- Happy birthday to me. I turn 34(h) today.
I spent most of Saturday reorganizing our data directories and pulling
a complete backup of the archive directories to DVD+R discs. In the
process of reorganizing the archive directories, I cut their size from
about 70 GB to about 20 GB. I did that by ruthlessly pruning,
particularly in the /archive/install directory, which held the bulk of
the data. There were several useless gigabytes of files in the
/archive/install/Microsoft directory alone, going all the way back to
Windows 3.1 for Workgroups.
I deleted such essential files as all of the NT4 service packs and
hotfixes, IE distribution files going back to IE2, Office 97, etc. etc.
In fact, after checking through the directories, I ended up simply
deleting the entire /archive/install/Microsoft directory and all its
subdirectories and files. I'll never go back to using Windows, so
what's the point of keeping all that historical gubbage?
I also deleted a ton of Linux ISOs, going back as far as Red Hat 4 and
Mandrake 5. The most recent ISOs I deleted from /archive/install/Linux
were the current ISOs for Simply MEPIS, Ubuntu, Kubuntu, and Mandriva,
all of which I'd downloaded in the last month or so. I don't need them
now, and I'm not likely to need them in the future. If I do, I'll
simply re-download them.
Once I got everything pruned and reorganized, I synched the /archive
directory on adelie,
Barbara's main system and our ad hoc
server, with /archiveback on newton,
my main system. I then pulled copies of the data from /archiveback to
DVD+R discs. That required only five discs, versus 16 or 18 before the
pruning. (I am really looking
forward to Blu-Ray writers, although I suppose the discs will be
I told K3b to verify each disc after burning, and in the process of
doing that I learned something, or actually confirmed something I
already knew. When you write thousands or tens of thousands of files to
a DVD disc using on-the-fly writing, what you get on the disc may be
only an approximation of the original file set, unless you're careful.
Of the five discs I burned, three compared without errors. Two had
compare errors, even though I was using a top-notch drive with
excellent media on a fast system. This reminds me of the bad old early
days of CD-R.
In my case, it's no big deal. I was backing up deep archive stuff,
which has been backed up before and will be backed up again. Basically,
the problem is that writable DVD lacks the massive redundancy of tape.
When you back up data to tape, a lot of the tape's capacity is used for
redundancy, including CRC data. If you subsequently restore the tape
and the drive has difficulty reading a particular block, it can
reconstruct that block from the redundant data elsewhere on the tape.
In other words, you can think of tape as having what amounts to
built-in RAID features. Writable DVD uses CRC data, but the emphasis is
on detecting errors rather than storing sufficient redundant data to
allow errors to be corrected.
Because I keep multiple copies of my backups, I have what amounts to
external redundancy. If I can't read an old file from one disc, chances
are very high that I'll be able to read it from an older backup set.
For newer files, which are created at arbitrary times and change
frequently, I have even more external redundancy. The working copies of
the files are on adelie, and
are frequently copied to newton.
All of the working data is copied to DVD+RW discs at least daily, and
the DVD+RW discs are setup in a grandmother-mother-daughter rotation.
(How's that for Politically Correct?) In other words, any file, new or
old, is likely to exist on at least half a dozen discs, so I have
little concern about losing data.
I did learn one thing, though. I cleared the on-the-fly checkbox in
K3b, which causes K3b to create an ISO image of the data to be copied
to DVD. Creating an ISO and burning it rather than individual files
allows the structure of the DVD to be set up ahead of time. Rather than
having to locate thousands of individual files, read them, and write
them one-by-one to the DVD, K3b can read and write just the one ISO
file. In theory, using an ISO image makes burn-time errors much less
likely to occur, and that seems to be true in practice. Every one of
the discs I burned from an ISO image compared without errors.
Creating the ISO image takes a few minutes, but you get back some of
that time during the actual burn because it's faster to burn one large
file than thousands of small ones. When I burned individual files using
an 8X burner and 8X media, the burn rate sometimes dropped as low as
3.67X. With the single large ISO file, the burner gets a steady stream
of data as fast as it can use it, and accordingly burns at or very near
its rated maximum speed.
