Monday, 10 June 2002
9:14 - There's a partial solar eclipse today, but it's not visible from our location. Because this eclipse straddles the International Date Line, the eclipse begins the day after it ends, which seems odd. The eclipse begins in Southeast Asia and Australia at sunrise on Tuesday morning, June 11th, and ends at sunset on Monday, June 10th in the central US, Southern Canada, and Central America. From our location, the eclipse will just be beginning at sunset.
You can find more information about the eclipse here. If the eclipse is visible in your location, whatever you do DON'T LOOK AT IT WITH YOUR NAKED EYE. Doing so can blind you almost instantaneously. If there's an astronomy club or science center near you, contact them to find out about safe viewing opportunities. Many will have organized public viewing sessions using equipment that allows viewing the eclipse safely.
This week I'm writing, all week, so updates here will be sporadic.
11:35 - Is it just me, or do other people sense an "Empire Strikes Back" campaign going on? A lot of articles have been published in the last week or two, all of them condemning Open Source Software (often damning it with faint praise) and praising Microsoft software, at least by implication. I sense the hand of Moriarty behind all this.
All of the articles I've read have originated with supposedly independent organizations that in fact accept funding from Microsoft (such as de Tocqueville anti-GPL security white paper released; white paper here), with advertising-supported publications (such as ZDNET's The very real limitations of open source), or with consultants, OEMs/VARs, systems integrators, and other people who earn their living with Microsoft software. I suspect that Microsoft is pulling out all the stops to encourage everyone over whom they have any influence to write "independent" criticisms of Open Source. Of course, anyone who knows anything will recognize that all of these articles and white papers are dubious at best, if not simply outright lies, but the danger is that these articles will be accepted as truth by those who don't know what's going on.
If you can't argue something on its merits, there's always FUD, and as far as I can see all we have here is a gigantic pile of FUD. I've said this before, but I really do think Microsoft is becoming desperate. They now appear to be attempting the Big Lie. What concerns me is that the Big Lie combined with Microsoft's increasing attempts to legislate OSS out of existence may eventually succeed.
I hope that Licensing 6.0 fails miserably, and that thousands of corporations hand Microsoft its head on a platter. Those organizations that refuse to sign on to Microsoft's Licensing 6.0 have a window of opportunity of one to two years to convert their operations to OSS, and I hope that many choose to do so. A year from now, OSS applications will be even better than they are now--and they're pretty much Good Enough for corporate use already. Microsoft has to win this war quickly or they'll have no chance to win it at all. Over the next few months, expect to see increasingly frequent and frantic attacks on OSS, all of them originating from Redmond.
Tuesday, 11 June 2002
9:17 - I've read the original version of the Alexis de Tocqueville Institution white paper entitled "Opening the Open Source Debate". That version was quickly withdrawn, and a new version was to have been posted by yesterday afternoon. If you haven't read it yet, don't bother. If I were Microsoft, I'd demand my money back. The paper is poorly reasoned and ungrammatical, and reads like a re-written Microsoft press release. It delivers nothing more than a pathetic rehash of Microsoft's foaming-at-the-mouth criticisms of the GPL. The Register has already posted a very good point-by-point rebuttal, so I won't bother even attempting to refute this tissue of lies and distortions.
11:55 - Well, I couldn't resist it. I downloaded the "new" version of the de Tocqueville white paper and checked a couple of things manually. I didn't do a line-by-line comparison, so I don't know what other changes were made for the new, improved Hatchet Job 2.0, but here's one example:
At least the Hatchet Job 2.0 sentence parses, which is more than can be said for the corresponding sentence in Hatchet Job 1.0. But, although they may have cleaned up the grammar, I'd be willing to bet they haven't done anything to repair the logical fallacies, unproven assumptions, and outright misstatements upon which the paper is based.
I find it hard to believe that any reputable organization would put its name on this white paper. I hope the Alexis de Tocqueville Institution made a lot of money for writing this paper, because it certainly destroyed its reputation and credibility by doing so. How can anyone now take seriously anything the Alexis de Tocqueville Institution writes or has written? As far as I'm concerned, their name is now just a bad joke.
12:19 - Roblimo posted an excellent article on Newsforge that counters the discredited Alexis de Tocqueville Institution Opening the Open Source Debate white paper with the year-old government-sponsored Mitre report, A Business Case Study of Open Source Software. I did a quick scan through this document and, as Roblimo says, as it stands the study is quite favorable to OSS, and it would likely be even more so if written today, given the improvements in OSS over the last 12 months. Whereas the AdTI report favors the interests of Microsoft and other proprietary software companies, the Mitre report favors the interests of users.
