It's easy enough to see how your offspring will
fare with some odd name. Change your own, and
see.
I know several people who have changed their
name due to death of a spouse, divorce, etc.
It's not difficult, costly, or illegal.
But it can be a nuisance making the switch.
Credit cards, license, etc.
Why not let your kids decide their own name
when they get older?
Maybe he just wants to be B0b.
Most of the dependencies go away when binaries
are compiled statically. Unfortunately, the
defaults have leaned so heavily towards shared
libraries, that people don't know how to compile
static anymore.
It makes one pine for the days of the a.out format.
I have one of those 128 MB USB dongles. This is
generally enough for a ripped CD. It'd be nice if
I could plug it into a player that knows how
to mount it, find music on it, and play it.
Then, I could "upgrade" my player by getting
another USB dongle. Further, if I could plug my
USB dongle into my car stereo, I wouldn't have
to burn CDs all the time. I bet a USB dongle
would never skip, no matter how deep
the pothole.
It's a feature that my USB dongle plugs into
a Mac, Windows, Linux without special drivers,
and it just works. I can put anything I want
on it, up to capacity.
While we're at it, I want a digital camera
where I can plug in a USB dongle.
I have three digital cameras. Kodak, Samsung
and Olympus. They're all USB. They all come
with USB cables. There are three different
camera side connectors. So, I can't just
keep the cables plugged into my PC, because
I can't spare three ports (I only have six).
The Samsung doesn't
talk to my Red Hat 9 (yet) - so there's
some sort of protocol change too. Feh.
It's got to work out of the box.
Accordingly, those Big Bang embers are now some 45 billion years out.
As the Universe is considered to be 13.7 +/- 0.2
billion years old, and things move at light
speed at maximum, how is anything over 27.4
billion years from anything?
The author benchmarks
a 2.8GHz Xeon with 533MHz FSB and 1MB of L3 cache.
and
a 3.2GHz P4C with an 800MHz FSB and 0.5MB of L3 cache.
He claims he doesn't want to compare the two, but
he does. Here are some other conclusions.
The Xeon has a slower clock, and yet outperforms
the higher clock P4C. This is further evidence
that MHz isn't everything.
The P4C has higher memory bandwidth (the FSB)
yet slower performance. This shows that on-chip
cache can be king over memory bandwidth too.
Some of my historic benchmarks fit completely
in the 486's cache, so not all applications
will benefit from more.
Alien searches (SETI@Home) appear to benefit
from large on-chip caches up to it's resident
set size (about 13 MB). The more the better.
My current favorite production application
has a resident set size of about 200 MB.
It isn't clear that on-chip cache size makes much
difference. It is clear that FSB bandwidth
makes all the difference.
As always, the best benchmark is your application.
Unfortunately, most of us can't run our
favorite application on a variety of machines
before buying one. I know I end up buying
something that appears cost effective.
This favors the low end processors, which
at the moment favors AMD in the X86 world.
I've been particularly highly impressed
with the Athlon's memory bandwith performance.
My Athlon 1800+ (1.3 GHz) performs better
than 1.5 Ghz P4's at work - primarily due to
having more than double the memory bandwidth.
It was also considerably cheaper.
I feel as if I got a good deal.
I personally have shown no brand loyalty,
purchasing a chip from a differant vendor
each time.
My 1987 Mac II has been in operation basically
every day since late September, 1987.
My boot drive died in about 1995.
In 1996, my laser printer died.
In 2001, the battery for the clock died.
Hardware upgrades on the Mac are trivial.
When Word 6 for the Mac came out, it was larger
than my hard disk. So, I stopped upgrading
at Word 5. I would have been just as happy
had I stopped with Word 4.
At first, I upgraded the OS whenever possible.
But with System 7.5.5, Apple said that was the
last. This has been a blessing.
I haven't performed a software upgrade in over
seven years. In fact, the only maintenance
is an occasional backup. I power up an external
SCSI drive, and copy one disk to the other.
In the meantime, I'm on my 4th x86 box
since 1987. The PC/xt was too lame to run Linux.
The 386/33 and the the PII both died after just
four or five years.
Feh. Hardware upgrades are driving me to
Linux upgrades. It takes over a year of admin
to get all the non-distribution stuff working
again on a new distribution. Feh.
Upgrades suck - and it's all because of
shared libraries. When are we going to learn
to distribute binaries compiled statically?
My oldest a.out static binaries still work.
Not so with newer stuff. Feh.
I don't want to pay for radio, have an mp3 CD
player in my car, and drive two plus hours a
day for comuting.
My new form of entertainment is free books.
Project Gutenberg had some mp3 books for a
bit - including a few from audiobooks4free.com
which are read by humans
(the site could be named audiobooks4cheap.com).
After downloading a few computer read books,
I decided that it was somewhat better to
roll my own.
I downloaded festival and notlame,
and got them to work under Linux. It takes
my Athlon 1800+ about a half hour to convert
a book to mp3 - consuming 60 MB to 150 MB.
You'll need at least 256 MB RAM.
Choose a book from over 10,000 books.
Download it - 300 KB - 500 KB. This is doable
with dialup.
Break it up into chapters. Festival's memory
requirements grow with the length of the
text to be converted. 512 MB is not enough
to convert most books. Also, my car's mp3
player does not remember where you were in
the track when you turn off the car. It's
impossible to find where you were in a 4
hour track.
