It just occured to me that the correct open source solution to the SCO problem is for someone to develop an open source replacement for restaurant management.
That way, McDonalds franchises can enjoy an open source solution intead of purchasing SCO products. This seems to be Darl's favorite customer, he mentions them in nearly everytime he talks.
Develop open source solutions for SCO's top 3 cusmtomers and their stock price will go back where it belongs...
I'm just thinking, if a few high volume websites were to include links to a 1-pixel GIF at bogus randomly generated domains, wouldn't sitefinder crumble, and zero the value of any collected statistics on domain name typos.
I think this would be much more entertaining than lawsuits or BIND hacks...
The thing I really hate about this HAVI stuff is that the CE companies are ignoring what the rest of the world knows about abstraction and layering network protocols.
You're missing the point to argue the merits of ethernet or 1394. The point is that this is a layer-4 protocol and should have nothing to do woth the physical or link layers. HAVI should be orthogonal to the physical layer.
Can you imagine the dark cave we would still be living in if TCP was somehow specifically bound to ethernet?
I think these HAVI guys are hung up on DRM, and feel that they might let the genie out of the bottle if they abstract the physical and link layers out of their protocols to run on any link layer.
As a result, this will wind up like previous CE standards effort and fail to provide a consistent interoperable system.
I'm amazed by the historical similarities between Apple/NeXT/Jobs and Palm/Handspring/Hawkins.
Jobs/Hawkins creates visionary company with great products, both are amazing achievements.
Frustrated by the big corporate thing Jobs/Hawkins goes off and starts NeXT/HS to do the next big thing enjoying much hype and early success.
Apple/Palm lose market share, start making boring products, bumble through some bad business decisions, change CPUs to PPC/ARM and generally suffer from a lack of vision.
Apple/Palm buys NeXT/HS for way more than it's worth to return the visionary guy to the fold.
So, I wonder if we can use this to predict the future:
Palm switches from ancient OS to Unix based OS. Palm puts out some new interesting products. Palm's core customers become bigger fans than ever. Palm becomes increasingly irrelevant as Microsoft increases market share.
(I really wanted to fit the big soft drink executive and the movie company into the story, but I just couldn't find the analogy.)
A really interesting guy on this topic is Tom Van Baak, the fellow that runs leapsecond.com. As a measure of the level of obsession a person can obtain, this guy has multiple cesium frequency standards, but he had to go out and buy a crazy russian hydrogen maser so he could get better than a microsecond a year accuracy. He's also got some interesting information about the leapsecond debate on his website.
This thing reminds me of an interview with Steven Spielberg when the "back to the future" movies came out in the 80's. He said that it's really easy to write about an apocolyptic future, but hard work to imagine a happy world in the future.
Maybe it's because we tend to idealize the past and forget about the horrible aspects of life 50-200 years ago. Maybe this sets a trend line where the past was great, the present is not as good, so the future must be hellish if we extrapolate far enough.
It's not quite that bad. The original EPIA boards use an integrates graphics/northbridge, the graphics core is a Trident Cyberblade I1. The xree86 acceleration works fine for this and is nice, but you are limited to 1280x1024x16.
The newer EPIA-M use an integrated northbridge/ graphics called the CLE266. The graphics core is some internal thing to VIA called Castlerock.
There's no external public documentation for castlerock, but there is a binary-only xfree86 available from VIA. I've been using this with redhat 8.0 after a little fiddling and it's reasonably fast.
Silly. The military has plenty of bandwidth, they buy up commercial bandwidth during conflicts to keep bad guys from using it. They probably buy up bandwidth and then sell it back to CNN and other friendly services, maybe at a loss, but it's just another mechanism for controlling information in the battlefield.
I haven't been there in about four years and at the time Ed Groethus, the owner, didn't seem to be in such great health, but I think the place is still there.
The Black Whole near Los Alamos, NM is quite a site. Most of the surplus stuff is from the nearby national lab. The place has been there maybe 40 years and is filled with crazy nuclear related stuff. Ed Groethus, the guy that owns the place seems to be very fond of much of his junk, so it can be tricky to get him to part with the good stuff.