13:23 - I've reconsidered my position. The AdTI report is right. We should all abandon this communistic OSS stuff and get with the plan. Why? Because the US can't afford the job losses that OSS will cause. Upon closer reading, I came across this shocking statistic in the AdTI report:
I didn't realize OSS could cause that much damage. Why, that's more than one job for every man, woman, and child in the U.S, and all of those jobs depend on stopping OSS. An OSS victory would be an economic disaster. None of us would have jobs. The economy would grind to a halt overnight. We'd all starve. Something has to be done to stop this.
Wednesday, 12 June 2002
9:54 - AMD says there's a bug in the Linux kernel. Hmmm. The last time this happened, it turned out there was a bug in the Athlon rather than in Linux. My recollection of the details of that last one is hazy, but this one sounds similar. Oh, well. I'm sure they'll get it worked out.
PC Hardware in a Nutshell hits the bookstores in less than two weeks. That means I need to get in some heads-down work on the HardwareGuys.com website, so I'd better get to work.
12:20 - Oh, good. I just got email from BellSouth saying that DSL is now available in my area. Let's see. It's been two years this month since we had the cable modem installed. At that, we were six months behind the rest of Winston-Salem in having cable modem service available, and Time-Warner was about two years late in deploying cable modems for most of Winston-Salem anyway. That means BellSouth is more than four years late in deploying DSL here. Geez. Talk about being late to the party.
And their service is nothing special, either. A 1.5 Mb/s download cap and a 256 Kb/s upload cap. No static IP address. And they charge $45/month if you supply your own modem. Geez. Too little, too late. From what my Canadian readers have told me, broadband access in most areas of the US is grossly overpriced. Something in the $20 to $25/month range might let BellSouth take customers away from RoadRunner. But $45/month is simply too much to allow them to make any sort of penetration into an existing market that's well served by cable-modems.
I'm sure that Pournelle and others who don't have broadband access at all would kill to get it for $45/month. But I predict BellSouth isn't going to be very successful showing up more than two years late to the party and charging $45/month. Nearly everyone in Winston-Salem who wants broadband already has it. BellSouth is going to pick up some customers who are outside the cable-modem service area and some new customers who've just moved in. But they're not going to steal many customers from RoadRunner.
Thursday, 13 June 2002
10:08 - There's an Interesting article on NewsForge about introducing Linux to folks at a retirement home. It looks like it may not be long before Pournelle's Aunt Minnie is using Linux herself.
Speaking of Linux, I occasionally get nastygrams from people taking me to task for referring to Linux as Linux rather than as GNU/Linux. I've read what Mr. Stallman has to say on the subject, and I'm not convinced. If Mr. Stallman had his way, I'd call the OS I'm running Red Hat GNU 7.3. According to Mr. Stallman, the "Linux" part isn't important because Linux is only the kernel. Only the kernel?
According to Mr. Stallman, the GNU project built most of the operating system, and lacked only the kernel. That strikes me as similar to saying that they'd built most of a computer, and lacked only the chipset, BIOS, and CPU. Certainly the GNU utilities are important to Linux, but without a kernel all they had was a collection of mostly reverse-engineered utilities, which Mr. Stallman refers to as the "GNU Operating System". An operating system without a kernel. Now there's a novel concept. The heart of Linux is the Linux kernel, so it seems to me appropriate to refer to the operating system itself as Linux, which I will continue to do. If that offends anyone, so be it.
11:24 - News on an interesting Mozilla bug:
If I understand the problem, it doesn't occur on Windows. I had several instances of Mozilla open on the task bar when I read your email. When I clicked on the link in the email, Mozilla opened it as a new instance. I'm not sure what "profile corruption" means or why opening a new instance from an embedded link would cause that to happen. Perhaps I'm risking corrupting my profile in Windows by clicking on links in email messages, but if so I haven't noticed any problems other than the usual Mozilla bugs.
Friday, 14 June 2002
9:00 - I suppose it's a commentary on modern life that it always comes as a shock to me when a business does the right thing despite the fact that it costs them some money to do it.
Barbara has pre-paid cell phone service from Cingular. Every three months we buy a phone card to add $20 or $30 worth of time to the phone. Any unused balance "rolls over" and so is not lost. Unless, that is, we let that balance expire by failing to add at least $20 to the balance within the 90-day deadline. Our balance, $49 or so, was due to expire on June 9th. On June 5th, Barbara and I went to Wilson Communications, the store where we'd bought the phone and each subsequent card. We bought a $20 card to refresh her balance.