Use festival to convert to wav files.
Use notlame to convert to mp3.
When you have enough books, cut a CD.
My script:
#!/bin/sh
# Convert a book (set of text files) to mp3.
# Current directory, *.txt => *.mp3
for i in *.txt; do
echo $i
A=`basename $i.txt`
sed s/_//g $A.t
time text2wave $A.t -o $A.wav
rm -f $A.t
time notlame --silent -h -mm $A.wav $A.mp3
rm -f $A.wav
done
I've glossed over getting festival compiled.
I went the route of downloading all available
tarballs. Create a new directory, and unpack
everything there. Don't forget speech_tools.
Compile speech_tools, then festival. Then
install.
Getting everything to work is painful enough
that it would be worth putting together
a knoppix disk that has everything built.
Then we'd need a site that can host a
binary of an mp3 encoder. I live in one
of those countries where I can only download
the source and compile it myself.
So, I'd have to put together a knoppix disk
that runs make... In any case, 256 MB RAM
might still be the low end for a knoppix
converter. You need a writable hard disk
partition (or big enough USB dongle) for
the mp3's. If you boot from CD, you need
a second drive to cut disks...
When you are all done, there's a learning
curve to understanding computer read text.
For me, it was three chapters.
The books I've converted run about 4 to 6
hours. I expect to 'read' one per week.
I get five or six books per CD.
That makes this about $.03 per book.
Less if I use rewritable CDs (for which
I have previously had little use).
You might notice that I encode books at
128 kbps - mono. From the arithmetic,
I expect this to allow
about 12 hours of books on a CD.
However, I get more like 24 hours.
Audiobooks4free has some 32 kbps books.
I find the artifacts
annoying, but improved the treble is
turned down all the way. I have some really,
really bad powered speakers from radio shack
(powered by 'C' sized batteries). 32 kbps
books sound great on it. This suggests that
a transform exists that could post-process
the 32 kbps decode to remove the artifacts.
If the speaker diameter is 2 cm, what would
the transform be?
A press release and a decimal expansion of the number are coming soon.
...the 1,521,561-digit prime...
53592^5054502+1
For Unix (and Linux) users, the bc script:
5359 * (2 ^ 5054502) + 1
produces the decimal expansion in about 5 minutes
on a PIII/800. As bc works, it's resident set
size increases, reaching a maximum of about
6.6 MB of RAM. So, a 386 with 8 MB RAM
running Linux could easily compute this result
in 2 to 3 hours.
One could publish the number easily enough.
A bzip2 compressed verion is about 656,116
bytes. However, just running the calculation
in bc seems reasonable at this point, avoiding
a slashdot effect.
That is, assuming you needed to know, for
some reason.
The article starts, I know this is wrong.
It's a start.
One way that the open DHCP configuration could
be exploited from anywhere on the internet
is if there is a Windows machine
on the same subnet. It could be broken into,
and used to own the Mac. This IS serious.
It appears best not to run Windows on the same
subnet with your Mac.
This is only a Major Mac Breach
in the sense that, it's the most serious
security issue on the Mac today. In the mean
time, someone anywhere on the internet can
do whatever they want with your machine just
because you run Internet Explorer on Windows.
And this isn't serious because no one has
bothered to exploit it. There are so many
other choices for which exploits are easily
available.
At work, I use Windows 2000 Pro. It works fine.
In the past 6 months, I haven't been infected
with a virus or worm. However, the network was
slowed to a crawl for about a week this summer
when a significant
fraction of identical machines were compromised.
The de-lousing and the patch downloads,
installs, failed installs,
reinstalls loaded the network down, and positive
non-zero work (for me) could really only be
accomplished on the Unix servers.
Since then, about twice a month, when I boot
Windows, some patch or other gets installed,
and, typically, my machine is rebooted, and
I have to log in again. It's a nuisance, costing
me only about an hour a month. I've only had
to call the help desk twice over patch issues.
The reason it works so well, is that the
company has a firewall, which is monitored
continuously by staff. Vendor patches are
monitored, tested on isolated test machines,
and automatically sent to end user systems
on reboot. This works, because, in Windows
land, frequent reboots are still manditory.
System stability is such that you can't simply
leave the system up and running for weeks
or months. This is still an improvement
over Windows 95, where in addition to insecurity,
installation of commercial software ran the
risk of creating software incompatibilities,
compromising stability. The company's response
was to have a staff perform compatiblity
testing, to ensure that I could get the right
versions of the tools I might need.
At home, I'm also connected to the internet
on a 24x7 basis. I don't have a staff to
monitor vendor patches and firewalls or
to man the help desk.
I can only look at the system for about a
half hour a week night. This is not enough
time for me to read my email, and persue
my interests. I still need security, so
Windows simply isn't an option. The options
are 1) a Mac, and 2) Unix (Linux in my case).
I'm on a limited budget, so I have opted to
use Linux as my firewall. So far, it's been
secure. I've shut down non-essential services,
I've applied both relevant patches so far
(I have not heard that there is an exploit
for either).
My email client has a built in spam filter,
and does not, by default display email.
I get to evaluate the subject line first.
Therefore, I generally do not get smut in my
face. My email client is incapable of executing
anything that is sent to me. It does not run
Java or Javascript even when I view a message.
