Microsoft's Windows Security Update CD is great in theory, but almost worthless in practice. The lead time for delivery is so long, by the time you get the CD, another batch of viruses/worms are out exploiting newly discovered vulnerabilities.
I'm a private pilot, and I flew for Washington Air Search And Rescue for a few years. I was always amazed at how hard it can be to find a downed aircraft, even though they are all required to carry functional emergency locator transmitters. In all of the missions I flew, not one had a "happy ending". In fact, in my first mission (a missing Cessna 152 with instructor and student) the plane was never found.
Finding an individual must be even more difficult. Someone on this thread suggested good cellular coverage, and they got modded "funny". It's not a bad idea. I've heard numerous stories of hikers lost in the Cascades getting found by using their cell phones.
There's a huge difference. IBM is (currently) trying to "play nice" with the Open Source community, and would probably prefer to not piss them off. Microsoft does not play nicely and apparently doesn't care who they piss off.
Predicting a "patent battle" between Microsoft and the Open Source development community is like predicting terrorist attacks. It's not a matter of "if", it's a matter of "when".
I cannot imagine IBM going after Open Source developers for patent infringement. Like other people on this thread have mentioned, it's just not in IBM's best interest, doesn't fit with their business model, etc. But Microsoft...
Microsoft has claimed repeatedly that their patent portfolio is "purely defensive". This is also the message spread internally to get uncooperative engineers to contribute to patent applications. At least, that was my experience.
The real question is: At what point will Microsoft feel the need to respond "defensively" to the Linux threat?
3/5 through the lifespan of the 2.6 kernel? WTF?? The kernel after 2.6.9 will be 2.6.10, not 2.7.0. For example, the current 2.4 kernel is 2.4.26...
2.6.0 is not as reliable as 2.4.26 because the latter has had 26 updates to get things "right". This is just the way things work in kernel development.
The 2.6 series has broken a few things, largely because:
There were a lot of major, architectural changes between 2.4 and 2.6. It probably should have been called 3.0.
A few kernel features that have been marked as "deprecated" in previous kernels have now been officially removed. This should have surprised no one. The kernel team is neither arbitrary nor capricious (for the most part:-).
The kernel is, obviously, a single point of failure. If a new kernel is deployed on 1,000,000 machines and only 0.1% have problems, that's still 1,000 people complaining on the mailing lists.
Would I run 2.6 on a mission-critical highly-buzzword-enabled enterprise server? Not yet. Do I run it on my desktop? Abso-fucking-lutely.
To summarize the essay: A suboptimal implementation of an optimal algorithm usually beats an optimal implementation of a suboptimal algorithm.
Also, the author is clearly enjoying the fruits of those "optimization aware" programmers that created Erlang, especially those "cycle-counters" who wrote the virtual machine.
So the Word team organized a special dev team that focused entirely on WordPerfect document import, "reverse-engineering" the WordPerfect file format (documentation for which was jealously guarded, as was the norm back then).
In all versions of Windows NT, whenever an unhandled exception (such as a NULL pointer deref) occurs in kernel-mode, the system blue screens. There is no attempt to cleanup or repair; the system is dead.
In Linux, whenever an unhandled exception occurs in kernel-mode, the kernel attempts to kill the offending process. Sometimes the system is so hozed that recovery is impossible. But sometimes recovery is possible.
I suspect this is one of the reasons Linux is more stable than Windows.
Way back in 1990, I joined Microsoft to work on OS/2 2.0. After I was there for just a few months, the "big split" happened, and IBM assumed all OS/2 development, and Microsoft focused on NT -- and I transferred to the NT group. There was, shall we say, a little animosity between some of ex-OS/2 folks and the NT folks.
One day, someone's computer fried, and the unmistakable stench of burned PC board wafted through the building. I saw a group of ex-OS/2 engineers returning from lunch. One stopped, sniffed the air, and said "Hmm... It smells like someone tried to boot NT on a 386SX."
In general (IIRC), a patent application must be filed with the patent office no more than one year after the invention was first released to the public. As someone in an earlier message noted, Microsoft released a beta version of Win95 with long filename support sometime in 1994, approximately 2 years before the patent application was filed. If these dates are indeed true, then the patent is invalid.
Of course, I can't speak for all groups at Microsoft, but the group I was in was very aware of this "one year grace period" for patents. We always noted exactly when a "patentable idea" was first mentioned to the public, even if it was just a Powerpoint slide at a trade show. The day the first non-Microsoftee hears about an idea is the day the patent clock starts running.
Again, I'm sure different groups within Microsoft handle this differently, but if the Win95 group didn't file the patent application within the one year grace period, then they fucked up in a big way.
Of course, in a refreshing turn of events, Microsoft's fuck up could turn out to be a blessing for the rest of the industry...
