Home » Daynotes Home » Week of 11 April 2005

Photograph of Robert Bruce Thompson
Daynotes Journal

Week of 11 April 2005


Latest Update: Friday, 15 April 2005 10:31 -0400
Free Speech Online - Blue Ribbon Campaign

Paying for this Place
Visit Barbara's Journal Page

Monday, 11 April 2005
[Daynotes Forums]    [Last Week]   [Mon]  [Tue]  [Wed]  [Thu]  [Fri]  [Sat]  [Sun]   [Next Week]    [HardwareGuys Forums]

11:43 - I'm back.

Astronomy Hacks is complete, at least from our point of view. Starting this morning, the O'Reilly production folks go to work converting our manuscript into a finished book. They're fast-tracking this one big time. Ordinarily, it might take four months from submission of the completed manuscript until the books hit the warehouses. For this book, they're going to do it in two months.

Our part of the work isn't completely finished. We'll have queries to respond to, galley proofs to check and correct, and so on. But it's all downhill from here, as far as our part is concerned.

We're pretty happy with the book overall. I said to Barbara the other night, "This would have been a better book if we'd had more time." She replied, "You always say that." And both are true. But at some point, you just have to declare the book finished, or it would go on forever.

I remember at one mystery conference encountering a lady who was hauling around the manuscript mystery novel she'd written, trying to get an agent or publisher to look at it. I asked her how long it had taken her to write her book, and she said almost twenty years. I remember thinking she'd need more than the usual amount of luck to talk an agent into representing her or a publisher into publishing her if the most they could expect from her was one book every twenty years. If I were writing novels, I'd be knocking out at least two or three a year.

It's funny. Early in her career, Lois McMaster Bujold had an elevator encounter with the guy who would eventually become her publisher. He told her, "If you write three books a year for me for seven years, I'll make you a star." She replied (or at least thought), "How about one book a year for twenty-one years?"

Although I've never written fiction, I've never understood why everyone claims it's so difficult and time-consuming. Writing consists of organizing your thoughts, constructing sentences and paragraphs from those thoughts, and then getting those sentences and paragraphs down on paper or up on the screen. It's not rocket surgery. I've always suspected that people who claim they can't write, or can write only slowly, in fact can't think, or can think only slowly.

It has been said of Isaac Asimov, who cranked out books like a chicken cranks out eggs, that his experience with writer's block was the worst ten minutes of his life. That means Asimov is ten minutes up on me, so far. I've never experienced writer's block, and I hope I never do. When I sit down at a keyboard, I just start writing. Now, sometimes what I write is crap, certainly, but I've never had any problem with churning out some sort of text.

I haven't done a word count on Astronomy Hacks, but I'd guess I probably averaged 2,000 to 2,500 words a day, seven days a week, during the 69 days of this project. In reality, I would have been cranking out about twice that if it hadn't been for all the time needed for non-writing tasks, such as shooting images, talking to vendors, interacting with O'Reilly, and so on.



The latest example of someone being screwed by proprietary software is ironic.

Linus Torvalds made a big mistake back in 2002 when he standardized on BitKeeper source-code management (SCM) software to maintain the Linux kernel. Stallman tried to tell him, but as usual Stallman used his lunatic-fringe political arguments instead of some very practical arguments that might have convinced Linus to choose an open source alternative.

The basic situation is this: BitKeeper was and is the best product on the market for maintaining large distributed programming projects with many developers. BitKeeper sells a commercial version of their product, but until recently also offered a free (as in beer, not as in speech) version that was used by many contributors to the Linux kernel. BitMover, the company that produces BitKeeper, just pulled the rug out from under Linux development by yanking the free version of their software, replacing it with a more crippled version that's inadequate for Linux kernel development. BitMover's motivation for the change was that at least one OSS programmer was writing an SCM that was interoperable with BitKeeper. Linus's only real choice now is to abandon BitKeeper and move to some open source SCM system. None of those are very good. Yet.

In the short term, Linux kernel development will be a bit chaotic. But that won't last long. As bad as Linus's decision to go with BitKeeper was, Larry McVoy, the owner of BitMover, really blew it. Presumably, he thinks Linux kernel development will move to the commercial version of BitKeeper. That ain't gonna happen, for several reasons. Instead, with one stupid decision, McVoy just destroyed his own company.

I mean, think about it. You have thousands of the world's best programmers working away on the Linux kernel, which epitomizes free and open source software. They're happily using the "free" version of BitKeeper, until McVoy yanks it out from under them. What are they going to do? Buy a copy of the commercial version of BitKeeper? Not bloody likely. They're going to turn, en masse, on McVoy's company, and destroy it. How? Well, I'll pose a theoretical question. How long will it take for a self-selecting subset of the 5,000 top programmers in the world to churn out an open source competitor to BitKeeper that matches and exceeds the abilities of the commercial product? In probably six months, or perhaps a year at the outside, McVoy will face a superior product that sells for $0. His company will go down the tubes.