On Sunday, I upgraded adelie,
Barbara's main system and our ad hoc
file server, from Xandros 2.5 Business Edition to Xandros 3.0 Business
Edition. I admit that I thought long and hard before doing this, first
because I hesitate to touch Barbara's system on general principles and
second because it is after all also our server. I think nothing of
installing experimental software, updates, and so on on any of the
other systems, including my own primary system. But Barbara's system is
an island of stability in an otherwise turbulent sea of hardware and
Still, it was time to do something. Barbara was happy with Xandros 2.5,
but it lacked a lot of features that would be useful to her, notably
transparent support for synching her PDA and working USB mass storage
support for her Creative MuVo MP3 player. It was time to get her off
the 2.4.x Xandros 2.x kernel and onto the 2.6.x Xandros 3 kernel.
- I sent off one web article to O'Reilly yesterday, Observing Deep Impact, and I'm
working on another today.
Jerry Pournelle forwarded me a copy of an email he'd received from a
high school student, with the comment that "perhaps we are not entirely
doomed..." With her permission, I'm reproducing that message here:
From: Susan Shepherd
Sent: Sunday, June 05, 2005 3:04 PM
To: Jerry Pournelle
Subject: The education mess: an insider's view
Hello. I am Susan Shepherd, a high school sophomore
from <redacted>, CA, whose school year just ended -- meaning that
anything I write to you will not come back to haunt me until the end of
August. I've been following your education discussions for about a year
and a half now, and (perhaps unsurprisingly) many of the conclusions
reached on your site were the same conclusions I came to during my
For example, reading, writing and math seem
de-emphazised, while science is gradually dumbed down for the hordes of
students who enter ninth and tenth grade not knowing d=r x t or how
cells divide. Some of this may seem like trivia, but there are many
students I have spoken to who do not know the basics of evolution, can
not explain cancer more than two months after the cancer unit has
passed, and who have no concept of what science (repeatable,
demonstrable) is or is not.
And that, sir, is from a GATE-level classroom. At
Edison, one of the top 100 high schools nationwide (at least two years
ago), there are about twice as many non-GATE or Prep students as there
are GATE students, and these know far less. I have spoken with P
students who have not read a book in their lives, who are stumped by
basic algebra, and who have no knowledge of politics, government or
history. Their grammar and spelling fall anywhere from middling to
atrocious. Some of the teachers seem to have given up, while others do
the best they can to teach classes full of 'students' who bring neither
paper nor pencil to school. Many of the parents do not seem to care.
And discipline is also an enormous problem; profanity and general
rudeness is the rule rather than the exception.
(An aside: There is a running joke about how an
observer can pick out P-students from a distance. They wear sports
jerseys, have tiny backpacks or small sports bags, and swear a lot.)
In the meantime, teachers bend over backwards to
fulfill 'the standards' and to teach what should have been taught
before. I have one English teacher in mind as a prime example of this.
We read two books and two Shakespeare plays in his class, one book/play
per quarter. The rest of the time, we read stories of various types
aloud in class. (! Isn't that for elementary school kids? Never mind;
certain kids read very slowly, so this may have been necessary.
Unfortunately.) During the entire school year, we did only three
forty-minute essay-type writing assignments. On Thursdays the teacher
brought out one of his old orange grammar books and retaught the
basics. "This is a noun/verb/apostrophe/semicolon." I lost interest in
his class rather early in the school year; perhaps from my writing you
can see why. But many other students liked his class. One of them said
that he taught her 'a lot' that she hadn't really learned before. What
he was doing ought to be taught in the third or fourth grade, not
tenth. Am I missing something?
The smartest students are seeking out
AP-classes to cover the loss. I switched to AP Biology midyear because
we were repeating seventh grade science. (Many of my former classmates
had forgotten the material, because the schools of today train students
to have short-term memory. The math classes are going ballistic over
this; students keep forgetting the basics over summer vacation.) Now
three or four freshman want to take AP Bio as sophomores. Others are
reaching out for AP English and AP European History. One or two take AP
Psychology. Why? My guess is, they aren't being stimulated enough in
their regular classes. This is No Child Left Behind, in all its macabre
One last lament. The way certain schools are run has
become ridiculous. My sister, a seventh grader, asked if she could
paint out graffiti on her locker when the school was closed. The
administrators kindly informed her that to do so would be a violation
of the Janitors' Union or some such. And when debate started over soda
machines, her middle school (Computech) replaced the soda machines with
ice cream machines. Edison did the same, but replaced two machines
instead of all of them. If this is an improvement, I must have missed
Thank you for listening. Part of this felt like a
rant, but I thought you might be intrigued by a student's perspective.