When I forward a message to my friends, the
default is not to forward attachments. As a
result, I do not participate in email viruses.
I can generally read my email even if my six
year old son is in the room.
My internet browser does not know how to
launch pop up (or under) windows. I haven't
missed it.
A coworker runs XP at home. His chat client
has compromised his machine so as to send him
to smut sites constantly. Ostensibly, he's
a computer professional, but he's not been
able to rid his computer of constant pop up
smut. As a result, he can't let his children
use his computer. Or even be in the same
room when it is on. I suggested the Linux
Patch disk.
The easiest way to make your Windows box
secure today is to never connect it to a
network. No anti-virus install. No patches.
It just works.
As a Linux and Mac
Alternative to 98
on
PC Annoyances
·
· Score: 2, Interesting
I bought and read the first Windows Annoyances.
That was despite the URL on the back to
the on-line complete and up to date version.
I even liked it. It had useful suggestions.
Certainly, there are alternatives to Windows
when you have a choice. I'm often not paid
to choose. My liking the book has to do with
the signal to noise ratio. Many books have no
signal.
The original book was a compilation - kind of
a blog. People submitted content to the site.
One scary piece was that, every now and then,
Windows would delete a folder containing an
MS competitor's product.
Not a problem - since essentially all Windows
users back up their systems on a regular basis...
Note that Windows XP is too large and slow to
run on many machines that run Windoze 98.
And, '98 is still a virus/worm nightmare.
For these smaller and slower machines,
the options are - get a firewall, antivirus,
etc., or, load it with Linux.
I still prefer Win 2000 pro over XP, in
a lesser of evils sort of way.
We have straitforward competition already, more
or less. Arian, Proton, H2,
Long March, Taurus, Pegasus,
Titan, Atlas, Delta, Shuttle, etc.
Lots of countries restrict export of payloads,
leading to restriction of competition.
One problem is that there aren't enough paying
payloads to support these launchers.
The biggest lifters, Saturn V and Energia
have been discontinued - despite economy of scale.
The shuttle has enjoyed continued funding despite
high cost. The ISS would have been MUCH cheaper
if lifted by Saturn V or Energia.
Also canceled were X-15 and Dyna-Soar concepts.
The X-15 made it to sub-orbital space flight.
Development of a new rocket is difficult and
expensive.
The N1 never successfully flew.
It's not because the Russians couldn't build
rockets. When we talk about something that's
difficult, we call it "rocket science" (it's
really technology - rocket engineering).
Several inovative projects have been aborted
due to costs, risks or technical difficulty.
Some of these promise cheaper rides. We don't
have a way of optimizing the funding of
projects in a way that ensures success.
A case in point. The National Aerospace Plane
project was canceled due to cost and/or
technical difficulties. EELV expects modest
gains, and funds two launcher families.
One project accepts high risks and fails,
the other accpets low risks, and double spends
to ensure that if one launcher fails, that
there is still a capabile indiginous launcher
available. The EELV project could not accept
higher risk within cost bounds.
Does anyone recall that, on something like the
25th anniversary of the moon landings,
President Bush Sr. called for a grand plan
to go to Mars? Are we there yet?
Just because the president says so, doesn't
mean that we're going.
JFK said so, and we went. Bush isn't JFK.
We choose to go to the moon, not because
it is easy, but because it is hard. -JFK
I'd like to hear, We choose to go to the
moon, not because it is hard, but to achieve
the following goals...
We're currently doing great things in space.
This month and next, a flotilla of space craft
will approach Mars. These are projects that
are objective driven. They cost a minor fraction
of what ISS or the space shuttle cost.
What is the mission of ISS? One answer has been
that it is a necessary step in going to Mars.
However, it does not solve problems of weighlessness
during the trip, or address radiation shielding,
or food growth or recycling.
Much of the science or technology that can be
gained from low Earth orbit is better accomplished
with untended platforms. For example, crystal
growth, and space telescopes require vibration
free environments, in addition to microgravity.
And ISS is expensive. The last estimate I heard
was $97e9. For what?
The Space Shuttle is a marvel, but also quite
expensive. It's primary unique capability -
bringing large things back from space - is
mostly unusued. It's pretty heavy lift, but
despite reuse, it's more expensive than
disposable rockets on a per-pound to orbit,
even if you ignore original construction costs.
The main problem with the Shuttle and ISS
is that they are goals unto themselves, rather
than the means to some end.
And, the programs are structured so that
it seems to make sense to keep spending to
keep the project alive.
I don't want to hear that we're going to create
a permanant station on the moon. I want to
hear about what we're going to do once we're
there.
It seems to me that it should be cheaper to
go to an asteroid like Eros than to the moon.
Eros has little gravity, so launching from it
does not require as much fuel. It still
costs $10,000 a pound to get to low Earth orbit.
The fuel has to get there somehow.
Meaning you have to explicitly provide all dependent RPMs during installation. This is the part that sucks.
To a large extent, what really sucks are the
dependencies themselves. A large fraction of
these dependencies stem from shared libraries.
Shared libraries introduce binaries that work
on one distribution, but not another - forcing
upgrades of non-distribution packages installed,
pointlessly.
What do shared libraries save? A little RAM?
Often, they don't save any RAM. Many packages
are the only users of a given shared library.
In this case, compiling statically produces
a smaller binary - as unused routines are not
included. Disk space? Shared libraries must
have symbols included. You need to have several
apps use a shared library to save anything.
