Monday, 13 December 2004
[Last
Week]
[Monday] [Tuesday] [Wednesday]
[Thursday]
[Friday] [Saturday] [Sunday]
[Next Week]
[Daynotes Journal
Forums] [HardwareGuys.com
Forums] [TechnoMayhem.com
Forums]
{Five
Years
Ago Today]
09:25
- Barbara returned late last night from her weekend bus tour to
South Carolina with her family. The dogs and I greeted her joyously.
She had so much fun that she plans to go again next year. When we heard
that, the dogs' ears drooped. My ears drooped, too. It's just not the
same when Barbara's gone.
I'm pretty busy from now until year-end, so there won't be much posted
here. I have deadlines approaching, year-end tax/business stuff to do,
the Winter Solstice holiday itself to prepare for and suffer through,
and so on. So don't expect much here until after the first of the year.
[Top]
Tuesday, 14 December 2004
[Last
Week]
[Monday] [Tuesday] [Wednesday]
[Thursday] [Friday] [Saturday]
[Sunday] [Next Week]
[Daynotes Journal
Forums] [HardwareGuys.com
Forums] [TechnoMayhem.com
Forums]
{Five
Years
Ago Today]
10:14
- I spent the afternoon over at Paul and Mary's house fettling
their network. Mary was still recovering from running a marathon down
in South Carolina last weekend. She ran an excellent race, and came in
at 3:31. Hurray for Mary! She kept up an 8-minute mile pace for the
entire race. What Mary does is beyond my imagining. Nowadays, I
couldn't run one 8-minute mile, let alone 26 of them in a row.
Barbara went to the 5:30 gym class and then picked up pizza and brought
it over to Paul and Mary's house. After dinner, we sat around for a
while discussing ideas for the astronomy book that Barbara and I plan
to write, and then headed home for an early evening.
Fred Reed has posted an interesting column about the
death of newspapers. Coincidentally the business section of our
paper this morning had an article about the decline in newspaper
advertising.
Frankly, I'm not sure why we continue to take the local paper. There's
not much in it worth reading. The national coverage is usually just a
rehash of stuff I'd read on the web the day before. The local coverage
is usually of little or no interest to me. Anything local I care about
I'll hear about from other people anyway. I read Dilbert, which I can
find on the web, and the bridge column, which I'm sure I could find on
the web if I cared to look. We read Dave Barry's column, which again is
available elsewhere. TV listings? The web is more up-to-date, and
anyway we watch very little TV. We have the Weather Channel and the web
for weather. There's just nothing in the local paper that makes it
worth getting.
Newspapers were dying before the Internet became ubiquitous, and the
web is hastening their demise. In 10 years, I suspect most local
newspapers will have folded.
12:13
- A tractor-trailer just showed up out front with a 260-pound
skid of UPS equipment from Falcon
Electric. Jerry Pournelle has sworn by Falcon UPSs for years, and
after talking at some length to a couple of the folks at Falcon I
understand why.
Years ago, there were a few engineering-driven companies that built the
best possible products, regardless of cost. Digital Equipment
Corporation, Hewlett-Packard, and Compaq come to mind. They used
milspec components, gold-plated connectors, and whatever else it took
to build products as perfectly as humanly possible. If there was an
alternative way to do something that was 10% better and cost twice as
much, they spent the extra money.
In those companies, engineering was king, and marketing was, if
anything, an afterthought. HP and DEC, for example, were both famous
for expecting their products to sell on their own merits and by word of
mouth. If you didn't appreciate how good they were, you didn't deserve
to have them.
Alas, most of those companies are gone now, or as with HP, have morphed
beyond recognition. Nowadays, if an alternative method is half as good
and costs $0.10 less, nearly all companies use the cheap method without
a second thought. So it was a pleasure for me to talk to the folks at
Falcon Electric. It's an old-fashioned company in the best sense. The
engineers still rule there.
During our conversation, I mentioned in passing that I'd noticed NewEgg
didn't carry Falcon Electric UPSs, and asked where my readers might buy
them. They were clearly puzzled that I expected their products to be
available in mass-market channels. "Well, we do have distributors,"
they said, "but you can just buy our products directly from us."