Note: if you think any of this is worthy of posting, please do. As I
said, I'm safe 'til late August, and none of my peers read your site.
Which is a shame.
To which I replied to Jerry:
What a remarkable young woman. She's
obviously destined to succeed in whatever she chooses to do.
On the other hand, people like her have never been the issue. The very
bright learn regardless of what obstacles are put in their paths. She's
clearly one of your "160+" people, who educate themselves without
regard to the quality of their schooling. It's the moderately bright
who are left behind. While the IQ 115 to 130 folks are not as important
in the overall scheme of things as the geniuses and high geniuses, they
are important nonetheless.
We are not doomed until the day that Harrison Bergeron becomes reality.
And I forwarded the message to some friends of mine with the following
Pournelle forwarded me this message. As
he says, the country is not doomed as long as it can produce
15-year-old girls like this.
To which one of them, an organic chemistry professor at Wake Forest
I'm not surprised. For all the
bitching and moaning about the system and poorly educated students,
there are some outstanding students. I've taught a number of students
who I have no doubt are more intelligent and will do better than
I. The real shame is that any students who have the intelligence,
but not the home life or opportunities aren't being picked up by the
schools and "rescued." I hope Jerry will respond to this girl and tell
her not to give in. Students who are independent and willing to put up
with some crap can prosper still.
To which I replied:
Alas, these students are outstanding
despite the schools, not because of them. Extremely bright kids always
learn, on their own if nothing else. It's the average kids and the
middling-bright ones, the IQ 115 to 130 group, whose needs are being
ignored by the current system.
Ultimately that will come back to haunt us. It's the geniuses who drive
things, of course. They're the ones who'll discover the cure for
cancer, found new industries, make great leaps in science, and so on.
They're the generals.
But no army can succeed with only top-notch generals. It needs
competent sergeants and junior officers, and it's exactly that group
that our current schools are ignoring. Over the next generation, we'll
produce just as many first-rate scientists, engineers, and physicians
as we always have. But I wonder where the lab technicians, draftsmen,
and nurses are going to come from.
In response to my request for permission to publish her message, Susan
Re: your email to Jerry Pournelle
Mon, 6 Jun 2005 23:02:14 EDT
Robert Bruce Thompson
Hello, Mr. Thompson. This is Susan Shepherd. The way I figure it, if
Mr. Pournelle forwarded that letter to you then you have to be of good
character, and I'd be honored to have my letter posted on your site as
well. It might be best if you removed the references to the city I live
in as well as my e-mail address, but I'd like to keep my name and the
names of the school in the story. I sincerely doubt that any harm could
come from administrators seeing the letter; there isn't much a school
wants to do to a student with as high a GPA as mine.
Thank you very much for offering to post my letter. I was pretty
surprised at the response I got, but I am glad about it too. It might
make people think a little more about what is really going on in
schools these days.
And this from a subscriber:
You and Jerry aren't listed in the top blog lists
Date: Sun, 05
Jun 2005 13:15:11 -0500
I wrote a
post on my blog which is a response to your question as to why I think
you and Jerry weren't on the recent list of blogs which was mentioned
on Slashdot. I have included a link to the post and the actual text
Bob Thompson and Jerry Pournelle should show up on
all of the lists of top bloggers. The reason that Bob and Jerry don’t
get listed in any of the compiled lists of bloggers that show up every
so often is quite simple. The authors of these lists use statistics
from Sitemeter, Technorati or some other
third party. Since I have never seen a Sitemeter or Technorati logo on
either Bob’s or Jerry’s site, I am going to assume that neither of them
are listed with either service. Also it should be noted that the
Sitemeter visit counts are notorius for being low compared to the visit
counts by programs which analyze logfiles like Webalizer. Sitemeter
reports just over 5000 visits to this blog in the last year. Webalizer
reports more visits than that to this blog in the last two months.
