10:00
- The taxes are done and off. I took Sunday off.
I've shifted gears from astronomy to PCs. I'm spending lots of time
emailing hardware vendors and on the phone with them, trying to get
samples of the latest and greatest stuff. There are some significant
changes coming this summer, not the least of which is Conroe, Intel's
new desktop processor.
Intel has had a tough time of it the last 18 months or so. Prescott was
disastrous. It was expected to scale to at least 5 GHz, and possibly 10
GHz. It ran out of headroom at 4 GHz. That, combined with power
consumption of up to 130W, made the Pentium 4 a sitting duck for AMD's
Athlon 64, which provided the same or better performance at half the
power consumption. With Conroe, that's all about to change. Conroe will
reclaim the performance crown, with power consumption about half that
of the best current Athlon 64 models.
AMD has no real answer for at least a year to 18 months. Their new AM2
socket really does little more than add support for DDR2 memory. That's
nice, but isn't likely to increase performance much, if at all. Current
Athlon 64 models have a very good embedded DDR memory controller, and
the current Athlon 64 processors aren't memory bound with standard DDR.
AM2 processors will use a different socket pinout and support DDR2
memory, but otherwise will differ little from current models, at least
in the short term. The latest Athlon 64 steppings draw less power than
earlier models, but they still won't be able to touch Conroe for power
efficiency.
It remains to be seen how quickly Intel will ramp up Conroe. Their
current dual-core Pentium D models are nothing to sneeze at. Certainly,
they don't have the performance or moderate power consumption
of fast Athlon 64 X2 models, but neither do they have the high
price tag. In fact, the Intel Pentium D 805, which sells for only about
$125, holds the current bang-for-the-buck title among all mainstream
processors, at least for people who want a dual-core processor for
multi-tasking.
It's going to be an interesting next year or two.
16:30 -
I guess they showed me. Yesterday, I emailed Wagg-Ed, Microsoft's PR
firm, to ask for review copies of the latest Vista beta and Media
Center Edition, intending to feature them in the new edition of Building the Perfect PC.
They emailed me back, asking for specifics, which I provided. Was I
going to review the software? No, just feature screen shots, a
narrative description of installing the product, and so on. How will
the product be used? Just for illustrating the process of installing it
on a newly-built PC. And so on.
This afternoon, I got email back from them, telling me that they "will
not be able to provide review copies of Windows Vista or Windows Media
Center Edition." I replied, "No problem. I'll just use Linux instead."
All of the system configurations in the new book will be
Vista-compatible, which we'll point out, but we'll also point out that
we don't run Vista on any of them, and why.
10:00
-
I'm starting not to dislike Ubuntu 5.10 as much as I did originally. At
this point, I'm in the "I can live with it" stage, having just
graduated the "grit my teeth and bear it" stage. Although I'd still
prefer to be using Xandros, Xandros 3 won't run on my current hardware.
Xandros 4 will support my new hardware, but isn't expected before June,
about the time that Ubuntu/Kubuntu 6.06, also known as Dapper Drake,
should be released.
By that time, it's possible I'll have "outgrown" Xandros. Xandros is
extremely well integrated, and isolates new Linux users from the
complexities of the OS. The price you pay for that is limited
flexibility and choice. As long as you're willing to live with the
choices that Xandros makes for you, everything just works. If you want
to go beyond what Xandros offers, things can get ugly. Conversely,
Ubuntu/Kubuntu is extremely flexible and offers unlimited choice, but
at the expense of less hand-holding.
I'll probably install Xandros 4 and Kubuntu 6.06 and play with them
extensively before I settle on one or the other. For my personal use,
that is. I've installed Xandros OCE for several friends who are not
computer people. I'll either leave them on Xandros 3 or upgrade
them to Xandros 4, once the OCE version is available. Although some
people sneer at Xandros as "Linux with training wheels", in my opinion
it can't be beat for Windows refugees, particularly those who have no
desire to learn a new OS.
09:17
- Here's an example of why I prefer Xandros to Ubuntu. Barbara's Xandros system, adelie,
also functions as our ad hoc file server. I created four SMB (Windows)
shares on that system: archive, barbara (her home directory), holding,
and usr. I would like to map those SMB shares to subdirectories of my
home directory on newton, my main desktop system. For example, the SMB share //adelie/archive should map to the local directory on newton /home/thompson/archive.
To accomplish this on Xandros, I simply fire up Xandros File Manager
and choose Tools->Map Network Drive. XFM presents a list of
available Windows shares. I click on, say, the share /archive on adelie.
Xandros fills in a suggested mount point, /home/thompson/archive, which
I can accept or override. When I click OK, Xandros takes care of
everything automatically, and the new share shows up as just another
directory in Xandros File Manager.