I love it. A company that knows its products are so good that people
will come looking for them. Glancing at the types of customers who buy
Falcon Electric products, it's easy to see that they're justified in
that belief. Falcon sells commercial-grade power protection equipment
to large corporations, the military, telephone companies, and so on. In
fact, to any organization that requires absolute reliability in power
protection. There's no consumer-grade stuff on the Falcon product list.
So I took a look at the Falcon price list, expecting to clutch my chest
and gasp. I was expecting Rolls-Royce prices, but I was pleasantly
surprised. Falcon equipment isn't inexpensive by any means, but neither
is it all that much more expensive than consumer-grade stuff with
similar capacities.
I have three Falcon units in for testing, a 1 kVA line-interactive
unit, and two on-line units, a 1 kVA and a 1.5 kVA. I plan to wring
them out over the coming weeks. Stay tuned.
I'll be very busy for the next few days, reviewing the QC1 draft of the
new PC Hardware Buyers' Guide
book and incorporating edits. O'Reilly has this one on the fast track,
and I want to get it as correct as possible and as current as possible
before it goes to the printer. There'll be one more QC pass in early
January, and then it'll be off to the printer.
[Top]
Wednesday, 15 December 2004
[Last
Week]
[Monday] [Tuesday]
[Wednesday] [Thursday] [Friday]
[Saturday]
[Sunday] [Next Week]
[Daynotes Journal
Forums] [HardwareGuys.com
Forums] [TechnoMayhem.com
Forums]
{Five
Years Ago Today]
8:39
- Someone emailed me the other day to ask me to explain why the
arguments of the Intelligent Design folks are pseudoscientific and
wrong. Basically, it's because they make a fundamentally flawed
assumption, which is that the current state of things is somehow
special. For example, Fred Reed, for whom I otherwise have a lot of
respect, writes:
Consider a thickish book
of, say, 200,000 words. By the newspaper estimate that there are on
average five letters per word, that's a million letters. What's the
likelihood that our monkey, typing randomly (ignoring upper case and
punctuation) will get the book in a given string of a million letters?
He has a 1/26 chance of getting
the first letter, times a 1/26 chance
of the second, and so on. The chance of getting the book in a million
characters is one in 26 to the millionth power. I don't have a
calculator handy, but we can get an approximation. Since 26 = 10
exp(log 26), then 26 exp(1,000,000) = 10 exp(log 26 x 1,000,000) .
Since log 10 = 1 and log 100 = 2, log 26 has to be between, somewhere
on the low end. Call it 1.2.
The monkey thus has one chance in
12 followed by 1,000,000 zeros. (OK,
999,999 for the picky.) That's what mathematicians call a BLG
(Brutishly Large Number). For practical purposes, one divided by that
rascal is zero. If you had a billion billion monkeys (more monkeys than
I want) typing a billion billion letters a second, for a billion
billion times the estimated age of the universe (10 exp 18 seconds is
commonly given), the chance of getting the book would still be
essentially zero.
He misses the point entirely, which is not surprising because
statistics and probability are slippery things. The question is not the
probability of his monkeys producing a
particular book, which is to say the current state of things,
but a book, which needn't be
in English or indeed recognizable by us as a book at all.
Let me give a concrete example. I used to play bridge, in which a deck
of 52 cards is dealt to four players, giving each player a hand of 13
cards. An old practical joke amongst bridge players is to stack the
deck so as to give an unsuspecting victim a hand of 13 spades. In a
truly random deal, the odds against him picking up a hand that contains
all 13 spades are more than 645 billion to 1. So, if a player picks up
a hand of 13 spades he suspects, almost certainly correctly, that
someone has stacked the deck.
In a nutshell, this is the argument of Intelligent Design. That the
likelihood of picking up a hand of 13 spades is so vanishingly small
that when we pick up such a hand it is impossible that we could have
received that hand as a result of a random deal. Some unseen prime
mover must be behind it. Superficially, that is a reasonable argument.
The problem is, it has a major logical flaw. To understand why, you
have to understand that the probability of receiving any specific hand
of 13 cards is exactly the same as the probability of receiving a hand
containing 13 spades. For example, in a random deal I am no more likely
to pick up
S: Q-9-4-2
H: K-J-7
D: A-10-8-3
C: 9-6
than I am to pick up
S: A-K-Q-J-10-9-8-7-6-5-4-3-2
By the arguments of Intelligent Design, therefore, if I sit down at a
bridge table, it is impossible for me to be randomly dealt any hand at
all, because the probability against my receiving that specific hand is
so vanishingly small. In other words, the logical flaw in the arguments
of Intelligent Design folks is that they assume the way things are is
the way they had to be. Their argument is circular, because it
presupposes a hidden guiding hand must exist for any result at all to
occur.