To add your
blog to Sitemeter, follow the instructions here. To add your blog
to Technorati, follow the instructions here. Sitemeter and
Technorati both require adding code to your site. With Technorati, you
only on the main page since I switched to Wordpress 1.5 shortly after
it came out.
Thanks. I don't think Jerry worries about what the blogger community
thinks of his page counts, and I know I don't care what it thinks about
Intererestingly, I see that Pournelle is now included on the Blogebrity B-list. (I'm
sure he'll be insulted, considering some of the names on their A-list.)
They still don't have me listed, even as a C-lister, although my two
primary sites generated 2,150 visitors and 10,620 page reads yesterday.
Oh, well. Pournelle's site generated 5,872 visitors yesterday and
21,519 page reads, and that buys him only B-list status.
- Here is--dare I say it?--a miscarriage
of justice. Texas has convicted a man, 19, of murder and sentenced
him to life in prison for helping his girlfriend, 17, miscarry by
stepping on her stomach while she punched herself. The young woman
faced no charges.
Regardless of one's feelings about abortion, this disparity in
treatment is shocking. The young man and young woman cooperated to
achieve a common goal. Neither is, or should be, any more culpable
under the law than the other. And yet, he now faces life in prison (and
could have faced the death penalty) while she was not even charged with
The law is an ass.
- I sent the following comments to Steven Vaughan-Nichols
response to his column
about Aptel being the death of desktop Linux.
Your columns are usually spot-on, but I
think you missed the boat
big-time on this one.
Apple isn't entering the generic X86 market. If they were, there might
be some cause for concern among desktop Linux vendors and users.
(Although Apple might take quite a while to have anything like
reasonable driver support for the plethora of PC components and myriad
configurations; they're used to running their OS on what amounts to a
But Apple's announcement yesterday really changes nothing. They've gone
from a totally proprietary platform that happens to use the PowerPC
processor to a totally proprietary platform that happens to run an
processor. You won't see OS X running on generic Intel boxes. Apple has
said as much, and I believe them.
I think this whole thing is much ado about nothing. Apple's market
won't increase by much, if at all. If anything, Apple has shot itself
the foot, as many Apple customers will hold off buying a PowerPC-based
Apple and wait for the Intel-based models. I expect to see Apple's
market share drop for the next two years or so, although it will
probably eventually recover to today's levels. But Apple systems will
always be as expensive as ever, and people will be no more likely to
an Intel-based Apple system than they are to buy a PowerPC-based one.
All of that said, I agree that Linux needs to get its butt in gear to
take advantage of Shorthorn delays. Unlike you, though, I don't see the
current Linux desktops as inferior to OS X, let alone XP. A year ago, I
converted all of our systems from Windows to Xandros, and I've just
recently completed the transition to Xandros 3 Business Edition. The
problem isn't Linux, per se. It's application availability.
If companies like Red Hat and IBM that have a major stake in open
want to promote Linux use on the desktop, what they should be doing is
throwing substantial resources at key applications categories and back
office stuff. We need a fully-compatible, interoperable OSS alternative
to the Exchange Server/Outlook pair, for example. We need to have a lot
more developers working on OpenOffice.org, Mozilla/Firefox/Thunderbird,
Evolution, and similar desktop apps. We need a true replacement for
FrontPage. And so on. The Linux desktop itself is good enough now.
And, in response to his reply:
You may be right. I hope not.
I think Jobs values the bird-in-hand margins from Mac hardware sales,
and fears the bird-in-bush margins from selling tens of millions of
copies to generic Intel box owners will be illusory when the additional
development and support costs of entering the wild-and-woolly anarchy
of having to support hundreds of video adapters, countless chipsets and
processors, and so on are taken into account.
