- It's bulky item pickup week. Barbara is in the process of a
serious spring cleaning, and she decided she was tired of all the
clutter. I knew she was serious when she said, "We had our new roof put
on months ago, and there's still sawdust up in the attic from where
they cut the ridge vent." Uh-oh. We have an unfinished attic.
So we've pretty much emptied out the attic and basement. The image
below shows part of the stuff stacked at the curb. Some of it was
painful for me to part with. For example, there's a 20+ year old
original IBM PC-XT carton (the blue one just to left of center.)
Fortunately, she let me keep the actual 20+ year old IBM PC-XT. I told
her I had plans for it, which I do. It'll involve some internal
modifications, but eventually I plan to install some serious hardware
in that IBM PC-XT case and put it on my desk. I think it'd be pretty
cool to have what looks like a 25-year-old system as my primary
desktop, but with modern components inside. Kind of like one of those
bootlegger cars that looks like a
junker but has a 700 horsepower engine under the hood.
Speaking of main systems, I completed the migration this weekend from
my older Antec Aria system to the Antec P160 system. It is now much
quieter in my office. The loud whine of the Aria system is gone,
replaced by the nearly inaudible hum of the new system. The loudest
sound in my office now is the fan noise from the Falcon Electric on-line UPS.
(Falcon UPS's, incidentally, are going to be recommended in the new
edition of PC Hardware in a Nutshell.)
The loud whine from the Aria system isn't Antec's fault. I crammed too
much stuff into that small box, including a hot-running Prescott
Pentium 4/3.2 CPU (underclocked to 2.8), and I made the mistake of
using the stock CPU cooling fan. That fan works fine in a larger
enclosure, but in the SFF Aria, it simply has to work too hard to keep
the CPU cool. When I checked the other day, the CPU fan was running at
5,700+ RPM, which is simply too fast to be quiet.
Barbara asked what I planned to do with the system, and I told her that
I'll pull the current CPU fan and use something else. What I'd like to
do is replace the whole heatsink with a better unit, such as the Thermalright
XP-120 with a Panaflo fan. Unfortunately, a large heatsink like the
XP-120 won't fit the Aria. Antec does supply a bracket with the Aria
that connects to the side of the power supply and provides a mounting
position for a standard 80mm or 92mm fan. What I may do is remove the
current CPU fan and install a larger, slower running fan blowing down
onto the current heatsink.
I see that Kansas
is about to retreat into the Dark Ages. It looks as though the
religious whackos in Kansas will be successful in their quest to
redefine science to support their religious superstitions. That'll
certainly make things a lot easier for "science" students in Kansas.
Rather than search for a natural, scientific explanation for observed
data, they can simply attribute what they see to the supernatural.
Perhaps while they're at it they should reconsider the phlogiston
hypothesis and begin anew the search for the Philosopher's Stone.
Religion is the enemy of reason, as cases like this make clear.
- Interesting day yesterday. Interleaved with final cleanup on Astronomy Hacks, I decided to start
ripping CDs and DVDs.
The DVDs were no problem. I have DVDshrink
installed on the Windows XP box, and I simply set it to ripping. One
thing became obvious immediately. A Sempron 2800+ is much slower than
a Pentium 4/3.4, at least when it comes to video encoding. I was
re-ripping a DVD that I'd ripped previously, so the results are
directly comparable. On the Pentium 4/3.4, analysis took about 45
minutes and transcoding another 45 minutes. On the Sempron 2800+,
analysis seemed to take longer, although I wasn't timing it. That
perceived slowness did lead me to time the transcoding/compression,
though, which took more than two hours. For that task, at least, the
Pentium 4/3.4 is about three times faster than the Sempron 2800+.
I'd ripped a few of Barbara's CDs earlier, but in the transition to the
new systems those files were overlooked. Not overlooked, actually,
because I'd encoded them as 320 Kb/s MP3s, which turned out to be much
higher quality than needed for Barbara's Creative MuVo M200 portable
MP3 player. She decided that 160 Kb/s was good enough for her portable
player, so I planned to
re-rip the dozen or so CDs I'd done as a test.