On top of this, RAM and disk are becoming
dirt cheap compared to what is saved. Really,
$105 US buys a 160 GB hard disk, $110 buys
512 MB RAM.
We can reduce this dependency nightmare nearly
overnight - compile statically.
As an experiment, download a copy of Netscape 2.
Attempt to surf the net for an hour. Does it work?
Now, try again with Netscape 3.
If I were into conspiracy theories, I'd say that
someone deliberately distributes
web page creation tools
that pointlessly use features that tickle
bugs in older browsers - eventually
forcing upgrades.
Warning - this page is old-browser dehanced.
Netscape 3 has all the features I want in
a browser, except one - it's buggy. It can
format text and graphics. It does forms.
It does ssl (security). It's small and very
fast (except that it's cache slows rapidly,
and you have to restart it frequently to
maintain speed).
People say, what does it matter? New browsers
are free. My response is that downloading
a new browser over a 28.8 modem is not nearly
free. Most users do not have what it takes
to make it happen. They get a browser upgrade
by buying a new system.
I need files that are larger than 4 GB.
I've needed this for 5 years - ever since
2.3 GB hard disks became affordable for
home use.
We could have this now, with 32 bit procesors.
Under Linux, the only way I know how to do
this is to stream to a raw disk (no filesystem).
I'm tempted to do this for backups - one disk
to another. However, I'd like more than one
backup image on the backup drive.
cd / ; tar cf - . | gzip -9 >/dev/hdb1
I need more than 32 bits of integer arithmetic.
bzip2 is limited to
512 MB files (2 ^ 29 bytes), because
it needs to calculate addresses to bits, and
is limited to 32 bits. It probably could
have been written using gcc's long long's
(which are 64 bits), but it would likely hamper
file portability. bzip2 can not produce
a single archive file that fills a single CD.
I need more than 32 bits of RAM addressing.
An app I wrote appears to
require about 8 GB RAM for a full production
run.
Yet, most of my applications ran fine on
16 bit machines.
I would have picked up an Alpha if I could
have afforded it. This shows that just
being the fastest machine on Earth does
not get you a spot on the desktop. Did
DEC offer last year's model for cheap?
The only low end Alpha's were on the used
market, as far as I know.
I would have picked up an AMD64 rather
than an Athlon, if I could have gotten one
cheap enough. My old machine died, so I
had to upgrade. It was a surprise, and
I didn't have much extra cash. I picked
up a low end 1800+, but with a good
motherboard, expandable to 3 GB RAM, and
3200+ chip. I may add RAM, but, so far,
I'm happy with the speed.
Ok, so go ask Average User how fast the CPU in the HP Pavilion 3000+ is. Odds are they'll say 3.0 GHz, which isn't true but is proof of AMD's success in "looking good on paper".
On the other hand, when the average user asks me what
the AMD model names mean, I tell them that
my AMD 1800+ is more than twice as fast as
the 1.6 Ghz Intel P4 at work. This P4 has
133 MHz RAM, which is less than half the
bandwidth available to my 1800.
My points is that AMD's suggestive model
numbers do not appear to be misleading.
If anything, they're conservative.
Yes, I understand that single numbers
are a poor guide to performance. I also
understand that the market wants single
numbers. AMD is using Intel's numbers (MHz),
and relating it to real performance
in a conservative way.
For me, this builds trust.
I still recall Intel's trust busting FDIV bug
handling fiasco. Really, tell me
again about how I don't care if my
math is inaccurate. Why should I buy
from them again?
Back to the comparison.
My company bought cheap P4's off the shelf.
What I did was buy a good motherboard. My
guess is that I spent $20 more on it than
was spent on the P4. I was able to do this
because I put the box together myself.
Had I picked up a "system", it wouldn't be
$20, it would be $500 more,
to get the same performance, as near as I
can tell.
My company has more than 30,000 seats.
They could easily assembly PCs themselves
and save money. In fact, there are many
things they could do to save money.
The moral here is that big companies
aren't necessarily very bright.
"We haven't talked to a single user who has said they're using [open source] because it's better." Tipp argued that it is more a case of sheer frustration with licensing and Microsoft's poor relationship with its customers over the last few years -- or simply the perceived cost benefits of open source -- driving users to migrate.
Open source licensing is better than Microsoft's
licensing. License servers are a continual
cause of downtime. Some sites have to hire
a full time body just to track how many installs
of various products are still on machines.
Help desks are often very unresponsive.
For many companies, a quick look at the source
code means that problems can be quickly solved.
Lower costs are better.
The idea that these features are somehow
seperate from code quality is fiction.
I use Unix, Linux and the Mac because they are
superior to Windows. I make the claim that
there is ALWAYS a superior product over Windows.
Don't tell me, "Thank you for choosing MicroSoft". If I'm using it, you can
bet it wasn't a choice.
Smart battery chargers made more difference
to me than increased capacity.
In the old days, you put your NiCad batteries
into the device, and ran it to zero. Then,
you'd put the batteries on the charger for
exactly six hours. If you forgot, you reduced
the life and performance of the batteries.
You ran them to zero, because otherwise you
didn't know how long to charge them.
Modern chargers know when the battery is full,
and stop. Leave the battery on the charger
for weeks and you now get fully charged batteries
when you want them. And, you get 1,000 cycles
instead of 50 to 100.