Either way, I think it's win-win for Linux. It's not all that much
harder for application vendors to support Lintel if they're supporting
Aptel, so if Apple stays proprietary desktop Linux should see a
flowering of commercial applications. If Apple does jump in with both
feet and makes OS X available on generic Intel boxes, Microsoft's
Windows monopoly goes away fast. All of a sudden, it's a much more
level playing field, with three reasonably equal competitors instead of
Microsoft and the two dwarfs."
Obviously, the shift to Intel is going to cost Apple market share over
the next 12 to 18 months. Jerry's comments about his own hesitance are
representative of what many Apple users will be thinking. Against that,
one has to weigh the increasing market share of notebook computers and
the benefit that Apple will receive from using Intel notebook CPUs.
That should boost their market share somewhat, but I think on balance
Apple's overall market share in two years is likely to be no higher
than it is now.
- Barbara comes home tonight. The dogs and I will celebrate.
Astronomy Hacks went to
bindery yesterday, which means it'll soon be in the distribution center
and should begin hitting the bookstores in a couple of weeks. O'Reilly
has agreed to do as they've done for prior books and send out some
number of review copies to my readers who express a willingness to read
the book and review it on Amazon.com and other on-line review sites.
In the past, I've made the review copy offer open only to subscribers.
This time, because the subject matter of the book is astronomy rather
than computers, I'm going to open it up to include all my readers, with
the proviso that if the number of people who respond is larger than the
number of review copies available, subscribers get preference.
If you'd like to receive a review copy of Astronomy Hacks, with the
understanding that you agree to read it and review it on Amazon.com and
whatever other on-line review sites you prefer, please send email to
thompson at wsal dot org with the subject line:
Astronomy Hacks review copy request
Please include the following information in the body of the message:
I'll filter these to a bin automatically, so I won't see any additional
comments you make, at least until I batch-process the requests. If you
have something to say to me, please send it as a separate message.
- Your full name
- Your full mailing address (not a P.O. Box; we need a street
- Your telephone number (UPS & FedEx require this)
It's been all over the tech news sites this week that the author of
DVDDecrypter got a take-down notice from an anonymous company, widely
believed to be Sony. He can't fight it. As he says, he has only 50
pounds in his bank account, so he'd have no chance against hordes of
corporate lawyers. He's gone belly-up entirely, and I don't blame him.
He's given his site to the company that issued the take-down notice and
walked away from his software entirely.
Still, the latest version of DVDDecrypter is all over the Internet, and
doubtlessly stored on millions of hard drives. I have no need for it,
but if you do it'd be as well to grab a copy while it's still easy to
find. I have no idea who sent me the following message, and I've
deleted and purged it on general principles.
Date: Wed, 8
Jun 2005 21:55:40 -0500
available at a Russian site
advising turning OFF the automatic update feature, as the
turned his whole site over to whoever has forced him on
those seeking automatic updates may be sniffed out.
I wonder how long it'll be before the OSS community picks up the ball
and runs with it. The author of DVDDecrypter has turned over his source
code to the company that issued the take-down order, but I'd imagine
there's at least one copy of it somewhere. And one copy is all it
takes. When are companies going to learn that the Internet regards
censorship as damage and automatically routes around damage?
- Barbara's home. The dogs did their little circle-dance to
greet her. Everyone's happy. We're having a steak dinner tonight,
which is better than the sandwiches and TV dinners I survived on while
Barbara was gone. Our friends Mary Chervenak and Paul Jones took pity
on me Tuesday and took me to a Chinese place for dinner. Otherwise, I
might not have survived.
Barbara is sitting in her office right now playing with her new email
system. I upgraded her Xandros 2.5 Business Edition system to Xandros
3.0 BE while she was gone. Alas, X3BE changed from Mozilla Suite to
Firefox and (optionally) Thunderbird. I don't much care for the Firefox
browser compared to the full Mozilla browser, but at least Firefox can
be enhanced with plug-ins to have most of the same functionality as the
full Mozilla browser. Thunderbird is a different story. It's a
lightweight mail client, much like Outlook Express, and there's not
much that can be done to fix it.
But with Mozilla Suite deprecated, I had to choose something. Firefox
seems a reasonable browser choice, but for mail I needed something with
at least most of the functionality of Mozilla Mail. I also wanted an
integrated PIM, which basically left me with two choices. Evolution is
very Outlook-like, but I didn't think much of it when I tested it in
the past. On Ron Morse's advice, I took a look at Kontact, which
integrates the Korganizer PIM with KMail.