So I put a CD in the drive. Xandros fired up its CD player applet,
kscd, which I immediately killed. I then fired up grip, which I'd
installed shortly before, and told it to start ripping the CD and
encoding the tracks as 160 Kb/s MP3s using LAME. The trouble was, it
was ripping at literally 0.1X. At that rate, it'd take me about 10
hours to rip one audio CD.
At first, I thought there might be a problem with that particular CD,
so I tried another. Same problem. So I rebooted the system, fired up
grip, and stuck in the original CD. This time, it ripped the CD at a
speed approaching the maximum DAE speed of the drive. Hmmm.
Once that CD finished ripping, it automatically ejected and I inserted
another CD. The Xandros CD player applet fired up automatically, and I
closed it. The grip app was hung, so I attempted to exit it without
success. I told Linux to kill the process, and then fired up grip
again. It started ripping the second CD, but again at about 0.1X. Hmm.
Obviously, there's something about the Xandros CD player applet that
conflicts with grip. I fired up Process Manager and located the CD
player applet, kscd, still resident in memory. I killed that process
and it went away. I was expecting grip to work at normal speed, but it
once again started ripping at 0.1X. Checking, I discovered that grip
was still resident in memory and showing up in Process Manager even
after I killed it. It's a process that won't die, and apparently that
instance of grip in memory is "poisoned" by the Xandros CD player app
and then refuses to exit.
I may just install Exact Audio Copy on the Windows box and use it to
rip Barbara's CDs. That box is, after all, named ripper.
- The BBC
proves it's clueless. They're running a pilot to test their new
offerings, DRM-laden TV and radio programs that can be downloaded and
watched or listened to, but not transferred electronically or written
to disc. The programs disappear after seven days.
Who in his right mind is going to sign up for this mess? Perhaps a few
technically-clueless consumers who don't have a teenager to advise
them, but that's about it. Britain is second only to Australia in
downloading television programs, and I can't imagine that this new BBC
effort will have any impact on that at all. Tastes terrible, more
What broadcasters refuse to recognize is that a television program,
once broadcast, for all intents and purposes enters the public domain.
It doesn't matter what the law says. People are reasonable about
copyright, and the overwhelming majority believe, with a great deal of
moral justification, that once a program has been made freely available
via broadcast for anyone to watch, anyone is entitled to make a
recording of that program for non-commercial use and distribute it to
others for non-commercial use.
Broadcasters better get used to it. They need to make their money on
the first showing, because that's all the money they'll ever make from
that program. And that's as it should be. Once a program has been put
on the public airways, it belongs to the public, at least in the sense
that they are entitled to record it, archive it, give copies away, make
derivation (commercial-free) versions, and so on. That's reality, no
matter how many laws claim the contrary.
They'd also better get used to the idea that the traditional commercial
is dead. It's too easy to edit out those 30-second snippets. Edited
versions with the commercials removed are posted to the Internet within
hours of the broadcast. Again, nothing is going to change that.
In fact, I'd like to see that process automated. Here's a feature that
MythTV should support in a future release: automated commercial
skipping. By that, I don't mean kludgy methods like looking for the
fade to black before a commercial. I mean that there should be central
servers that track the exact beginning and end times of commercial
breaks. The first few people who record a particular episode of a
program should be able to mark the beginning and end of the commercial
breaks simply by pressing a button on their remote controls. The PVR
software should record those time ticks as a small text file, and then
upload the data to the central servers. Any attempt to "game" the
system by broadcasters could be rendered useless with a simple
voting/trust system, whereby anomalous results were discarded. When
someone else who had recorded that particular episode began watching
it, the PVR software would access the central server, download the
commercial data for the episode, and skip the commercial automatically
Broadcasters and producers haven't come to terms with the fact that
their current economic model is obsolete. They can kick and scream all
they want, but it's gone for good. They need to implement a new model
or models, any number of which have been proposed. Instead of fighting
bittorrent, they should embrace it as a no-cost option for program
delivery. Instead of resisting new techology, they should be doing
everything they can to take advantage of it. They should welcome their
programs being freely available for download, and indeed do everything
they can to encourage that. Instead, they're stuck in the past, unable
to conceive that the economic model they've depended on for fifty years
or more is deader than the dinosaurs.