Everyone calls the place "The Black Hole", but I think the real name is something more boring like "Los Alamos Salvage". It's definitely worth the trip if you're within 500 miles and are mesmerized by bits of shiny metal.
Beautiful. For Valentine's day I think you should put an LCD in the lid of the case and make a post-modern laptop out of the thing. I've got some parts for you if you're up for it.
I've been thinking about this a bit and I think I'm okay with the printer companies. I think it's probably the fault of media companies like PC Magazine or CNET that there's not a standard benchmark for printer speed or cartridge lifetime. If one of the well read publications would set a standard for reviews, the printer companies would fall in line and report the standard benchmarks, without silly legislation.
Sure, the printer companies run a racket, but I'm okay with them because the rules to their racket are basically clear when you buy the printer: We make cheap printers and expensive cartridges, we have a private interface between the printer and the cartridge, we're tricky people and we will do our best to keep this interface private so we can make more money.
I kind of like it, there is plenty of competition and plenty of good companies to choose from. Before I spend $70 on a cartridge I look and see what the other companies are up to. It seems to be an efficient market and the printer I have now is about a million times better than the printer I had 10 years ago.
Now, compared to something like the music industry, the printer guys are saints. The rules to the music industry are something like: We sign young artists to ripoff contracts. We pay radio stations a lot of money through independent producers to play songs over the free airwaves many many times, but we will sue you if you give a copy of a song to your friends. We think it's legal to make an analog cassette of a CD but illegal to make a digital copy of a CD. We would like to legislate that copying bits isn't always legal. Now I'm getting wound up...
Wall's religion reminds me of the joke at
the beginning of the film Annie Hall.
The one about the man who goes to a psychologist and says "Doc, My brother thinks he's a chicken." The Doctor says, Bring him him, I will cure him." The guy says, "I can't, we need the eggs."
Evidently this was written for a 200MHz strongARM. Does anyone how much of the CPU is required on a Zaurus? More specifically, will it run on the 74MHz Cirrus ARM7 used in the Rio600, Rio Receiver, etc.
I think the whole thing is a case of odd behavior caused by steroid use. The Schwarzenegger film, the workout equipment, the obsessive workout schedule, angry physical outbursts. I would love to see before and after pictures of those guys. I bet they were using steriods.
The next time someone says "... on steroids", this is what they're talking about!
It's important to differentiate between architecture optimizations and CPU specific optimizations. The ARMv5 instruction set is a relatively minor architectural tweak to the ARMv4 instruction set. The names give you the impression that it's some grand change between v4 and v5, if a technical guy did the naming it would be ARMv4 and ARMv4.01. ARM is playing some games with architecture naming to protect their business position with patents in a silly way.
ARMv5 adds a couple of new instructions over v4, an instruction to count leading zeros in a register (which a compiler would likely never use), and a better method of switching between the ARM instruction set and the 16-bit Thumb instruction set. The later isn't relevant for PocketPC since Thumb mode isn't supported. I think v5 might having a new debugging hook as well.
The new XScale parts are ARMv5te, the T is for the 16-bit Thumb instruction set, which no one seems to care about. The "E" adds some DSP oriented instructions that are pretty interesting for media codecs and such. They are the MMX equivalent for the ARM world. They likely won't improve performance of the general purpose aspects of the platform.
I think it's a red herring to chase Microsoft for not optimizing for the ARMv5, the changes are really small and I don't see any performance impact, certainly not if you have to maintain another version for all of the strongARM based products.
Now, as far as CPU specific optimizations for the PXA250 (XScale) implementation of the ARM architecture. IMHO Intel chased MHz and left behind a lot of good sense about system performance. The high order bit is bus performance as others have already pointed out.
In addition to the bus performance, Intel made many tradeoffs to optimize for clock speed: The 7-stage pipe has a 4-clock penalty for a mis-predicted branch. This is compared to the circuit design heroics in the strongARM that implements "all branches are 2-cycles". The Xscale approach is much more complicated, it probably doesn't perform any better, but you get a high clock speed.
Intel adds clock cycles to all load/store-multiple instructions in Xscale. This is a pretty big deal in ARM since they are used in the entry and exit of most C functions, in memcpy(), and any time you are moving chunks bigger than a register.