Is the battery at an unknown charge level, or
known to be half charged? No problem. Just
throw it on the charger.
1,000 cycles is a cycle a week for 20 years.
Modern chargers do this for old batteries too.
So, with the old chargers, the batteries
required vigilent maintenence. With new
chargers, the batteries are at your
convenience.
My digital camera takes about 100 pictures
on a smart media card. It also takes about
100 pictures on one cycle of batteries.
When I'm done shooting, I dump the pictures
on my computer, and swap batteries to another
set. If it was $2 for new batteries, I would
take fewer pictures. For me, it's about a
penny.
Take AA sized rechargables. I have NiCad, from 20 years ago. They didn't last as long as disposable per charge, but were cheaper in the long run due to reuse.
Then NiMh's came out, aproaching disposables in per-charge performance. They were something like 1400 milli-amp hours. I bought them.
Then, better chargers came out, increasing the number of cycles, not just for the new batteries, but for my old ones too. I bought a new charger.
Then, the 1600's came out. I bought them.
Then 1700's - I have some of those.
Then 1800's - yep.
Now I've seen 2200's. I'll be buying those soon.
At the store, the cashier says, "but no one buys recharagables, compared to disposables". Of course not. They last 20 years instead of a few days.
I get 8 hours out of my walkman for about a penny. I'm not going back.
Two things make modern NiMH batteries
superior to old NiCad.
Modern NiMH formulations store about double
what NiCads did per charge.
Modern chargers know when the battery is
charged, and stop. So, now it really is
typical to get 1,000 cycles out of them
rather than less than 100. Modern chargers
can do this for NiCad too - mine does.
Heat kills rechargable batteries. This can
come from the charger, through overcharging.
This can come from use - through high
discharge rates.
There's an engergizer charger sold at Meier's
for under $30. With it, NiMH batteries last
1,000 cycles. That's a charge a week for 20
years. The charger knows when the battery
is charged, and does the right thing. You
can swap between two sets and run for a really
long time.
It seems that we are collectively engaged in
a climate control experiment. At the moment,
there is no goal. It looks like time to
buy property at high elevations.
Slashdot Mirror
I know several people who have changed their name due to death of a spouse, divorce, etc. It's not difficult, costly, or illegal.
But it can be a nuisance making the switch. Credit cards, license, etc. Why not let your kids decide their own name when they get older? Maybe he just wants to be B0b.
It makes one pine for the days of the a.out format.
It's a feature that my USB dongle plugs into a Mac, Windows, Linux without special drivers, and it just works. I can put anything I want on it, up to capacity.
While we're at it, I want a digital camera where I can plug in a USB dongle.
I have three digital cameras. Kodak, Samsung and Olympus. They're all USB. They all come with USB cables. There are three different camera side connectors. So, I can't just keep the cables plugged into my PC, because I can't spare three ports (I only have six). The Samsung doesn't talk to my Red Hat 9 (yet) - so there's some sort of protocol change too. Feh. It's got to work out of the box.
As the Universe is considered to be 13.7 +/- 0.2 billion years old, and things move at light speed at maximum, how is anything over 27.4 billion years from anything?
The Xeon has a slower clock, and yet outperforms the higher clock P4C. This is further evidence that MHz isn't everything.
The P4C has higher memory bandwidth (the FSB) yet slower performance. This shows that on-chip cache can be king over memory bandwidth too.
Some of my historic benchmarks fit completely in the 486's cache, so not all applications will benefit from more. Alien searches (SETI@Home) appear to benefit from large on-chip caches up to it's resident set size (about 13 MB). The more the better. My current favorite production application has a resident set size of about 200 MB. It isn't clear that on-chip cache size makes much difference. It is clear that FSB bandwidth makes all the difference.
As always, the best benchmark is your application. Unfortunately, most of us can't run our favorite application on a variety of machines before buying one. I know I end up buying something that appears cost effective. This favors the low end processors, which at the moment favors AMD in the X86 world. I've been particularly highly impressed with the Athlon's memory bandwith performance. My Athlon 1800+ (1.3 GHz) performs better than 1.5 Ghz P4's at work - primarily due to having more than double the memory bandwidth. It was also considerably cheaper. I feel as if I got a good deal. I personally have shown no brand loyalty, purchasing a chip from a differant vendor each time.
When Word 6 for the Mac came out, it was larger than my hard disk. So, I stopped upgrading at Word 5. I would have been just as happy had I stopped with Word 4.
At first, I upgraded the OS whenever possible. But with System 7.5.5, Apple said that was the last. This has been a blessing.
I haven't performed a software upgrade in over seven years. In fact, the only maintenance is an occasional backup. I power up an external SCSI drive, and copy one disk to the other.
In the meantime, I'm on my 4th x86 box since 1987. The PC/xt was too lame to run Linux. The 386/33 and the the PII both died after just four or five years. Feh. Hardware upgrades are driving me to Linux upgrades. It takes over a year of admin to get all the non-distribution stuff working again on a new distribution. Feh.
Upgrades suck - and it's all because of shared libraries. When are we going to learn to distribute binaries compiled statically? My oldest a.out static binaries still work. Not so with newer stuff. Feh.
My new form of entertainment is free books. Project Gutenberg had some mp3 books for a bit - including a few from audiobooks4free.com which are read by humans (the site could be named audiobooks4cheap.com). After downloading a few computer read books, I decided that it was somewhat better to roll my own.