Microsoft did something like this (on a much smaller scale) about 8 years ago. A few hundred idle computers scattered around the MS campus created the Chicken Crossing animation that was presented at Siggraph 96.
In fact, UI is not hard anymore (since we don't have to use the Xt object model, the most overengineered piece of object-oriented crap that ever came out of an ivory tower). Instead we have simple UIs and simple object -event models like KDE's components and QT's slots to hide the complexity (most of the time), and vastly more examples of consistent and market-persistent UI designs since back in the day, making UI design and implementation so dead simple the bulk of the time that any barely or even not quite competent coder is without excuse.
If, as you say, UI is not hard anymore, why are there so many applications around with sucky UIs? You seem to be implying the "hard part" of UI is getting pixels on the screen. And you're right -- that was hard, now it's easy.
But good UI design is hard. It is based on psychology and a deep understanding of human cognition. Good UI design is based on solid theory, backed up by formal usability testing in controlled conditions.
Apple has a 20 year history of great UI design because they can afford to hire people like Bruce Tognazzini and Donald Norman. They can also afford to staff a great usability lab, and they take the time to include feedback from the lab in their product's design cycles.
The OSS community needs more people like Tognazzini and Norman.
Download the LFS book, plus the all of the GNU/Linux source tarballs, patches, etc.
Build you own basic Linux system, from scratch.
If you're feeling adventurous, you can also read the "Beyond Linux From Scratch" book and build X11, Gnome or KDE (or both), etc.
You will certainly learn a ton about Linux systems in general just by following the LFS recipe. Plus, you'll end up with a system that has exactly what you want on it, and if you don't like the way something works, you'll know exactly what to fix.
You are correct about the language skills. My niece is in the US Navy. Her areas of expertise are cryptanalysis and Arabic language. Guess where she's been spending her time lately?
Out of curiousity, who *does* buy retail MS software in Poland?
Excellent question. Honestly, I have no idea. I'm sure many people do, but the ones I know cannot afford it.
I have in front of me a flyer from a local computer retailer. Their low-end budget system has a 1.4GHz AMD Duron, 128M RAM, 40GB HD, 52X CDROM, etc. Pretty basic stuff. The cost (without monitor or OS) is 1099 ZL (about $289 US). The cost with Windows XP Home is 1538 ZL (about $404 US). Is Windows really worth nearly 40% of the cost of the hardware? Hardly.
The high local price for XP may be due, in part, to Microsoft's licensing policies for OEMs. The local builders don't generate the volume to qualify for Microsoft's lowest pricing, so those people who want Microsoft get screwed.
Luckily, most system builders around here make thier computers available without an OS, so there's no "Microsoft Tax" if one wants to run Linux or BSD. (I guess there's also no Microsoft Tax if one wants to run an illegal copy of Windows...)
"Piracy" (or whatever you want to call it) is a real problem here in Poland. I honestly think most people here would prefer to own a legitmate copy of Windows. However, when a retail copy of Windows XP Pro costs more than your monthly apartment rent, it's hard to ignore the $5 copy available at the local open-air market.
And, FWIW, a retail copy of the latest Office Pro costs about 20-25% of the average yearly salary here.
This is one thing (of many) that I really hate about Microsoft. They don't adjust their prices based on the local economy (at least they don't do it here). Many other products sold here seem to have their prices adjusted based on the local standard of living. Not Microsoft software.
I think this is a great potential growth area for free/open source software: make quality software available to people who simply cannot afford Microsoft. For most of us here on Slashdot (I suspect) $45 US is probably our latte budget for the week. For many families here in Poland, $45 US (~170 Zl) is their grocery budget for a week or two.
The oh-so-lame FTP client was always based on BSD. It's pretty gross code, and no one wanted to touch it. Only the minimum work was done to make it build and (sort of) run under NT.
The telnet client has a strange history. At one point (back in NT 3.1 and 3.5, I think) it was basically just the normal Windows terminal app. Someone had the bright (?) idea to have it talk to a virtual serial port. So, a driver implemented the virtual serial port, which talked to a user-mode helper service, which issued the socket calls to talk to the remote telnet server. Now, just imagine the ring transitions, imagine the context switches, imagine the buffer copies. And, of course, imagine the total lack of performance.
Later (NT 3.51 maybe?) the virtual serial port bullshit was ripped out, and the terminal app was hacked into being a (slightly less lame) telnet client.
For Win2K (maybe) and XP (for sure) a "new" telnet client was shipped. I've never seen the source to this version, so I do not know its origin. However, looking at its structure in a debugger, I would guess that it's not BSD based. It uses threads and async I/O, handles different language locales, etc. In other words, it appears very un-BSD like. But that's just my hunch.