The load-use penalty is bigger in Xscale. This is a pretty big deal in ARM. The ARM instruction set is pretty compact. It is a RISC processor, but the combination of shifting operations combined with ALU operations makes it possible for a good compiler to generate reasonably compact code. As a result, it's harder for a compiler to put instructions between a load and instructions that use the destination of the load. This is another trade-off in Xscale that allows a higher clock speed but hurts performance otherwise.
I go on too long, but the DEC designed strongARM used in the SA1100 is a tour-de-force of clean implementation and balanced system performance. It's amazing that core was designed in 1993 (I think, someone please correct me) and is still the leader for handheld apps. The Intel guys went after clock speed at the expense of everything else in Xscale and it will probably never optimize well for a platform like PocketPC.
I wonder if amateur athletes could have a
GPL-like license for the depecition of
their image in competition? Anyone can show
images of the althlete in competition provided
that they don't restrict the ability of others
to distribute the image to others, something
like that.
I like the way the MAD MPEG audio decoder is
licensed. It's GPL code, but the author is open about licensing it to commercial closed-source projects. I think this is a nice arrangement:
You know the funny thing. Porter and Duff still work for Pixar and share the same CEO as Apple...
I think the Porter/Duff paper is the last word on compositing.
jeff
There are two versions to collect statistics
on
More Copy Protected CDs?
·
· Score: 5, Informative
It's pretty clear to me why records companies are distributing both clean and copy protected versions of the same CD. They are collecting statistics on consumer acceptance for the copy protection scheme. They release two versions and then look at the return rates for the two versions to figure out how many people have CD players that reject the copy protection scheme.
Slashdot Mirror
It just occured to me that the correct open source solution to the SCO problem is for someone to develop an open source replacement for restaurant management.
That way, McDonalds franchises can enjoy an open source solution intead of purchasing SCO products. This seems to be Darl's favorite customer, he mentions them in nearly everytime he talks.
Develop open source solutions for SCO's top 3 cusmtomers and their stock price will go back where it belongs...
jeff
I'm just thinking, if a few high volume websites were to include links to a 1-pixel GIF at bogus randomly generated domains, wouldn't sitefinder crumble, and zero the value of any collected statistics on domain name typos.
I think this would be much more entertaining than lawsuits or BIND hacks...
jeff
The thing I really hate about this HAVI stuff is that the CE companies are ignoring what the rest of the world knows about abstraction and layering network protocols.
You're missing the point to argue the merits of ethernet or 1394. The point is that this is a layer-4 protocol and should have nothing to do woth the physical or link layers. HAVI should be orthogonal to the physical layer.
Can you imagine the dark cave we would still be living in if TCP was somehow specifically bound to ethernet?
I think these HAVI guys are hung up on DRM, and feel that they might let the genie out of the bottle if they abstract the physical and link layers out of their protocols to run on any link layer.
As a result, this will wind up like previous CE standards effort and fail to provide a consistent interoperable system.
jeff
I'm amazed by the historical similarities between Apple/NeXT/Jobs and Palm/Handspring/Hawkins.
Jobs/Hawkins creates visionary company with great products, both are amazing achievements.
Frustrated by the big corporate thing Jobs/Hawkins goes off and starts NeXT/HS to do the next big thing enjoying much hype and early success.
Apple/Palm lose market share, start making boring products, bumble through some bad business decisions, change CPUs to PPC/ARM and generally suffer from a lack of vision.
Apple/Palm buys NeXT/HS for way more than it's worth to return the visionary guy to the fold.
So, I wonder if we can use this to predict the future:
Palm switches from ancient OS to Unix based OS.
Palm puts out some new interesting products.
Palm's core customers become bigger fans than ever.
Palm becomes increasingly irrelevant as Microsoft increases market share.
(I really wanted to fit the big soft drink executive and the movie company into the story, but I just couldn't find the analogy.)
jeff
It's the same marketing thought process that puts "fat free" on a box of sugar.
jeff
A really interesting guy on this topic is Tom Van Baak, the fellow that runs leapsecond.com. As a measure of the level of obsession a person can obtain, this guy has multiple cesium frequency standards, but he had to go out and buy a crazy russian hydrogen maser so he could get better than a microsecond a year accuracy. He's also got some interesting information about the leapsecond debate on his website.