Barbara's been using Kontact/KMail for at least five minutes now, and I
haven't heard any screams, so perhaps it'll be all right. I haven't
upgraded my own primary desktop to Xandros 3 BE, but when I do I'll
also need to migrate my mail. I'd planned to install the full Mozilla
Suite, but that apparently is not an option with X3BE, or at least not
an option I want to risk. When I opened Xandros Networks and searched
for the Mozilla Suite, it was listed, but said "Not installable from
current application sources." I don't want to risk borking a fresh
install of X3BE, and the Mozilla Suite is no longer actively supported
anyway, so now is a reasonable time to make a change.
It's just that I hate migrating my mail to a new client. I'll probably
archive all my current mail to the archive folders and then import
those. That'll give me all my mail back to the time I changed over to
Mozilla Mail. Counting listserv traffic, that's probably a million
messages, give or take.
I wasn't entirely clear on how to accommodate multiple mail accounts
with Kmail, but I got it worked out on Barbara's system. Basically, you
set up identities first. In Barbara's case, she actually does have two
real mail accounts. All of her mail on our server comes to her
researchsolutions.net mailbox. Mail sent to *@fritchman.com and other
domain names is autoforwarded to her researchsolutions.net mailbox.
Then she has a Roadrunner account that she also uses, which is a true
The concern was that when she receives mail addressed to one of her
fritchman.com accounts, she wants her reply to be from that email
address and with a signature reflecting that email address; when she
gets mail to her researchsolutions.net account, she wants her reply to
be from that address with a matching signature, and so on. That turns
out to be easy to accomodate in KMail. I simply created an identity for
each of her email addresses, and then created a top-level folder for
each identity. Part of the folder creation process allows you to map
the folder to an identity. Replies sent from that folder automatically
are set to come from the email address/identity associated with that
folder, and with an appropriate signature. When she sends new mail, she
can choose the identity from which it's sent.
My major remaining concern is archiving. I haven't discovered yet how
to move older mail out of the main working data directories to an
archive area, if indeed that is even possible. Also, a minor annoyance
is that Kmail stores its data in the /mail directory under the user
home directory. I haven't figured out how to change the location of the
mail folders, if indeed that is even possible. I want to store
Barbara's mail in the home/thompson/usr/barbara/mail directory on my
main system, newton. That
directory is actually a network share to /home/barbara/usr on adelie, Barbara's main system. With
it mapped, I can easily back up her data to DVD as a part of my regular
backup routine on my local system. I can still back up her mail data,
of course, but I'll have to map a directory on my local system to
/home/barbara/mail on adelie,
and then remember to include it in the backup set.
- I am really, Really, REALLY tired of getting phishing emails
A quick check check of the message source reveals, of course, that the
supposedly secure https:// PayPal URL in fact points to an http://
numeric IP address located who knows where. Not that it was necessary
to do even that minor investigation to verify that this is a phish. The
message text is sufficiently garbled to make it clear that it was
written by someone who is not fluent in English.
A year or 18 months ago, when phishing first started to hit big-time,
Pournelle and I were talking about it. We agreed that the problem with
phishing, as with the Nigerian scam and many other scams, is that the
victims are self-selecting and that the majority of people aren't
stupid enough to fall for them. If a phisher sends out one million
emails and gets 100 responses, he can be pretty sure that those 100
responses include valid information. And he'll make use of that
information, because law enforcement can't do much to stop him, what
with national boundaries and all.
But what if instead of only 100 responses, the phisher got 1,000,
10,000, or even 100,000 responses? Only 100 of them would contain valid
information, but how would he know which 100? Nowadays, when I get a
phishing scam like this one, I'm very, very tempted to click on the
link and give the guy complete garbage. If everyone did that, phishing
would quickly disappear entirely. There'd be no point to phishing if
there was no way to separate the few valid responses from the many
1999, 2000, 2001, 2002, 2003, 2004, 2005 by Robert Bruce Thompson. All