The Intelligent Design folks make much of the argument that evolution
is "only a theory", with the implication that a theory is only a guess.
In science, a theory is far from a guess; it is the next best thing to
a fact. Scientists will not state that evolution is a fact, because
scientists place great value on facts. For something to be accepted as
a fact, it must pass a very high bar. In theory, for example, the sun
will rise tomorrow. But that is not a fact until it has been proven to
have occurred.
The theory of evolution is supported by all of the facts we have. If
someone discovered and confirmed just one fact that was fundamentally
incompatible with the theory of evolution, that theory would be
discarded. Scientists would devise new hypotheses to explain what we
observe and would then test those hypotheses against known facts. But
unless and until such a fact is discovered, the theory of evolution is
the best and far most likely explanation we have for the facts we
observe. If you think of evolution as being just a guess, you're simply
proving that you understand nothing about science.
[Top]
Thursday,
16 December 2004
[Last
Week]
[Monday] [Tuesday]
[Wednesday] [Thursday] [Friday]
[Saturday]
[Sunday] [Next Week]
[Daynotes Journal
Forums] [HardwareGuys.com
Forums] [TechnoMayhem.com
Forums]
{Five
Years
Ago Today]
10:27
- I don't usually care for Christmas stories, but here's one worth reading.
It's the story of Ebenezer Broadcaster.
As the story says, "Mass marketing died when the Internet was born, and
media is now all about consumer choice." Broadcasting is deader than
buggy whips. It just doesn't realize it yet. Oh, the entrenched
broadcasting interests will struggle mightily to maintain the status
quo ante, but they'll inevitably lose. As Canute learned, no one can
hold back the tide.
Roadrunner just announced that it's boosting download speed caps for
residential users from 3 Mb/s to 5 Mb/s in January, with business users
jumping from 6 Mb/s to 8 Mb/s. Internet2 can transfer an entire DVD in
four seconds. Technologies like WiMax hold the promise of delivering
inexpensive, fast broadband access even in thinly populated areas.
Soon, anyone anywhere who wants a fast broadband connection will be
able to have one at an affordable price. That's the death knell for
broadcasting.
Every evening now, our Roadrunner connection slows down. As more and
more kids discover BitTorrent, they're spending their evenings
searching Suprnova.org for torrents to download the latest movies and
TV programs. Roadrunner continues to build out their networks, so the
slowdown is temporary, but what those kids are doing is permanent.
They've gotten out of the habit of watching broadcast television,
instead looking to the Internet to deliver entertainment. They'll never
return, nor will their children or their children's children. The
phrase "network television" has taken on a new meaning.
Companies that are committed to the old order are doomed, no matter how
many congressmen and judges they bribe or how many laws they buy.
Wishing things were different isn't a viable business plan. It's the
newer, smaller companies that are fast on their feet and willing to
recognize and exploit the new reality that will prosper.
Ubiquitous high-speed Internet access is the largest agent for change
we've ever seen, not just in how we play but in how we work, at least
so-called "knowledge workers." Many of us still commute to work every
morning, to a downtown office, but that is already changing and the
pace can only accelerate.
This idea of centralization developed in the late 19th century, when
typewriters were costly and telephones rare. At that time,
centralization made sense, because it was cheaper and more efficient to
move people to a central location where they could share scarce,
expensive resources than it would have been to provide those resources
in employees' homes. Nowadays, the converse is true, and yet many of us
still drive to work every morning.
The cost of that is outrageous. We waste millions of man-hours per day
on needless commuting and spend trillions of dollars on the
infrastructure to get commuters to and from work and to provide
workplaces for them. This morning, when I finished reading the
newspaper in the den, I did my usual 15-second commute to my office.
Barbara left to drive the 10 miles or so to the law firm where she
works downtown. She starts work very early to avoid the morning rush
and find a good parking place, which is telling.
Even disregarding Barbara's commuting costs in time, gasoline, vehicle
depreciation, and parking fees, the equation simply doesn't balance.