Me, I'm a simple guy, I just need to keep NTP locked to a couple of microseconds to sleep well.
jeff
This thing reminds me of an interview with Steven Spielberg when the "back to the future" movies came out in the 80's. He said that it's really easy to write about an apocolyptic future, but hard work to imagine a happy world in the future.
Maybe it's because we tend to idealize the past and forget about the horrible aspects of life 50-200 years ago. Maybe this sets a trend line where the past was great, the present is not as good, so the future must be hellish if we extrapolate far enough.
jeff
It's not quite that bad. The original EPIA
boards use an integrates graphics/northbridge,
the graphics core is a Trident Cyberblade I1.
The xree86 acceleration works fine for this and
is nice, but you are limited to 1280x1024x16.
The newer EPIA-M use an integrated northbridge/
graphics called the CLE266. The graphics core
is some internal thing to VIA called Castlerock.
There's no external public documentation for
castlerock, but there is a binary-only xfree86
available from VIA. I've been using this with
redhat 8.0 after a little fiddling and it's
reasonably fast.
jeff
Silly. The military has plenty of bandwidth, they buy up commercial bandwidth during conflicts to keep bad guys from using it. They probably buy up bandwidth and then sell it back to CNN and other friendly services, maybe at a loss, but it's just another mechanism for controlling information in the battlefield.
I haven't been there in about four years and at the time Ed Groethus, the owner, didn't seem to be in such great health, but I think the place is still there.
The Black Whole near Los Alamos, NM is quite a site. Most of the surplus stuff is from the nearby national lab. The place has been there maybe 40 years and is filled with crazy nuclear related stuff. Ed Groethus, the guy that owns the place seems to be very fond of much of his junk, so it can be tricky to get him to part with the good stuff.
Everyone calls the place "The Black Hole", but I think the real name is something more boring like "Los Alamos Salvage". It's definitely worth the trip if you're within 500 miles and are mesmerized by bits of shiny metal.
jeff
Here's another media player for the GBA. It
looks like this one is shipping.
Songpro (annoying website warning)
jeff
Beautiful. For Valentine's day I think you
should put an LCD in the lid of the case and
make a post-modern laptop out of the thing.
I've got some parts for you if you're up for
it.
jeff
Captain Opteron - "Dump as a bag of hammers..."
jeff
I've been thinking about this a bit and I think I'm okay with the printer companies. I think it's probably the fault of media companies like PC Magazine or CNET that there's not a standard benchmark for printer speed or cartridge lifetime. If one of the well read publications would set a standard for reviews, the printer companies would fall in line and report the standard benchmarks, without silly legislation.
Sure, the printer companies run a racket, but I'm okay with them because the rules to their racket are basically clear when you buy the printer: We make cheap printers and expensive cartridges, we have a private interface between the printer and the cartridge, we're tricky people and we will do our best to keep this interface private so we can make more money.
I kind of like it, there is plenty of competition and plenty of good companies to choose from. Before I spend $70 on a cartridge I look and see what the other companies are up to. It seems to be an efficient market and the printer I have now is about a million times better than the printer I had 10 years ago.
Now, compared to something like the music industry, the printer guys are saints. The rules to the music industry are something like: We sign young artists to ripoff contracts. We pay radio stations a lot of money through independent producers to play songs over the free airwaves many many times, but we will sue you if you give a copy of a song to your friends. We think it's legal to make an analog cassette of a CD but illegal to make a digital copy of a CD. We would like to legislate that copying bits isn't always legal. Now I'm getting wound up...
jeff
The one about the man who goes to a psychologist and says "Doc, My brother thinks he's a chicken." The Doctor says, Bring him him, I will cure him." The guy says, "I can't, we need the eggs."
jeff
Evidently this was written for a 200MHz strongARM.
Does anyone how much of the CPU is required on
a Zaurus? More specifically, will it run on the
74MHz Cirrus ARM7 used in the Rio600, Rio Receiver,
etc.
jeff
I think the whole thing is a case of odd behavior caused by steroid use. The Schwarzenegger film, the workout equipment, the obsessive workout schedule, angry physical outbursts. I would love to see before and after pictures of those guys. I bet they were using steriods.