Yesterday, Barbara called to say that her phone had stopped working. When she called 611 on it, the automated attendant told her that she had $0.00 and 0 minutes on the phone. As soon as Barbara told me that, I suspected that the time on the new card we'd bought on 5 June hadn't been installed properly. In the past, the guy we bought the card from had used the card we just got to re-fill the phone. We thought they did that automatically, since they punch up the serial number on the card into a little electronic thingee near the cash register, and each time we'd bought a card in the past we hadn't had to enter it ourselves.
That turned out not to be the case. They were pretty busy when we went in to buy the card, and apparently the guy we bought it from didn't enter it into the phone. We never even thought about it until Barbara's phone died. So I called them and Gary told me to bring in the phone and they'd make it right. While we sat there, the guy from Wilson Communications called Cingular and explained the problem. He wasn't able to get anywhere with them, of course. So he entered the $20 card Barbara had bought on the 5th and bought and added a $50 card out of his own pocket.
As Gary and I shook hands, I told him that although he'd just eaten $50, that might be the best $50 he'd ever spent on advertising. We'll never go anywhere else for phone stuff, of course, and we'll also tell all our friends how Wilson Communications does business. They'll make that $50 back many times over.
13:10 - The first crack in the dam? Wal*Mart is now shipping eight system configurations with Lindows pre-installed. Microsoft must be gnashing their collective teeth. Microsoft hates it when obscure little garage shops ship boxes with Linux installed. How much more must they hate it when a high-profile mass marketer like Wal*Mart is now advertising and shipping PCs that have no Microsoft software installed on them?
Of course, Wal*Mart is rather exaggerating the capabilities of Lindows when they claim Lindows has "the ability to run most Microsoft programs" (and in bold face, yet). That's misleading at best and many would consider it a fraudulent claim. I suspect Microsoft's lawyers will be all over Wal*Mart before the day ends.
I also find it odd that Wal*Mart is shipping PCs with an OS that's still in beta. When I checked the Lindows web site just now, I found that they still aren't shipping their final version. Also, I can find no link to download Lindows, which is rather strange for a Linux-based OS. Michael Robertson, the CEO of Lindows, has already drawn the ire of some OSS folks for failing to make source code available and for the Lindows End User License Agreement, which flies in the face of OSS and FSF by restricting distribution of the product.
And speaking of teeth-gnashing, AMD hates it when people advertise their Athlon processors using the actual clock speed rather than AMD's hokey Performance Rating model number. I see that Wal*Mart is doing that, which won't make AMD happy.
14:03 - A year ago, I got very few questions about how to reduce the noise of PCs. That subject is getting a lot more popular, possibly because people are using faster processors and more memory, which in turn requires higher output power supplies. I must confess that I've done little research or experimentation on quiet PCs. I sometimes use PC Power & Cooling Silencer power supplies in situations where low sound levels are necessary, but other than that I haven't worked with any of the sound-reducing technologies. Here's an interesting report by a reader who decided to build a quiet PC for himself. If you're interesting in reducing the noise level of your PC it's worth reading.
Saturday, 15 June 2002
10:00 - Microsoft says there's a bug in the way Windows XP handles the registry that results in blue screens. Apparently, XP Home and XP Pro can munge their own registries, rendering the system unbootable.
Microsoft has issued a hotfix for the problem, but apparently doesn't trust it much because they recommend you not apply the hotfix unless you are severely affected by the problem. They say "if you are not severely affected by this problem, Microsoft recommends that you wait for the next service pack that contains this fix." Of course, they don't say when that SP will be available, so it appears you have the choice of (a) applying a hotfix that Microsoft itself doesn't trust or (b) going without and risking your system becoming unbootable, at which point you have presumably become "severely affected" and in retrospect should have applied the hotfix. Talk about damned if you do and damned if you don't. Microsoft logic is not Earth logic.
See http://support.microsoft.com/search/preview.aspx?scid=kb;en-us;Q318159 for further details.
My own view is that anyone who is running Windows XP should upgrade to Windows 2000 immediately.