Her law firm spends thousands of dollars per year to provide a
workplace for her, which amounts to a desk and chair, a computer, a
network connection, and a telephone. She has all of that at home.
How difficult would it be for her firm to virtualize the
centralization? Not very. Instead of paying thousands of dollars per
year for space for her, all they need do is provide a fast Internet
connection with VPN, a computer (which they're providing anyway), and
some minor office equipment. They could also provide an IP phone, which
could be virtualized as a part of their internal phone system. If they
want face-to-face conversations, install a webcam. That's it. Multiply
those cost savings by the number of staff they employ and it comes to a
huge number.
Every study of telecommuting I've seen says that the upsides far
outweigh the downsides. Sure, her firm might have to hire a few more IT
people, because they'd be making house calls. So what? They could cut
down on the amount of expensive downtown office space they pay for by
about 90%. Their space needs would be minimal. Some staff and client
meeting rooms, a few offices, facilities for IT equipment, and that's
about it.
Telecommuters are happier and more productive than people who commute
to offices. They get more work done, get it done faster, and that work
is of higher quality. Telecommuters cost the company less, much less,
than office-bound workers. Some companies fear that telecommuters will
be harder to supervise and less productive than office-bound workers,
but the converse has proven to be true. Some would-be telecommuters
fear the "out of sight, out of mind" issue, thinking that by missing
out on water-cooler conversations they will become isolated and perhaps
overlooked at salary-review and promotion time. Again, that fear has
never materialized. A good worker is a good worker, period, no matter
where he happens to do his work.
In fact, I wonder if there's a business opportunity in office
outsourcing. I can visualize a company that hosts businesses that are
changing to a telecommuting model. Provide office space dedicated to
the needs of telecommuting-based businesses. Small suites, some with
dedicated meeting rooms and others for those satisfied with common
meeting rooms. Centralized server and telecomm facilities that would
eliminate the need for companies to employ much dedicated IT staff.
Redundant datacomm set up to meet the requirements of a large number of
telecommuters. Technicians who could be dispatched to the homes of
telecommuters who had computer or telecomm problems. Consultants who
would help companies design and deploy a comprehensive telecommuting
plan. And so on.
It's interesting to think about all this, but it'll be more interesting
to actually do something about it. If I can get Barbara comfortable
with the idea, I plan to cancel our cable television and telephone
service and run our house on IP. That won't happen immediately, but it
is one of my long-term goals. We'll see how it all works out.
Sometimes I just want to strangle someone...
Mozilla
Firefox
This is one more example, albeit trivial, of why Firefox annoys me.
What possible reason
could there be for arbitrarily switching the buttons? Confusing
long-time Mozilla Browser users?
And speaking of annoyances, one thing that really outrages me is when
software attempts to force me to do something its way because the
programmers have certain beliefs. The first time I ran into that was
with Eudora, which refused to allow me to request a return
receipt as the default action. I needed return receipts for what I was
doing at the time, but the only way to get them with Eudora was to
request a return receipt on a message-by-message basis. Eudora even
provided a sanctimonious explanation in the help file, saying that it
was a waste of bandwidth to request return receipts. Give me a break.
That was enough to make me swear off using Eudora. I've never used it
since, and I've recommended against it any time anyone asked me.
I've run into a similar situation with Mozilla Composer and its
derivative N|vu. When I insert an image, as I just did above, these
programs prompt me to enter alternative text. I don't want to enter
alternative text. I don't care if my site complies with government or
other standards for accessibility. I don't have the time or the
inclination to sit there thinking up alternative text entries. But
Composer and N|vu default to prompting for alternative text, and
there's no option to change that default.
If that were merely an oversight, I'd be annoyed but not outraged. But
I'm sure it's no oversight. Someone decided it would be a Good Thing to
help the blind by making it difficult for everyone to avoid entering
alternative text. In other words, that someone decided that it was
acceptable to make my life more difficult to benefit a group for which
he feels sympathy. And I deeply resent that.
Set the default to require alternative text. Fine. But if I once mark
the radio button for "Don't use alternative text", that should change
the default automatically and permanently. To force me to take that
extra step each time I insert an image is simply obnoxious. Enough so
that I'm looking for another HTML editor.
And here's a response from an anonymous correspondent, who points out
something I should have mentioned about Intelligent Design.