The next time someone says "... on steroids", this is what they're talking about!
jeff
It's important to differentiate between architecture optimizations
and CPU specific optimizations. The ARMv5 instruction set is a
relatively minor architectural tweak to the ARMv4 instruction set.
The names give you the impression that it's some grand change between
v4 and v5, if a technical guy did the naming it would be ARMv4 and
ARMv4.01. ARM is playing some games with architecture naming
to protect their business position with patents in a silly way.
ARMv5 adds a couple of new instructions over v4, an instruction to count
leading zeros in a register (which a compiler would likely never
use), and a better method of switching between the ARM instruction
set and the 16-bit Thumb instruction set. The later isn't
relevant for PocketPC since Thumb mode isn't supported. I think
v5 might having a new debugging hook as well.
The new XScale parts are ARMv5te, the T is for the 16-bit Thumb
instruction set, which no one seems to care about. The "E" adds
some DSP oriented instructions that are pretty interesting for
media codecs and such. They are the MMX equivalent for the ARM
world. They likely won't improve performance of the general
purpose aspects of the platform.
I think it's a red herring to chase Microsoft for not optimizing for
the ARMv5, the changes are really small and I don't see any
performance impact, certainly not if you have to maintain another
version for all of the strongARM based products.
Now, as far as CPU specific optimizations for the PXA250 (XScale)
implementation of the ARM architecture. IMHO Intel chased
MHz and left behind a lot of good sense about system performance.
The high order bit is bus performance as others have already
pointed out.
In addition to the bus performance, Intel made many tradeoffs
to optimize for clock speed: The 7-stage pipe has a 4-clock penalty
for a mis-predicted branch. This is compared to the circuit
design heroics in the strongARM that implements "all branches
are 2-cycles". The Xscale approach is much more complicated, it
probably doesn't perform any better, but you get a high clock speed.
Intel adds clock cycles to all load/store-multiple instructions
in Xscale. This is a pretty big deal in ARM since they are
used in the entry and exit of most C functions, in memcpy(),
and any time you are moving chunks bigger than a register.
The load-use penalty is bigger in Xscale. This is a pretty big
deal in ARM. The ARM instruction set is pretty compact. It is a
RISC processor, but the combination of shifting operations
combined with ALU operations makes it possible for a good compiler
to generate reasonably compact code. As a result, it's harder
for a compiler to put instructions between a load and instructions
that use the destination of the load. This is another trade-off
in Xscale that allows a higher clock speed but hurts performance
otherwise.
I go on too long, but the DEC designed strongARM used in the SA1100
is a tour-de-force of clean implementation and balanced system
performance. It's amazing that core was designed in 1993 (I think,
someone please correct me) and is still the leader for handheld
apps. The Intel guys went after clock speed at the expense of
everything else in Xscale and it will probably never optimize well
for a platform like PocketPC.
jeff
Okay, but I'm only meeting in buildings where
the architectural drawings are freely available
by a license that RMS approves of.
jeff
I wonder if amateur athletes could have a
GPL-like license for the depecition of
their image in competition? Anyone can show
images of the althlete in competition provided
that they don't restrict the ability of others
to distribute the image to others, something
like that.
jeff
Rep. Boucher has a brilliant idea here. Send him a note of support, especially if you live in the 9th district of Virginia: Rep. Rick Boucher jeff
I like the way the MAD MPEG audio decoder is
licensed. It's GPL code, but the author is open about licensing it to commercial closed-source projects. I think this is a nice arrangement:
http://www.mars.org/home/rob/proj/mpeg/
jeff
"You've got hunger!"
You know the funny thing. Porter and Duff still work for Pixar and share the same CEO as Apple...
I think the Porter/Duff paper is the last word on compositing.
jeff
It's pretty clear to me why records companies are distributing both clean and copy protected versions of the same CD. They are collecting statistics on consumer acceptance for the copy protection scheme. They release two versions and then look at the return rates for the two versions to figure out how many people have CD players that reject the copy protection scheme.
jeff