I downloaded festival and notlame, and got them to work under Linux. It takes my Athlon 1800+ about a half hour to convert a book to mp3 - consuming 60 MB to 150 MB. You'll need at least 256 MB RAM.
Choose a book from over 10,000 books.
Download it - 300 KB - 500 KB. This is doable with dialup.
Break it up into chapters. Festival's memory requirements grow with the length of the text to be converted. 512 MB is not enough to convert most books. Also, my car's mp3 player does not remember where you were in the track when you turn off the car. It's impossible to find where you were in a 4 hour track.
Use festival to convert to wav files.
Use notlame to convert to mp3.
When you have enough books, cut a CD.
My script: .txt`
#!/bin/sh
# Convert a book (set of text files) to mp3.
# Current directory, *.txt => *.mp3
for i in *.txt; do
echo $i
A=`basename $i
sed s/_//g $A.t
time text2wave $A.t -o $A.wav
rm -f $A.t
time notlame --silent -h -mm $A.wav $A.mp3
rm -f $A.wav
done
I've glossed over getting festival compiled. I went the route of downloading all available tarballs. Create a new directory, and unpack everything there. Don't forget speech_tools. Compile speech_tools, then festival. Then install.
Getting everything to work is painful enough that it would be worth putting together a knoppix disk that has everything built. Then we'd need a site that can host a binary of an mp3 encoder. I live in one of those countries where I can only download the source and compile it myself. So, I'd have to put together a knoppix disk that runs make... In any case, 256 MB RAM might still be the low end for a knoppix converter. You need a writable hard disk partition (or big enough USB dongle) for the mp3's. If you boot from CD, you need a second drive to cut disks...
When you are all done, there's a learning curve to understanding computer read text. For me, it was three chapters.
The books I've converted run about 4 to 6 hours. I expect to 'read' one per week. I get five or six books per CD. That makes this about $.03 per book. Less if I use rewritable CDs (for which I have previously had little use).
You might notice that I encode books at 128 kbps - mono. From the arithmetic, I expect this to allow about 12 hours of books on a CD. However, I get more like 24 hours. Audiobooks4free has some 32 kbps books. I find the artifacts annoying, but improved the treble is turned down all the way. I have some really, really bad powered speakers from radio shack (powered by 'C' sized batteries). 32 kbps books sound great on it. This suggests that a transform exists that could post-process the 32 kbps decode to remove the artifacts. If the speaker diameter is 2 cm, what would the transform be?
5359 * (2 ^ 5054502) + 1
produces the decimal expansion in about 5 minutes on a PIII/800. As bc works, it's resident set size increases, reaching a maximum of about 6.6 MB of RAM. So, a 386 with 8 MB RAM running Linux could easily compute this result in 2 to 3 hours.
One could publish the number easily enough. A bzip2 compressed verion is about 656,116 bytes. However, just running the calculation in bc seems reasonable at this point, avoiding a slashdot effect. That is, assuming you needed to know, for some reason.
Windows is better than nothing.
Windows will be better than totally secure!
One way that the open DHCP configuration could be exploited from anywhere on the internet is if there is a Windows machine on the same subnet. It could be broken into, and used to own the Mac. This IS serious. It appears best not to run Windows on the same subnet with your Mac.
This is only a Major Mac Breach in the sense that, it's the most serious security issue on the Mac today. In the mean time, someone anywhere on the internet can do whatever they want with your machine just because you run Internet Explorer on Windows. And this isn't serious because no one has bothered to exploit it. There are so many other choices for which exploits are easily available.
At work, I use Windows 2000 Pro. It works fine. In the past 6 months, I haven't been infected with a virus or worm. However, the network was slowed to a crawl for about a week this summer when a significant fraction of identical machines were compromised. The de-lousing and the patch downloads, installs, failed installs, reinstalls loaded the network down, and positive non-zero work (for me) could really only be accomplished on the Unix servers.
Since then, about twice a month, when I boot Windows, some patch or other gets installed, and, typically, my machine is rebooted, and I have to log in again. It's a nuisance, costing me only about an hour a month. I've only had to call the help desk twice over patch issues.
The reason it works so well, is that the company has a firewall, which is monitored continuously by staff. Vendor patches are monitored, tested on isolated test machines, and automatically sent to end user systems on reboot. This works, because, in Windows land, frequent reboots are still manditory. System stability is such that you can't simply leave the system up and running for weeks or months. This is still an improvement over Windows 95, where in addition to insecurity, installation of commercial software ran the risk of creating software incompatibilities, compromising stability. The company's response was to have a staff perform compatiblity testing, to ensure that I could get the right versions of the tools I might need.
At home, I'm also connected to the internet on a 24x7 basis. I don't have a staff to monitor vendor patches and firewalls or to man the help desk. I can only look at the system for about a half hour a week night. This is not enough time for me to read my email, and persue my interests. I still need security, so Windows simply isn't an option. The options are 1) a Mac, and 2) Unix (Linux in my case). I'm on a limited budget, so I have opted to use Linux as my firewall. So far, it's been secure. I've shut down non-essential services, I've applied both relevant patches so far (I have not heard that there is an exploit for either).