So how do they make money on freely-distributable content? I can think
of a lot of ways off the top of my head, and I haven't spent much time
thinking about it. How about turning that bug in the corner of the
screen into a revenue producer? Instead of displaying the NBC logo or
whatever, why not sell that prime real estate to advertisers? Why not
return to the days when one company sponsored an entire show? Instead
of the NBC logo, how about a Ford logo or a Wal*Mart logo? For that
matter, if one company doesn't want to pay the freight for the entire
program, why not alternate the contents of the bug, changing it at each
Then there are sponsored product placements, which the networks already
depend heavily on. When you see a can of Coke sitting on the kitchen
table in a TV program, it isn't there by accident. Coca-Cola paid for
that placement. When you see a character driving a Ford Explorer, you
can bet that Ford paid for that placement. In fact, sponsored product
placements are so widespread nowadays that if you see any identifiable
product in a television program, you can bet that the company paid for
it to appear. So why not do more with those placements? Work it out
with the TiVo and MythPC folks so that those placements become live
And people will pay for timeliness. Instead of broadcasting a new
episode, go direct-to-Internet with it. Charge a dollar for the
download (with no DRM), and millions of people will pay that dollar
just to have the episode right now. Who cares what happens to it next
week or next month? It'll be all over the web and freely downloadable.
So what? You've made your money already, both directly from the people
who paid for it and from those bugs and sponsored placements that will
The way to make money is to give people what they want, the way they
want it. People don't want commercials or DRM, and any intelligent
broadcaster or producer would recognize that. Unfortunately, the
industry is stupid. As Heinlein said, it's raining soup and they're too
stupid to hold up their bowls.
- Child pornographers are one of the lowest primate life forms.
I have nothing but contempt for people who produce or buy child
pornography. And yet I have to come to their defense. One of the
fundamental precepts of the American justice system, perhaps the fundamental precept, is that a
person charged with a crime is presumed innocent until proven guilty
beyond a reasonable doubt.
Over the last couple of decades, the government has been attempting to
turn that precept on its head, requiring those accused to prove their
innocence. Here's yet another
example. The government will no longer be required to prove that
accused child pornographers in fact used children as their subjects.
The accused pornographers will be required to prove that they didn't.
It's difficult to have any sympathy for someone accused of child
pornography, of course, but the fact remains that accused child
pornographers should have the same rights as anyone else. Namely, that
the government should be required to prove its case. The government
argues that it may be difficult or impossible to prove that the subject
of a pornographic image was in fact underage when the image was made,
or even to locate the subject. So be it. If they can't prove their
charges, the accused should be acquitted.
The mere fact that an image appears to represent an underage person
should be insufficient to gain a conviction. The prosecution should
have to produce evidence, not supposition. As things stand now, the
government can even convict someone of producing child pornography
without being required to prove that the image in question represents a
real person. Think about that. A computer-produced image that does not
represent a real person is sufficient evidence to convict someone of
child pornography. That's a violation of several Constitutional rights,
and is simply unacceptable.
We're on a very slippery slope here.
As it turns out, my new main system has two 160 GB Seagate S-ATA drives
in it. I'd forgotten that I'd installed the second drive. When I
installed Xandros on the system some weeks ago, the installation
procedure displayed both drives, but defaulted to creating partitions
only on the first drive. I'd intended to go back later and enable the
second drive, but I forgot about it until yesterday.
Xandros has no equivalent of Windows' Disk Manager, so I needed to open
a console and configure the second drive from the command line. That
was easy enough. I logged on as root, ran cfdisk /dev/sdb, partitioned
the second drive as a primary partition of type 82 (Linux), and wrote
the new partition information to the disk. I then ran mkreiserfs /dev/sdb1 to format
the new volume. After a reboot, the new volume showed up in Xandros
File Manager as a new filesystem named sdb1. I suppose I should have
given it a better name, but I couldn't think of one, so I just accepted
So now I have 320 GB of disk space on my main system. That'll come in
handy for medium-term storage of audio and video rips, as well as
providing a second spindle for backing up my working and archive data.