More on silent PCs from reader Michael Hipp:
Thanks. I must confess that I've never paid much attention to PC noise levels, other than using a PC Power & Cooling Silencer power supplies for systems that are to be used in noise-sensitive environments. It seems to me that a PC that uses a Silencer and a very quiet hard drive like a Seagate Barracuda ATA IV makes so little noise that there isn't a lot of point to looking for further noise reductions, but perhaps I'm just less sensitive to noise than some people. Perhaps at some point I'll build a "Quiet PC" project system.
The weather is to be good tonight, so Barbara and I will probably head up to Pilot Mountain to observe. Nautical Twilight isn't until 21:49 and Astronomical Twilight at 22:36, so it'll be another late night. Luna is at about 28% illumination, and moonset is at 00:11, so we'll have reasonably dark skies for observing deep-sky stuff.
Sunday, 16 June 2002
11:20 - When Pournelle first reported problems downloading the latest ATI drivers, I decided to do him a favor and just download the drivers via my cable modem and then upload them to Pournelle's server. To my surprise, that turned out not to work. I've now been trying for two days to download the Windows 2000 drivers, and I can't get them. They're something like 14 MB, and each time I try to get them the download process blows up after anything from 2 or 3 MB up to 6 or 7 MB. Nor does attempting to restart the download work. Instead of restarting the download where it left off, the restart thinks about for a minute or so and then starts downloading the file from the beginning. This is very annoying.
Word on ADSL costs in Spain:
Thanks for the kind words. Those prices seem a bit higher than typical US prices, but not tremendously so. Of course, ADSL prices here tend to vary according to what alternatives are available. In areas served by cable modem, DSL prices are typically a lot lower than for areas that don't have cable modem as an alternative. When BellSouth first offered me DSL three years ago (even though they really couldn't provide it at my location) they were offering 1.5 Mb/s down and 384 Kb/s up for $70/month, but that was because cable modems hadn't been deployed here yet. Also, the speeds they quote for ADSL are "typical" and aren't guaranteed. Those at the far end of the loop are likely to see slower speeds.
Business school teaches a concept called price elasticity of demand. That's a fancy way of saying that more people will buy something at a lower price than will buy it at a higher price. If the product you're selling has high fixed costs and low marginal (unit) costs, you'll probably make more money by selling more product at a lower price than you will selling fewer units at a higher price.
Broadband Internet service, cable or DSL, has very high fixed costs and very low marginal costs, so the idea is (or should be) to get as many people signed up and paying by the month to help offset those high fixed costs. (Before any writes to tell me, I realize that if you take a long enough view "fixed costs" aren't really fixed. It's safe to call broadband infrastructure costs "stepped costs" in the sense that if you keep adding customers you'll eventually have to upgrade your infrastructure--faster servers and routers, a bigger pipe to the backbone, etc.--so your "fixed costs" do increase over time.)
It seems to me that cable modem and DSL providers have done a pretty poor job of optimizing revenues. One sure sign that you've not done price elasticity of demand calculations correctly is that your product has poor penetration, and broadband has very poor penetration even in markets where one would expect it to have high penetration. I've read various figures, but some indicate that broadband penetration in many middle-class and upper middle-class neighborhoods is as low as 5%. That in areas where nearly every household has a PC, many have multiple PCs, and most of them connect to the Internet by some means, such as AOL dialup. So for every twenty households, one uses broadband and perhaps 15 use dial-up. Why? Because dial-up costs them $25 a month, whereas broadband might be $50 or $60 a month.
The advantages of broadband over dialup are clear--higher speed and always-on. But it's obvious that most people aren't willing to pay twice or more as much for those advantages. If a broadband company already has the infrastructure in place (and has already paid the costs of putting it there) it's simply stupid to let so many potential customers continue using dialup. If the broadband companies dropped that differential to $10/month, they'd get a lot more customers, enough more I suspect to make substantially larger profits on their investments. If they dropped the differential to $5/month, they'd get most of the dialup customers, and it they dropped it to $0/month they'd get essentially all of them. Their revenue from those 20 households would increase from $60/month to $400/month, and their costs would not increase proportionately.
I think the problem is that the broadband companies are still focused on the early adopter phenomenon. With any new technology there is going to be a certain percentage of the population that is willing to pay high prices to have that technology first. That's all well and good from the point of view of the company that produces the technology. They get very high per-unit profits, which helps them pay the development and deployment costs. But if they make the mistake of trying to maintain those high prices too long, they miss the boat on mass market acceptance. Which is what I think is happening with broadband.
Copyright © 1998, 1999, 2000, 2001, 2002, 2003, 2004 by Robert Bruce Thompson. All Rights Reserved.