My email client has a built in spam filter, and does not, by default display email. I get to evaluate the subject line first. Therefore, I generally do not get smut in my face. My email client is incapable of executing anything that is sent to me. It does not run Java or Javascript even when I view a message. When I forward a message to my friends, the default is not to forward attachments. As a result, I do not participate in email viruses. I can generally read my email even if my six year old son is in the room.
My internet browser does not know how to launch pop up (or under) windows. I haven't missed it.
A coworker runs XP at home. His chat client has compromised his machine so as to send him to smut sites constantly. Ostensibly, he's a computer professional, but he's not been able to rid his computer of constant pop up smut. As a result, he can't let his children use his computer. Or even be in the same room when it is on. I suggested the Linux Patch disk.
The easiest way to make your Windows box secure today is to never connect it to a network. No anti-virus install. No patches. It just works.
As a Linux and Mac
The original book was a compilation - kind of a blog. People submitted content to the site.
One scary piece was that, every now and then, Windows would delete a folder containing an MS competitor's product. Not a problem - since essentially all Windows users back up their systems on a regular basis...
Note that Windows XP is too large and slow to run on many machines that run Windoze 98. And, '98 is still a virus/worm nightmare. For these smaller and slower machines, the options are - get a firewall, antivirus, etc., or, load it with Linux.
I still prefer Win 2000 pro over XP, in a lesser of evils sort of way.
Sikorsky had a program awhile ago. X-Wing
Lots of countries restrict export of payloads, leading to restriction of competition.
One problem is that there aren't enough paying payloads to support these launchers.
The biggest lifters, Saturn V and Energia have been discontinued - despite economy of scale. The shuttle has enjoyed continued funding despite high cost. The ISS would have been MUCH cheaper if lifted by Saturn V or Energia.
Also canceled were X-15 and Dyna-Soar concepts. The X-15 made it to sub-orbital space flight.
Development of a new rocket is difficult and expensive. The N1 never successfully flew. It's not because the Russians couldn't build rockets. When we talk about something that's difficult, we call it "rocket science" (it's really technology - rocket engineering).
Several inovative projects have been aborted due to costs, risks or technical difficulty. Some of these promise cheaper rides. We don't have a way of optimizing the funding of projects in a way that ensures success.
A case in point. The National Aerospace Plane project was canceled due to cost and/or technical difficulties. EELV expects modest gains, and funds two launcher families. One project accepts high risks and fails, the other accpets low risks, and double spends to ensure that if one launcher fails, that there is still a capabile indiginous launcher available. The EELV project could not accept higher risk within cost bounds.
Just because the president says so, doesn't mean that we're going. JFK said so, and we went. Bush isn't JFK.
We choose to go to the moon, not because it is easy, but because it is hard. -JFK
I'd like to hear, We choose to go to the moon, not because it is hard, but to achieve the following goals...
We're currently doing great things in space. This month and next, a flotilla of space craft will approach Mars. These are projects that are objective driven. They cost a minor fraction of what ISS or the space shuttle cost.
What is the mission of ISS? One answer has been that it is a necessary step in going to Mars. However, it does not solve problems of weighlessness during the trip, or address radiation shielding, or food growth or recycling.
Much of the science or technology that can be gained from low Earth orbit is better accomplished with untended platforms. For example, crystal growth, and space telescopes require vibration free environments, in addition to microgravity.
And ISS is expensive. The last estimate I heard was $97e9. For what?
The Space Shuttle is a marvel, but also quite expensive. It's primary unique capability - bringing large things back from space - is mostly unusued. It's pretty heavy lift, but despite reuse, it's more expensive than disposable rockets on a per-pound to orbit, even if you ignore original construction costs.
The main problem with the Shuttle and ISS is that they are goals unto themselves, rather than the means to some end. And, the programs are structured so that it seems to make sense to keep spending to keep the project alive.
I don't want to hear that we're going to create a permanant station on the moon. I want to hear about what we're going to do once we're there.
It seems to me that it should be cheaper to go to an asteroid like Eros than to the moon. Eros has little gravity, so launching from it does not require as much fuel. It still costs $10,000 a pound to get to low Earth orbit. The fuel has to get there somehow.
To a large extent, what really sucks are the dependencies themselves. A large fraction of these dependencies stem from shared libraries.
Shared libraries introduce binaries that work on one distribution, but not another - forcing upgrades of non-distribution packages installed, pointlessly.
What do shared libraries save? A little RAM? Often, they don't save any RAM. Many packages are the only users of a given shared library. In this case, compiling statically produces a smaller binary - as unused routines are not included. Disk space? Shared libraries must have symbols included. You need to have several apps use a shared library to save anything. On top of this, RAM and disk are becoming dirt cheap compared to what is saved. Really, $105 US buys a 160 GB hard disk, $110 buys 512 MB RAM.
We can reduce this dependency nightmare nearly overnight - compile statically.
If I were into conspiracy theories, I'd say that someone deliberately distributes web page creation tools that pointlessly use features that tickle bugs in older browsers - eventually forcing upgrades.
Warning - this page is old-browser dehanced.
Netscape 3 has all the features I want in a browser, except one - it's buggy. It can format text and graphics. It does forms. It does ssl (security). It's small and very fast (except that it's cache slows rapidly, and you have to restart it frequently to maintain speed).
People say, what does it matter? New browsers are free. My response is that downloading a new browser over a 28.8 modem is not nearly free. Most users do not have what it takes to make it happen. They get a browser upgrade by buying a new system.