With Ubuntu, the process is a lot more complicated. Brian Bilbrey, a
Linux guru, was kind enough to spend an hour on the phone with me last
night getting Ubuntu set up. After some unsuccessful mucking about, it
became clear that Ubuntu was going to let us mount the SMB shares, for
which we'd already manually created directories in my home directory.
After some head scratching, Brian said it sounded like I didn't have
smbfs installed and suggested I run the following command:
sudo apt-get update && sudo apt-get install smbfs
That worked, although for the life of me I don't understand why Ubuntu
didn't install smbfs by default. With smbfs installed, I was able as
root to issue the command:
mount -t smbfs //adelie/archive /home/thompson/archive
which indeed worked. I'd already created the directory
/home/thompson/archive, and when I clicked on that directory in
Ubuntu's file manager, I indeed saw the subdirectories and files
contained in that share on adelie.
I noticed, though, that the icons associated with those files and
directories made it clear that they didn't belong to me. Checking
permissions, I found out that they were owned by root. That makes
sense, since I did the mount command as root. So I used the command:
umount //adelie/archive
to unmount the SMB share. I then logged in as myself (rather than root)
and issued the mount command. Alas, Ubuntu returned an error message,
telling me I had to be root to mount something. Duh. That makes sense,
but it left me with the permissions problem. Also, of course, I was
issuing the mount commands at a command line, which means they wouldn't
survive a reboot. Presumably, I can add a line like the following to
fstab:
//adelie/archive /home/thompson/archive smbfs user 0 0
to mount the SMB shares automatically, but I haven't gotten around to doing that yet.
09:35
- And that did it. Thanks to email from Brian Bilbrey, I added the following lines to /etc/fstab.
//adelie/archive
/home/thompson/archive smbfs
rw,username=barbara,password=<barbara's-password>,uid=thompson,gid=thompson
0 0
//adelie/barbara
/home/thompson/barbara smbfs
rw,username=barbara,password=<barbara's-password>,uid=thompson,gid=thompson
0 0
//adelie/holding
/home/thompson/holding smbfs
rw,username=barbara,password=<barbara's-password>,uid=thompson,gid=thompson
0 0
//adelie/usr /home/thompson/usr smbfs rw,username=barbara,password=<barbara's-password>,uid=thompson,gid=thompson 0 0
And everything now works as it should. I've saved a copy of fstab in my never-delete directory.
12:13 - European consumer electronics giant Philips has filed a patent application
worthy of Dogbert. If implemented, this technology would prohibit
television viewers from changing the channel or muting the sound during
commercials, both on live TV and on programs recorded to a DVR. This
would be accomplished by inserting a signal to mark the beginning and
end of commercial breaks.
It might surprise my readers to learn that I'm all in favor of this
technology being implemented widely. Current automated methods of
detecting and eliminating commercials, such as those used by MythTV,
are only about 80% effective. Those methods depend on algorithms that
detect anomalies in the program stream, such as the fade to black that
occurs between a program segment and a commercial segment. Having
unambiguous markers in the program flow would make it much easier to
detect and eliminate commercials automatically.
Some years ago, I proposed a cooperative commercial-eliminating
mechanism similar to CDDB, by which users of PVR applications such as
MythTV could flag commercials manually and have the timing data
uploaded automatically to a database, which later viewers could access
automatically to flag their recordings to eliminate commercials. But
having the networks do the flagging for us would be better still.
Of course, there are a few problems with this. Mainstream viewers are
screwed, because they depend on commercial products, none of which will
delete the commercials for them. Those of us with video capture cards
and access to open-source apps like MythTV will benefit while most
viewers will suffer.
Also, the networks may have a cunning plan: inserting flags in the
midst of the actual program, which they hope our PVR software will
automatically delete, believing that material to be a commercial, and
thereby ruin the recording. But, as is always true of the cunning plans
of the copyright pigs, this simply won't work. Our open-source PVR apps
won't assume that everything flagged as a commercial is in fact a
commercial, but merely (a) that all commercials are likely to be
flagged as such, and (b) anything flagged is quite likely to be a
commercial. In combination with the current algorithms for detecting
commercials, this should boost accuracy from about 80% to more like
99.9%.
Thanks, Philips.
09:51
-
One of the annoying things about writing PC hardware books is that
hardware and software comes and goes so quickly. For a while, Barbara
and I had a running joke that our endorsement of a product was the kiss
of death. It happened repeatedly. We'd recommend a product, such as the
OnStream series of tape drives, only to find that the company had
entered bankruptcy, often between the time we submitted the final
manuscript draft and the time the book hit the stores.