Excellent
argument. The only other thing I would point out is that
disproving a hypothesis or a theory is not proof that an alternative is
correct. Supporters of Intelligent Design seem to think that
poking holes in the theory of evolution proves them correct, when in
truth, it says nothing at all about the validity of Intelligent Design.
That's what happens when I sit down and start writing off the top of my
head. I forget to mention important points. Thanks for the catch.
[Top]
Friday,
17 December 2004
[Last
Week]
[Monday] [Tuesday]
[Wednesday] [Thursday]
[Friday]
[Saturday] [Sunday] [Next
Week]
[Daynotes Journal
Forums] [HardwareGuys.com
Forums] [TechnoMayhem.com
Forums]
{Five
Years
Ago Today]
10:05
- I finished up the QC1 pass through the PC Hardware Buyer's Guide book
yesterday and sent my comments off to O'Reilly. I'm still awaiting
comments from my "kitchen cabinet", but I suspect there won't be too
many.
I was looking back over my goals for 2004. Some I completed partially,
some I didn't even start on, and a few I finished completely. Notable
among the latter was one of my more important goals: "Migrate to Linux:
This Year For Sure". I completed that one in spades. All of our primary
desktop systems, including Barbara's, are now running Xandros Desktop
Linux. Other than my notebook, a PVR system, a temporary file server,
and several test-bed systems, we have no Windows systems in the house.
The notebook will be converted to Linux as soon as there's a workable
Linux version of the Cartes
du Ciel planetarium program available for Linux. Of course, there's
always Xephem.
It's powerful, but I just can't get used to the Motif interface.
Perhaps I should make more of an effort to learn Xephem, because at the
rate the Cartes du Ciel Linux conversion is going, it may be years
before there's a release version.
The PVR system (actually, there are two) will probably be converted to
Linux as well. I haven't looked at MythPC, Freevo, or any of the other
Linux-based PVR packages lately, but I'm willing to bet they've
improved since I last looked at them. Of course, now that Barbara is
down to watching exactly one weekly program, Left Wing, it may not make sense to
have a PVR system at all, other than to write about. Especially since I
want to drop our cable TV service. Perhaps I should build an automated
BitTorrent-based PVR system. Nah, I guess that'd get the MPAA upset
with me.
The file server is still running Windows 2000 from inertia more than
anything else. I'll probably convert it to Xandros before the end of
the year. Either that, or I may decide we don't really need a file
server at all. In that case, I'll strip it down to bare metal, install
SuSE, and donate it to Senior Services. They'd probably appreciate it.
It's a Pentium 4 box with 512 MB of RAM and a DDS-3 tape drive. It'd
make a nice small server for them.
One of my goals for 2005 is to quietize this place. I'm tired of the
whine of CPU coolers and exhaust fans. For a lot of our systems, that
won't require much more than replacing stock CPU coolers with quiet
units from Thermalright or Zalman. For others, I'll have to do a bit
more work.
I have three books on the schedule for 2005, so it'll be a busy year.
[Top]
Saturday,
18 December 2004
[Last
Week]
[Monday] [Tuesday]
[Wednesday] [Thursday]
[Friday]
[Saturday] [Sunday] [Next
Week]
[Daynotes Journal
Forums] [HardwareGuys.com
Forums] [TechnoMayhem.com
Forums]
{Five
Years
Ago Today]
[Top]
Sunday,
19 December 2004
[Last
Week]
[Monday] [Tuesday]
[Wednesday] [Thursday]
[Friday]
[Saturday] [Sunday] [Next
Week]
[Daynotes Journal
Forums] [HardwareGuys.com
Forums] [TechnoMayhem.com
Forums]
{Five
Years
Ago Today]
09:28
- I was just checking my web access logs this morning, and I see
that for December to date only about a third of my traffic reports the
agent as Internet Explorer. The remainder is almost entirely Mozilla
and Firefox. What really amazed me, though, is that about a sixth of my
traffic comes from Linux desktops.
When you consider that a lot of my visitors browse this site from their
offices rather than their homes, which usually means they're running
Windows/IE, those numbers are truly staggering.
[Top]
Copyright
© 1998, 1999, 2000, 2001, 2002, 2003, 2004 by Robert Bruce
Thompson. All Rights Reserved.