I need files that are larger than 4 GB. I've needed this for 5 years - ever since 2.3 GB hard disks became affordable for home use. We could have this now, with 32 bit procesors. Under Linux, the only way I know how to do this is to stream to a raw disk (no filesystem). I'm tempted to do this for backups - one disk to another. However, I'd like more than one backup image on the backup drive. /dev/hdb1
cd / ; tar cf - . | gzip -9 >
I need more than 32 bits of integer arithmetic. bzip2 is limited to 512 MB files (2 ^ 29 bytes), because it needs to calculate addresses to bits, and is limited to 32 bits. It probably could have been written using gcc's long long's (which are 64 bits), but it would likely hamper file portability. bzip2 can not produce a single archive file that fills a single CD.
I need more than 32 bits of RAM addressing. An app I wrote appears to require about 8 GB RAM for a full production run.
Yet, most of my applications ran fine on 16 bit machines.
I would have picked up an Alpha if I could have afforded it. This shows that just being the fastest machine on Earth does not get you a spot on the desktop. Did DEC offer last year's model for cheap? The only low end Alpha's were on the used market, as far as I know.
I would have picked up an AMD64 rather than an Athlon, if I could have gotten one cheap enough. My old machine died, so I had to upgrade. It was a surprise, and I didn't have much extra cash. I picked up a low end 1800+, but with a good motherboard, expandable to 3 GB RAM, and 3200+ chip. I may add RAM, but, so far, I'm happy with the speed.
On the other hand, when the average user asks me what the AMD model names mean, I tell them that my AMD 1800+ is more than twice as fast as the 1.6 Ghz Intel P4 at work. This P4 has 133 MHz RAM, which is less than half the bandwidth available to my 1800.
My points is that AMD's suggestive model numbers do not appear to be misleading. If anything, they're conservative. Yes, I understand that single numbers are a poor guide to performance. I also understand that the market wants single numbers. AMD is using Intel's numbers (MHz), and relating it to real performance in a conservative way.
For me, this builds trust. I still recall Intel's trust busting FDIV bug handling fiasco. Really, tell me again about how I don't care if my math is inaccurate. Why should I buy from them again?
Back to the comparison. My company bought cheap P4's off the shelf. What I did was buy a good motherboard. My guess is that I spent $20 more on it than was spent on the P4. I was able to do this because I put the box together myself. Had I picked up a "system", it wouldn't be $20, it would be $500 more, to get the same performance, as near as I can tell.
My company has more than 30,000 seats. They could easily assembly PCs themselves and save money. In fact, there are many things they could do to save money. The moral here is that big companies aren't necessarily very bright.
Open source licensing is better than Microsoft's licensing. License servers are a continual cause of downtime. Some sites have to hire a full time body just to track how many installs of various products are still on machines.
Help desks are often very unresponsive. For many companies, a quick look at the source code means that problems can be quickly solved.
Lower costs are better.
The idea that these features are somehow seperate from code quality is fiction.
I use Unix, Linux and the Mac because they are superior to Windows. I make the claim that there is ALWAYS a superior product over Windows.
Don't tell me, "Thank you for choosing MicroSoft". If I'm using it, you can bet it wasn't a choice.
In the old days, you put your NiCad batteries into the device, and ran it to zero. Then, you'd put the batteries on the charger for exactly six hours. If you forgot, you reduced the life and performance of the batteries.
You ran them to zero, because otherwise you didn't know how long to charge them.
Modern chargers know when the battery is full, and stop. Leave the battery on the charger for weeks and you now get fully charged batteries when you want them. And, you get 1,000 cycles instead of 50 to 100. Is the battery at an unknown charge level, or known to be half charged? No problem. Just throw it on the charger. 1,000 cycles is a cycle a week for 20 years.
Modern chargers do this for old batteries too.
So, with the old chargers, the batteries required vigilent maintenence. With new chargers, the batteries are at your convenience.
My digital camera takes about 100 pictures on a smart media card. It also takes about 100 pictures on one cycle of batteries. When I'm done shooting, I dump the pictures on my computer, and swap batteries to another set. If it was $2 for new batteries, I would take fewer pictures. For me, it's about a penny.
20 years ago. They didn't last as long as
disposable per charge, but were cheaper in
the long run due to reuse.
Then NiMh's came out, aproaching disposables
in per-charge performance. They were
something like 1400 milli-amp hours.
I bought them.
Then, better chargers came out, increasing
the number of cycles, not just for the new
batteries, but for my old ones too. I bought
a new charger.
Then, the 1600's came out. I bought them.
Then 1700's - I have some of those.
Then 1800's - yep.
Now I've seen 2200's. I'll be buying those
soon.
At the store, the cashier says, "but no
one buys recharagables, compared to disposables".
Of course not. They last 20 years instead
of a few days.
I get 8 hours out of my walkman for about
a penny. I'm not going back.
Modern NiMH formulations store about double what NiCads did per charge.
Modern chargers know when the battery is charged, and stop. So, now it really is typical to get 1,000 cycles out of them rather than less than 100. Modern chargers can do this for NiCad too - mine does.
Heat kills rechargable batteries. This can come from the charger, through overcharging. This can come from use - through high discharge rates.
NiMH batteries perform better in cold weather.
Kyoto holds the promise of a goal.