I'm working on the media center PC for the new edition of Building the Perfect PC
right now. I've been looking at various PVR applications, such as
BeyondTV, Sage, MythTV, and so on. In the first edition of the book, we
looked at MyHTPC. We found a lot to like, but concluded that it wasn't
quite ready for prime time. In the time since then, there's been a lot
of development on the software, and it was released as a commercial
product from a company named Meedio. It looks a lot like Microsoft
Windows Media Center Edition, but is more featureful, less encumbered
by DRM, and much more configurable. We planned to recommend it (with
several competing products) to people who wanted to run Windows on
their media center boxes.
I'd done a nice write-up of Meedio Pro, with screen shots, details
about its features, and so on. Then yesterday I went back to the Meedio site
and found only an announcement that Meedio is no more. It's been bought
by Yahoo, and the Meedio software is no longer available or supported.
The electronic program guide stops working on 1 July, so anyone who's
bought Meedio is screwed.
There's probably a lesson here. In one sense, the loss of Meedio is no
big deal. Many competing products remain available, and some of them
are quite good. At least one of the good ones, GB-PVR,
is also free-as-in-beer, although not open source. But all of those
products share a common drawback. If the company goes out of business
or decides to discontinue the product, its users are stuck. It's quite
possible they'll need to strip their systems down to bare metal and
start again from scratch.
That doesn't happen with open-source applications like MythTV,
at least projects that have a large and active community. If the
original project leader abandons the project or gets run over by a
truck, others will continue the project. If the original project leader
starts taking the application in directions that the user base doesn't
like, someone will fork the project. No matter what, the risk of users
being abandoned is very small.
When we looked at MythTV in the first edition of the book, we concluded
that it wasn't yet ready for prime-time. We haven't looked at MythTV
lately, but we plan to do so. Just a quick perusal of the feature list
tells us that things have changed a lot in the last couple years. And
Linux is much, much easier to install and configure than it was a
couple years ago, so it's quite possible that MythTV is a practical
choice nowadays even for non-Linux folks. We'll have to see.
12:38 - This from Mark Huth:
From: Huth Mark
To: Robert Thompson
Subject: web pages
Date: Thu, 20 Apr 2006 09:00:16 -0700 (12:00 EDT)
Bob,
This is an odd comment. As
you know, I’ve been reading your web pages for a long time and
have never been critical of the pages themselves. Now I find that
the backgroup, the fine mottling you use as a background makes it
difficult to read the pages. Of interest, this doesn’t seem
to have anything to do with my eyes, but may well have something to do
with the use of high resolution LCD monitors! I tried a little
experiment, I looked at your web pages on my desk lcd’s and then
went down the hall to use a much older and smaller CRT. It seemed
much less distracting and was much easier to read the text on the
CRT. I then got a couple of partners, one nurse, and two medical
assistant to look at the pages (partners older, nurse and medical
assistants MUCH younger). All agreed that the text was very hard
to read on the LCD and much easier on the CRT.
Possible
conclusions…..give up reading the web page, throw away the lcd?,
ask Thompson to take a look at it and see if he can figure it you.
Grin,
mark
That is interesting. I'm looking at my page right now on a 19" Samsung
LCD, which has very high brightness and contrast, and the background is
so unobtrusive it almost disappears. Years ago, I used a flat white
background (255,255,255), but that seemed a bit glaring. I started
using the background image to tone things down a bit. I'll ask my other
readers what they think. Perhaps I should just drop the background
image and return to using a pure white background.
So, what does everyone else think? Post your thoughts on the messageboard.
08:43
- Paul Thurrott, whom no one could accuse of being anti-Microsoft or anti-Windows, has just issued a scathing indictment of Microsoft and Vista. From the article:
"Since
the euphoria of PDC 2003, Microsoft's handling of Windows Vista has
been abysmal. Promises have been made and dismissed, again and again.
Features have come and gone. Heck, the entire project was literally
restarted from scratch after it became obvious that the initial code
base was a teetering, technological house of cards. Windows Vista, in
other words, has been an utter disaster. And it's not even out yet.
What the heck went wrong?"
In fact, the situation is worse than even this article portrays. By the
traditional definition, feature completeness, Vista is not yet even in
alpha. The OS is fundamentally broken, primarily because the new
DRM features that Microsoft has embedded so deeply in Vista simply do
not work. Microsoft is pulling out all the stops to ship
something--anything--it can call "Vista" by the currently-promised
February 2007 release date, but I think it's very unlikely that it will
meet that date. Summer 2007 is more likely, and even then I expect the
first release of Vista to be disastrously bad.
Meanwhile, Linux just keeps chugging along.
00:00
-
00:00
-
Copyright
© 1998,
1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006 by Robert Bruce
Thompson. All
Rights Reserved.