If McVoy had been smart, he would have offered a free copy of his commercial version to anyone working on Linux kernel development, or indeed on any free/open-source project. He would never have been able to sell his product to those people anyway, but by giving it to them he would have eliminated their motivation to produce their own competing product. McVoy could have continued selling the commercial version to commercial and private software developers. McVoy would have been a hero to the free/OSS folks, and could have continued making good money on his commercial version. Now, he's screwed. The entire market, free/OSS, commercial, and private developers, will eventually switch to the free version-tracking software, and probably sooner rather than later. Why not? It'll be as good or better than BitKeeper, and it won't cost anything. BitMover, RIP.



[top]

Tuesday, 12 April 2005
[Daynotes Forums]    [Last Week]   [Mon]  [Tue]  [Wed]  [Thu]  [Fri]  [Sat]  [Sun]   [Next Week]    [HardwareGuys Forums]

08:47 - We have a new cover for the book.

Astronomy Hacks Cover

Yesterday was occupied with the usual last-minute stuff. The book actually went to production about lunch time. I experienced the usual let-down when a book reaches completion, going from full-speed ahead and dozens of balls in the air to having nothing left to do. Now it's pretty much up to O'Reilly's production folks to get the book ready and out the door. It should hit the bookstores around late June. I'm already thinking about the second edition.



Now it's back to the grind. I have to get the taxes done this week, and I have to catch up on a lot of things I'd let slide while we completed the book. Barbara has a list of things that need to be done around the house.

My Windows XP system died last week, as the book was in its final stages. Well, it didn't die, but the grace period on activation ran out. It still worked, but I knew that would change if I rebooted. Every couple hours, it'd pop up a message to tell me the grace period had expired. When I finally rebooted it yesterday, it wouldn't let me log in. Oh, well.

I'll probably convert that system into my new primary desktop. My current primary desktop runs Xandros 2.5 Business Edition on an Antec Aria SFF system with an Intel D865GRH motherboard with a Pentium 4/3.2 processor. The dead XP system is an Intel D925XCV motherboard with a Pentium 4/570 processor and an nVIDIA GeForce6 6800 in an Antec P160 case. I'll install Xandros 3 on that system, and convert the Aria system to a Windows 2000 test-bed, assuming I can find one of my Windows 2000 CDs. I may not even bother to connect that system to the network. I'll just use it to rip and burn CDs and DVDs. If dvdshrink were available for Linux, I wouldn't bother having a Windows system at all, other than test-bed systems for doing screen shots for the hardware books.

I was also thinking about building a PVR system around MythTV, but I don't think I'll bother. Barbara and I dropped everything but basic cable TV service in February, and we haven't missed anything. There's almost nothing on we want to record, so it seems pointless to build a PVR system. I may still build one, but if I do it'll be for the hardware books and as an intellectual exercise rather than to fill any real need.

And there are a bunch of other projects I need to get to work on.



I got some interesting mail about the BitKeeper issue. Most people seem puzzled about the position Linus Torvalds has reportedly taken, and I confess I'm among that group. Some, however, blame Tridge for the crisis. Here's a representative message from someone who thinks Tridge is responsible.

-------- Original Message --------
Subject: bitkeeper
Date:    Mon, 11 Apr 2005 17:05:03 -0500
From:    Christensen, Chris (Aspen Research)
To:      Robert Bruce Thompson

Robert:

You said "Linus Torvalds made a big mistake back in 2002 when he standardized on BitKeeper "

You could also argue that in 2-3 years with bitkeeper, the Linux kernel has made much greater strides (in what is becoming and ever more complex piece of software) than in any similar period before bitkeeper, and the kernel developers are certainly no worse off now than they were before bitkeeper.  I wouldn't call it a big mistake: it was a choice that had consequences.

I would say that Larry McEvoy made the larger mistake.  Andrew Tridgell whose obstreporous antagonism toward bitkeeper is the reason for the current state of affairs has, I think, something to answer to the rest of the kernel developers for.  The experience with bitkeeper has given expectations to some very smart software developers, so I agree that there should soon be a viable open-source/free competitor.

There may not be an english language verision of it:

http://www.theinquirer.org/?article=22440

and

http://www.wired.com/news/business/0,1367,67181,00.html

I think it's like the old Fram oil filter commercial, "Pay me now, or pay me later." Three years ago, Linus could have chosen not to go with a commercial SCM system. If he'd done that, we'd have had a usable OSS SCM within a few months. The Linux kernel might be a few months behind where it is now. But we'd have an OSS infrastructure in place, and wouldn't be dealing with a sea change today. Better to have bitten the bullet back then, when things were simpler, than throw a monkey-wrench into the works now.