- O'Reilly sent me the final QC galleys for Astronomy Hacks yesterday
afternoon, so I'll be madly scanning through them to find any remaining
errors. The book has already been paginated, so I can no longer make
any significant changes, but I do still have the opportunity to make
minor correctors, fix typos, etc. This is it, though. After this pass
O'Reilly does the final preparation to send the book to the printer. It
should be in the bookstores about a month from now, give or take.
People sometimes ask me why I recommend Plextor optical drives
so highly. The question usually takes the form, "Is the Plextor
<whatever> really worth $<however-many> more than
<competing-drive>?" Well, maybe not, but probably so. If you
seldom use your optical drive, you can probably get away with using
something other than Plextor. But if you depend on your optical drives,
as I do, Plextor is the only way to go.
I've been using Plextor optical drives for years now, and I don't
remember ever having one fail. That despite the fact that I hammer them
heavily. For example, I've used one Plextor drive to rip CD-Audio discs
all day long for a couple weeks solid, with nary a problem. No hardware
problems, and clean rips every time. Similarly, I've had other Plextor
drives ripping DVDs for eight hours a day or more, day after day, with
not one problem. I can't say the same for cheaper drives, which may
survive light use, but tend to fail quickly when one hammers on them.
If you visit any commercial short-run CD/DVD duplicating company,
you'll find racks and racks of burners, all of them Plextor. There's a
reason for that. Companies who burn optical discs for a living quickly
learn which brands of optical burners are reliable and which aren't.
Which is why they all use Plextor drives, and another reason I
drives so highly.
Back in the days when optical drives were very expensive, there was
some justification for choosing Plextor drives only when their higher
reliability justified the price premium. But nowadays, with even
Plextor burners selling in the $100 range, there's seldom a good reason
choose anything less. Sure, I sometimes install an NEC ND-3520A DVD
burner rather than a Plextor PX-716A, but only when the drive is
unlikely to be used heavily, and only when the higher cost of the
Plextor is a major issue, such as for a budget system.
For corporate/business systems, it almost never makes sense to use a
cheaper drive. The high cost of an on-site visit to replace a failed
drive means that it's actually cheaper overall to buy the more
expensive Plextor drive in the first place. With a fleet of 100 or
1,000 systems, it doesn't take many service calls to exceed the total
cost differential of installing Plextors in all of the systems.
The same thinking generally holds true for SOHO systems, which often
use the optical drive for backing up. In that environment, where the
drive absolutely, positively has to work and work reliably, choosing
Plextor is a no-brainer.
So, although we have lots of other brands of optical writers around
here, we don't depend on any of them. My dead drive shelf was full of
optical writers made by other companies, or was until Barbara started
her radical cleaning efforts recently. I pitched a stack of dead
optical writers from companies like LG (never again), Lite-On, Samsung,
Toshiba, Sony, HP, and NEC, just to name some of the better-known brand
names. And not one of the dead writers was a Plextor, despite the fact
that we have more Plextor drives than all other brands combined. That
should tell you something.
- I'm still working hard on the final QC pass of Astronomy Hacks. My editor mailed
me yesterday to say that pre-orders for the book are strong, better
than expected. What's really important, of course, isn't the "sell-in"
or pre-orders, but the "sell-through", or re-orders. We won't know how
those are going until three months or so after the book hits the
stores. But I'm quite optimistic.
After I finish up this QC pass, I've essentially "signed off" on the
book. It then goes through the final production stages before being
sent to the printers. The book should hit the warehouses and
distribution centers around mid-June and the bookstores not long after
And I have to write two or three web articles for O'Reilly's web site
to help promote the book. I'll get started on those as soon as I finish
the QC pass.
1999, 2000, 2001, 2002, 2003, 2004, 2005 by Robert Bruce Thompson. All