as I know all of them depend on brute-force decryption, which means that if you use a non-dictionary password of at least seven or eight characters you should be safe.

Expect Pentium 4 motherboards to be in relatively short supply for the next few months. Intel has ramped up Pentium 4 production much faster than expected. That, combined with the precipitous Pentium III ramp-down recently announced, has left third-party motherboard and component makers holding the bag. No one can get enough components to meet the expected demand for Pentium 4 motherboards as we approach the fall selling season. That means that Pentium III motherboards are likely to become a drug on the market. That might not be a problem, except that Intel has set the price for their new Tualatin-core Pentium III processors very high to prevent the Tualatin from cannibalizing Pentium 4 sales. The mass market PC OEMs want the higher clock speeds of the Tualatin, but no one is going to be willing to pay the high premium the 1.2 GHz Tualatin will sell for relative to a 1.5 GHz or higher Pentium 4. Which is exactly what Intel planned to encourage Pentium 4 sales.

Eventually, of course, the Tualatin will be used in the Celeron and the Pentium III will fade away except in the mobile market, where there's no prospect of a mobile Pentium 4 for quite some time. Smart buyers over the next few months will take advantage of low Pentium III motherboard prices. With PC133 SDRAM selling for $0.25 per megabyte and Coppermine-core Pentium IIIs selling at reasonable prices, it's possible to put together a competent system for an incredibly low price, even if you stick to all Intel-branded components. I'd recommend that unless you have good reason to want a Pentium 4, you stick with the Pentium III and SDRAM for now. A Pentium 4 with RDRAM is an expensive solution, and using SDR-SDRAM with a Pentium 4 with SDR-SDRAM simply chokes it for memory bandwidth. The Pentium 4 will begin to make more sense once Intel releases their DDR-SDRAM chipset.

I just got a press release saying that Fujitsu will depart the desktop hard drive market later this year, leaving only IBM, Maxtor, Samsung, Seagate, and Western Digital as suppliers of desktop hard drives. I wasn't surprised at Maxtor's acquisition of Quantum, but I didn't expect Fujitsu to be the next chip to fall. I think we can expect continuing consolidation in this sector. By this time next year, we'll probably be down to three--IBM, Maxtor, and Seagate.

Well, my advice remains the same. Buy Seagate and Maxtor hard drives.

Click here to read or post responses to this week's journal entries

Click here to read or post responses to the Linux Chronicles Forum

[Top]

Friday, 3 August 2001

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Messageboard]  [HardwareGuys.com Messageboard]

Hmm. Nothing much to write about this morning. I have a bunch of chapters in progress and should have several posted on the subscribers-only page soon, possibly the first of next week.

 

 

Click here to read or post responses to this week's journal entries

Click here to read or post responses to the Linux Chronicles Forum

[Top]

Saturday, 4 August 2001

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Messageboard]  [HardwareGuys.com Messageboard]

A month or so ago, Seagate announced the Barracuda ATA IV drive, but I somehow missed that announcement. My contact at Seagate emailed me last night to tell me about the IV and to say that he was sending me one. From the press release, it sounds like the drive to buy. It stores 40 GB per platter, and will be sold in 20, 40, 60, and 80 GB versions. The internal transfer rate is nearly 70 MB/s, which means that it's the first drive that will actually show a real performance benefit from using the ATA/100 interface. According to Seagate, the drive is fast enough to stream eight DVD movies simultaneously without dropping a frame. By that, I presume they mean the drive can stream eight discrete DVD files rather than one interleaved file. That's one fast drive. 

Just as important for some applications is that it's a very quiet drive. The 20 GB and 40 GB single-platter models generate only 20 dB at idle and 24 dB when seeking, which is inaudible. The dual-platter models presumably generate somewhat more noise, but the level is not specified. Even so, I'm sure the dual-platter units will also be extremely quiet drives. Along with the power supply and cooling fans, the hard drive is the major source of noise in a PC. Using a PC Power & Cooling Silencer power supply and a Seagate Barracuda ATA IV drive, it should be possible to build a high-performance PC that is inaudible (or nearly so) while running. For some applications, that may be important.

I just noticed an interesting editorial over on Storage Review. Scroll down to the 20 July entry titled On Pop-ups and Alternative Advertising. There'll you'll find that Storage Review's ad revenues as of 5/19/01 were down by 90% from 1/1/2001. A drop to 10 cents on the dollar in less than six months. They're now making barely enough to continue running their servers, with almost nothing left over to pay the people that generate the content. No wonder ad-supported tech sites have been dropping like flies. Many of those that remain, like AnandTech, are pale shadows of their former selves. No doubt even big-name sites like BYTE are also suffering. It all comes down to a simple fact that I first stated years ago. Advertising doesn't work and can't work on the Internet, and sites that depend on ad revenue are doomed to fail. It's simply not a sustainable business model.

Ironically, the editorial praises the "free Internet" but complains about  "those who use software to suppress advertising and to protect their God-given right to get something for nothing." People who own ad-supported web sites often argue that using ad-blocking software is somehow the moral equivalent of theft. It's not, of course, any more than taping a movie and zapping the commercials is theft. People are free to pick and choose what they want to look at, and are under no obligation to view your presentation as a whole. Pop-ups, pop-unders, and similar obnoxious ad methods are simply the last gasp of Internet advertising. They're actually counterproductive in that the increasing use of intrusive ads simply encourages more people to install ad-blocking software. 