As to the other, I'm not particularly worried. Linux in particular and OSS in general are usually represented as a world-wide effort, but the truth is that US and western European programmers are responsible for about 99.9% of the software we use, both OSS and commercial. I haven't seen much coming out of Africa, South America, or Asia, unless you count things like Ubuntu, which is pretty much a repackaging of the work of US and western European coders. As to there not being an English-language version, any such product would be doomed to fail.



14:43 - I've been cleaning up the mess that always results when I write a book. This time, there are only a few hundred files, most of them images, and a total of a couple gigabytes of disk space occupied. I was going to move everything to my "holding" directory, which is kind of an intermediate storage area between active and archive, but I noticed that holding is already at 4 GB.

I use holding so that I don't have to back up my main archive directories frequently. Stuff remains in our working directories until they no longer fit on one DVD, at which point I sweep everything not actively in use to the holding directory and burn a new DVD of the holding directory (actually, I burn two copies, both to have a spare and in case a file isn't readable from one or the other DVD.) When holding gets full, it all gets swept to the archive directories and I then burn a new set of archive backups to DVD. Since holding is just about full now, I might as well sweep everything--holding and unused active files--directly to archive.

But that means I'll need to burn a new set of archive DVDs. Urk. I just checked, and the archive directories are currently about 45 GB. That's 10 or 11 DVDs per copy, times two for the spare copy. Hmmm. Checking further, I see that about half of the disk space is being eaten by /archive/install, which contains installation files, ISO images, etc. And I wonder if I'll ever again need Office 97, Windows NT SP1, or Red Hat Linux 5.2, to name just a few.

I freely admit that I am a packrat, but surely even a packrat has its limits. It sounds stupid, but I'm afraid if I delete something I'll later find I need it desperately and it's no longer available anywhere. That actually has happened in the past, although admittedly not often. Well, okay. Once.

But looking at these directories I think to myself that some of these files may literally no longer exist anywhere except on my hard drives. I mean, I have full sets of patch files for Novell NetWare 3.X. I have .arj files, for heaven's sake, not to mention .arc files that predate the creation of the .zip format. And I just found an executable with a 1985 date, and some data files from 1981. If I save these files, archaeologists hundreds or thousands of years in the future may give thanks when they dig up my hard drive.

I am a keeper. Barbara is a thrower-away. I've often wondered about married couples who are both one or the other. If they're both keepers, they must constantly have to buy larger homes. If they're both throwers-away, they must spend a lot of money replacing things they should have kept.

I remember years ago being in a museum, looking at an artifact. I don't remember what it was, but I do remember that it was described as "the only known example of...". And I remember thinking at the time that sometime back in about 1515 that artifact was owned by a married couple. The wife kept encouraging the husband to throw it out, but he resisted. And, because he insisted on keeping it, we now have "the only known example" sitting in a museum.



[top]

Wednesday, 13 April 2005
[Daynotes Forums]    [Last Week]   [Mon]  [Tue]  [Wed]  [Thu]  [Fri]  [Sat]  [Sun]   [Next Week]    [HardwareGuys Forums]

08:30 - Bob Walder is back. I used to read Bob's journal page every day--one of my few regular daily visits to websites--and I was sorry to see him abandon it in mid-2002. But Bob is back now, so presumably we'll be reading more stories about him, Lynne, Benson, and a new family member.



Chuck Waggoner has this to say over on the messageboard:

It is inconceivable to me that there would not be an English version of any significant program.  EVERYBODY in my German classes in Deutschland, speaks English--and they come from all over the globe (amazingly, among the best English-speakers are the French).  A lot of advanced technical classes in Europe are taught only in English.  My experience has been that the best engineers and technicians speak flawless English, and use it quite a bit in their work--either in dealing with reference information or to communicate across borders, even if the communication is strictly between European countries.  And, just so, developers from around the world, speaking different languages, use English as the common language for communication.  Even in China, with its dozens of different dialects which make many millions unable to communicate with each other, English is becoming the common language to bridge that gap.

Programs like Exact Audio Copy and IrfanView--both created in Germany--have English as the primary language for the program.

Even if the program--unbelievably--turned out to be developed in one non-English speaking country, there would be a super-quick translation to English, done by someone.

English became the default world language a long time ago. Someone who does not speak, read, and write English is at a severe disadvantage in today's international economy. I've heard it argued that English became dominant because of the economic and cultural power of the United States, and before it the United Kingdom. That may be true in part, but it goes further.