Slashdot Mirror
Microsoft's Windows Security Update CD is great in theory, but almost worthless in practice. The lead time for delivery is so long, by the time you get the CD, another batch of viruses/worms are out exploiting newly discovered vulnerabilities.
I'm a private pilot, and I flew for Washington Air Search And Rescue for a few years. I was always amazed at how hard it can be to find a downed aircraft, even though they are all required to carry functional emergency locator transmitters. In all of the missions I flew, not one had a "happy ending". In fact, in my first mission (a missing Cessna 152 with instructor and student) the plane was never found.
Finding an individual must be even more difficult. Someone on this thread suggested good cellular coverage, and they got modded "funny". It's not a bad idea. I've heard numerous stories of hikers lost in the Cascades getting found by using their cell phones.
There's a huge difference. IBM is (currently) trying to "play nice" with the Open Source community, and would probably prefer to not piss them off. Microsoft does not play nicely and apparently doesn't care who they piss off.
Predicting a "patent battle" between Microsoft and the Open Source development community is like predicting terrorist attacks. It's not a matter of "if", it's a matter of "when".
...and his name is Microsoft.
I cannot imagine IBM going after Open Source developers for patent infringement. Like other people on this thread have mentioned, it's just not in IBM's best interest, doesn't fit with their business model, etc. But Microsoft...
Microsoft has claimed repeatedly that their patent portfolio is "purely defensive". This is also the message spread internally to get uncooperative engineers to contribute to patent applications. At least, that was my experience.
The real question is: At what point will Microsoft feel the need to respond "defensively" to the Linux threat?
3/5 through the lifespan of the 2.6 kernel? WTF?? The kernel after 2.6.9 will be 2.6.10, not 2.7.0. For example, the current 2.4 kernel is 2.4.26...
2.6.0 is not as reliable as 2.4.26 because the latter has had 26 updates to get things "right". This is just the way things work in kernel development.
The 2.6 series has broken a few things, largely because:
Would I run 2.6 on a mission-critical highly-buzzword-enabled enterprise server? Not yet. Do I run it on my desktop? Abso-fucking-lutely.
...they hate that.
To summarize the essay: A suboptimal implementation of an optimal algorithm usually beats an optimal implementation of a suboptimal algorithm.
Also, the author is clearly enjoying the fruits of those "optimization aware" programmers that created Erlang, especially those "cycle-counters" who wrote the virtual machine.
Of course, the odds are the computers used back in 1963 had UPPER CASE letters only. Real programmers don't need lower case!
I tried to enter this message in just upper case, but it pissed off the lameness filter. Lame.
I have a theory. No, really!
In all versions of Windows NT, whenever an unhandled exception (such as a NULL pointer deref) occurs in kernel-mode, the system blue screens. There is no attempt to cleanup or repair; the system is dead.
In Linux, whenever an unhandled exception occurs in kernel-mode, the kernel attempts to kill the offending process. Sometimes the system is so hozed that recovery is impossible. But sometimes recovery is possible.
I suspect this is one of the reasons Linux is more stable than Windows.Reminds me of a story...
Way back in 1990, I joined Microsoft to work on OS/2 2.0. After I was there for just a few months, the "big split" happened, and IBM assumed all OS/2 development, and Microsoft focused on NT -- and I transferred to the NT group. There was, shall we say, a little animosity between some of ex-OS/2 folks and the NT folks.
One day, someone's computer fried, and the unmistakable stench of burned PC board wafted through the building. I saw a group of ex-OS/2 engineers returning from lunch. One stopped, sniffed the air, and said "Hmm... It smells like someone tried to boot NT on a 386SX."
Let me guess -- there's a computer buried in there somewhere, and you're stress-testing it's ability to withstand static discharge?
Anyone interested in the virtualization problems in the x86 architecture should read "Analysis of the Intel Pentium's Ability to Support a Secure Virtual Machine Monitor". Interesting, but I don't think the situation is as dire as minus_273 seems to think.
In general (IIRC), a patent application must be filed with the patent office no more than one year after the invention was first released to the public. As someone in an earlier message noted, Microsoft released a beta version of Win95 with long filename support sometime in 1994, approximately 2 years before the patent application was filed. If these dates are indeed true, then the patent is invalid.
Of course, I can't speak for all groups at Microsoft, but the group I was in was very aware of this "one year grace period" for patents. We always noted exactly when a "patentable idea" was first mentioned to the public, even if it was just a Powerpoint slide at a trade show. The day the first non-Microsoftee hears about an idea is the day the patent clock starts running.
Again, I'm sure different groups within Microsoft handle this differently, but if the Win95 group didn't file the patent application within the one year grace period, then they fucked up in a big way.