Such software has gone from being a niche product, like Internet JunkBuster, used only by knowledgeable computer folks to mainstream products, like Norton Internet Security, which is installed and used by ordinary people. I belong to several non-computer mailing lists that are frequented by ordinary people interested in things like astronomy and mystery fiction. For them, computers are just a tool, and not something they think a lot about. A year or so ago, there weren't any references on those lists to ad-blocking software. Nowadays, there are frequent threads about ad-blocking software, including free stuff like WebWasher. If everyone isn't already running ad-blocking software, they soon will be.

Instead of railing about ad-blocking software and accepting ever more intrusive ads, the owners of this site need to recognize that Internet advertising is dead and convert their site to a sustainable revenue model, which is to say subscriptions. The ad-supported mindset, left over from the days when banner advertisers paid per impression, is that readers are an unmixed blessing. In those days, page reads translated directly to revenue. That hasn't been the case for a long time, though. Advertisers got smart. They realized early that impressions didn't count for much. What really counted was click-throughs. But nowadays click-through rates have plummeted to ridiculously low levels. Most sites would be delighted to have a 0.5% click-through rate, and 0.1% rates are probably closer to the norm. That means you may have to serve 1,000 page reads or more to get one click-through, and serving those 1,000 page reads costs serious money. So visitors who read your pages without clicking on an ad cost you money, whether or not they're using ad-blocking software. Running a high-volume site with a very small click-through rate is a fast way to lose your shirt. All those readers generate higher costs, but not the higher revenues needed to support the volume.

The owners of the site obviously recognize already that having many readers per se doesn't guarantee adequate revenues. What they need to understand is that there's no way to fix that. The mechanism itself is unsustainable. And in fact their niche may be too small to be viable on a subscription basis as well. But instituting a subscription model is their best hope. They'll have to be careful about pricing. Something in the $25/year range is reasonable for a site one visits daily. But I don't think Storage Review is that kind of site. I visit it once a month or so, and I suspect that's closer to the average. I think they're likely to find that most of their frequent visitors are of the free-loading sort. They visit daily to cruise the forums and so on, but are unlikely to be willing to pay a $25/year subscription to do that. The people most likely to be willing to pay a $25/year subscription are the folks who visit them once a month or whatever, looking for information to help them make good buying decisions. But there are relatively few of those. 

In order to maximize revenue, Storage Review should probably charge something like $5/year. At that level, essentially 100% of the serious information seekers will subscribe, and perhaps 10% or more of the casual visitors. If they tried charging $25/year, they'd probably get 50% of the serious information seekers and perhaps only 0.1% of the casual users. I don't know what numbers Storage Review generates, although their Advertising page claims 3,000,000 visitors per month. I think they're talking about page reads or even requests rather than actual visitors, but even so that's substantial traffic. With that number of users, a properly solicited $5 annual subscription should generate more revenue than they're generating now from ads. Not to mention giving them added credibility. Consumer Reports has recognized since its inception that one can't accept ad revenue and remain unbiased. The same is true for sites like Storage Review. They'd be doing themselves and their readers a service by instituting a subscription plan.

And speaking of subscribing, if you're a regular reader and you haven't yet subscribed, why not? Click here for instructions about how to subscribe.

The UMAX 3400U scanner continues to function flawlessly. The scan quality is not as good as that provided by the HP 6200C (when I could get it working), but it's more than good enough for what I need to scan, and better than I expected from a $70 scanner. The scanning software has a hokey interface, but it does the job. If you need a cheap scanner to use under Windows 98, you could do a lot worse than the UMAX 3400U.

 

Click here to read or post responses to this week's journal entries

Click here to read or post responses to the Linux Chronicles Forum

[Top]

Sunday, 5 August 2001

[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week]
[Daynotes Journal Messageboard]  [HardwareGuys.com Messageboard]

Boy, talk about a rip-off. Dell and Gateway obviously hope their customers don't know the price of memory.

I was just over on the Dell site configuring an example system. The basic configuration came with 128 MB, which is ridiculously little with the current price of memory. Dell offers upgrades to 256 MB or 512 MB, but their prices are outrageous. Upgrading the standard 128 MB to 256 MB costs $110. Upgrading the standard 128 MB to 512 MB costs $310. I checked the Gateway site to see what they were charging for memory upgrades. Gateway was only a bit better, at $100 for the 128 MB to 256 MB upgrade and $300 for the 128 MB to 512 MB upgrade. 

So I checked the Crucial web site, where I found that a 128 MB DIMM was going for about $22, a 256 MB DIMM for about $40, and a 512 MB DIMM for about $99. That means it should cost about $18 rather than $100/$110 to upgrade from 128 MB to 256 MB, or about $77 rather than $300/$310 to upgrade from 128 MB to 512 MB. Actually, less than that for the 512 MB upgrade, because Dell or Gateway would certainly install two $40 256 MB DIMMs in preference to one $99 512 MB DIMM.

It's appears Dell and Gateway hope to make large profits by selling memory upgrades, and I wouldn't be surprised if their salespeople are pushing those upgrades. But charging people four or five times what something is worth is a good way to lose customers. Certainly those who learn the real cost of memory before they order, but even more so those who learn it after they've ordered. If you'd paid Dell or Gateway $300 for a memory upgrade only to learn that it should have cost you $60, would you ever buy a system from them again?

The Romans had advice for this situation. Caveat Emptor.

 

 

Click here to read or post responses to this week's journal entries

Click here to read or post responses to the Linux Chronicles Forum

[Top]

 

Copyright 1998, 1999, 2000, 2001, 2002, 2003, 2004 by Robert Bruce Thompson. All Rights Reserved.