English is dominant because it is inherently superior to other languages. The reason becomes clear when you compare the number of words in the English language to the number of words in other languages. Other languages typically have only a fifth to a tenth the number of words available in English. Where other languages make one word serve to convey many subtly different meanings, English offers many related but nuanced words. An English speaker who has mastered the vocabulary can communicate fine distinctions more readily than a speaker of any other language.

Obviously there are exceptions to this generalization. The Eskimos, for example, supposedly have twenty (or fifty) words for snow, and some languages have very specific words for family relationships that are not available in English. In English, for example, your mother's second-oldest brother is your uncle, just as is any other sibling of your mother or father. In some languages, I'm told, there is a specific word to convey that exact relationship. But in general English offers by far the widest choice of words, and so is the most powerful and flexible language on the planet.

Americans are often criticized for their inability to speak other languages. I think provincialism has nothing to do with it. I think most Americans, reasonably enough, think they already speak the best language available and so see no point to learning another.




Taxes today and tomorrow. Ugh.


[top]

Thursday, 14 April 2005
[Daynotes Forums]    [Last Week]   [Mon]  [Tue]  [Wed]  [Thu]  [Fri]  [Sat]  [Sun]   [Next Week]    [HardwareGuys Forums]

08:32 - Taxes. Ugh.



16:21 - Tax returns are done, except for recopying them neatly, photocopying, and actually sending them off. Double ugh.

The smartest thing the government ever did was making permanent the "temporary withholding" instituted "for the duration of the emergency" during World War II. Nowadays, withholding is a matter of personal safety for politicians.

I'd like to see us return to the days of no withholding, when on 15 April everyone had to write a check for the full amount of taxes due. Make all taxes due that day: income taxes, social security (the full amount, including the concealed half), Medicare, property taxes, sales taxes, vehicle taxes, gasoline taxes, tobacco taxes, alcohol taxes, excise taxes, and so on. There would be a revolution. We peasants would storm the legislatures with pitchforks and torches held high. We'd hang politicians from lampposts. And they would deserve it.

It's no wonder that I'm an anarchist. The wonder is that everyone isn't.


[top]

Friday, 15 April 2005
[Daynotes Forums]    [Last Week]   [Mon]  [Tue]  [Wed]  [Thu]  [Fri]  [Sat]  [Sun]   [Next Week]    [HardwareGuys Forums]

10:31 - The taxes are done and off, and I have a whole year to calm down. Well, except for the quarterly estimated taxes, but sending in a few thousand dollars here and there gets me less outraged than the big tax day.



I keep reading about these girls being abducted, raped, and murdered. The perpetrator nearly always turns out to be a convicted "sex offender". And I read about calls for a national sex offender registry and think to myself that these people are missing the point entirely. We don't need local registries of sex offenders, let alone a national registry. What we need is for these bastards to be locked away permanently or, better, killed.

Part of the problem is the politically-correct "sex offender" label, which covers everything from people who haven't really done anything wrong at all to the most vicious, predatory child-rapists and murderers. An 18 year old guy who engages in voluntary sex with his 14 year old girlfriend, for example, may be lumped into the "sex offender" group with a guy who abducts, rapes, and murders a six year old girl, as may a flasher or a peeping Tom.

Considering only those criminals who truly deserve the name sex offender--the violent rapists and child molesters--it's acknowledged by everyone except the hopelessly politically correct and soft-hearted and -brained liberals that the recidivism rate is essentially 100%. If you release these people from prison, they're going to do it again. Period.

The solution is pretty obvious. Don't release them. Stop treating statutory rape as rape. It isn't. But establish Draconian penalties for true rape with violence, and enforce them without exception. At the least, any violent sex offender who is caught and convicted should be sentenced to life in prison without possibility of parole. Better still, execute him, and do it quickly. Or simply declare him an outlaw in the original wolfshead sense, and toss him to the family and friends of his victim or victims.

It seems to me that the parents of these murdered girls should have recourse against the people who freed their daughters' killers. They knew or should have known that it was inevitable that these bastards would abduct, rape, and murder again, and yet they knowingly released them into the community. The police aren't responsible. Had it been up to them, these criminals would have ceased breathing long ago. But the parole boards, courts, and legislators should be held personally responsible for these preventable tragedies.


[top]

Saturday, 16 April 2005
[Daynotes Forums]    [Last Week]   [Mon]  [Tue]  [Wed]  [Thu]  [Fri]  [Sat]  [Sun]   [Next Week]    [HardwareGuys Forums]

00:00 -



[top]

Sunday, 17 April 2005
[Daynotes Forums]    [Last Week]   [Mon]  [Tue]  [Wed]  [Thu]  [Fri]  [Sat]  [Sun]   [Next Week]    [HardwareGuys Forums]

00:00 -



[top]

Copyright © 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005 by Robert Bruce Thompson. All Rights Reserved.