Of course, in a refreshing turn of events, Microsoft's fuck up could turn out to be a blessing for the rest of the industry...
Wolfpack is already dead, and it already has an OS. Sorry.
Microsoft did something like this (on a much smaller scale) about 8 years ago. A few hundred idle computers scattered around the MS campus created the Chicken Crossing animation that was presented at Siggraph 96.
If, as you say, UI is not hard anymore, why are there so many applications around with sucky UIs? You seem to be implying the "hard part" of UI is getting pixels on the screen. And you're right -- that was hard, now it's easy.
But good UI design is hard. It is based on psychology and a deep understanding of human cognition. Good UI design is based on solid theory, backed up by formal usability testing in controlled conditions.
Apple has a 20 year history of great UI design because they can afford to hire people like Bruce Tognazzini and Donald Norman. They can also afford to staff a great usability lab, and they take the time to include feedback from the lab in their product's design cycles.
The OSS community needs more people like Tognazzini and Norman.
Shakespeare was not using logic, he was using algebra. He was answering the question "what is the square root of 4*b^2?"
Or...
- Install a live CD (like Knoppix).
- Visit the Linux From Scratch project site.
- Download the LFS book, plus the all of the GNU/Linux source tarballs, patches, etc.
- Build you own basic Linux system, from scratch.
- If you're feeling adventurous, you can also read the "Beyond Linux From Scratch" book and build X11, Gnome or KDE (or both), etc.
You will certainly learn a ton about Linux systems in general just by following the LFS recipe. Plus, you'll end up with a system that has exactly what you want on it, and if you don't like the way something works, you'll know exactly what to fix.On a related note, LANL has announced a plan to begin above-ground weapon testing in the Salt Lake City area...
You are correct about the language skills. My niece is in the US Navy. Her areas of expertise are cryptanalysis and Arabic language. Guess where she's been spending her time lately?
Excellent question. Honestly, I have no idea. I'm sure many people do, but the ones I know cannot afford it.
I have in front of me a flyer from a local computer retailer. Their low-end budget system has a 1.4GHz AMD Duron, 128M RAM, 40GB HD, 52X CDROM, etc. Pretty basic stuff. The cost (without monitor or OS) is 1099 ZL (about $289 US). The cost with Windows XP Home is 1538 ZL (about $404 US). Is Windows really worth nearly 40% of the cost of the hardware? Hardly.
The high local price for XP may be due, in part, to Microsoft's licensing policies for OEMs. The local builders don't generate the volume to qualify for Microsoft's lowest pricing, so those people who want Microsoft get screwed.
Luckily, most system builders around here make thier computers available without an OS, so there's no "Microsoft Tax" if one wants to run Linux or BSD. (I guess there's also no Microsoft Tax if one wants to run an illegal copy of Windows...)
"Piracy" (or whatever you want to call it) is a real problem here in Poland. I honestly think most people here would prefer to own a legitmate copy of Windows. However, when a retail copy of Windows XP Pro costs more than your monthly apartment rent, it's hard to ignore the $5 copy available at the local open-air market.
And, FWIW, a retail copy of the latest Office Pro costs about 20-25% of the average yearly salary here.
This is one thing (of many) that I really hate about Microsoft. They don't adjust their prices based on the local economy (at least they don't do it here). Many other products sold here seem to have their prices adjusted based on the local standard of living. Not Microsoft software.
I think this is a great potential growth area for free/open source software: make quality software available to people who simply cannot afford Microsoft. For most of us here on Slashdot (I suspect) $45 US is probably our latte budget for the week. For many families here in Poland, $45 US (~170 Zl) is their grocery budget for a week or two.
That's recreational firearm enthusiast, you insensitive clod!
The oh-so-lame FTP client was always based on BSD. It's pretty gross code, and no one wanted to touch it. Only the minimum work was done to make it build and (sort of) run under NT.
The telnet client has a strange history. At one point (back in NT 3.1 and 3.5, I think) it was basically just the normal Windows terminal app. Someone had the bright (?) idea to have it talk to a virtual serial port. So, a driver implemented the virtual serial port, which talked to a user-mode helper service, which issued the socket calls to talk to the remote telnet server. Now, just imagine the ring transitions, imagine the context switches, imagine the buffer copies. And, of course, imagine the total lack of performance.
Later (NT 3.51 maybe?) the virtual serial port bullshit was ripped out, and the terminal app was hacked into being a (slightly less lame) telnet client.
For Win2K (maybe) and XP (for sure) a "new" telnet client was shipped. I've never seen the source to this version, so I do not know its origin. However, looking at its structure in a debugger, I would guess that it's not BSD based. It uses threads and async I/O, handles different language locales, etc. In other words, it appears very un-BSD like. But that's just my hunch.