Let's just say that I admire how much resources it takes under NT to spawn one new process. In fact I'm positively astonished. A good thing? I think not.
Can I ask: in your tests of resource usage for process spawning in NT, which API did you use? There are multiple process spawning APIs available, and to get good performance you will need to avoid using the backwards compatible "CreateProcess" or "CreateProcessEx" APIs and go directly to the kernel level variants.
The CreateProcess family of calls are not designed for the specific case of forking an existing process, but instead are intended as a 'fork & exec' kind of call. Some systems (e.g. cygwin) have been known to use them to emulate fork, but this is not really a good way of doing it.
I'm also in awe of the way the NT kernel is virtualized and compartimentalized. Wait, it's not. You do know, don't you, that a Sun E15k with an arbitrary number of CPUs under Solaris can be split any which way (dynamically even) as virtual computers?
OK, so Solaris supports a feature that NT doesn't. I'm afraid I don't get your point. Different systems have different feature levels, and different price points. You can't expect total equivalence of such things. (Out of interest: can Solaris do this on x86 machines?)
Is it the TCP/IP stack that you admire? Hmm, where was that taken from again?
The Windows TCP/IP stack is not taken from BSD code, at least not the version in present editions of windows.
In fact what's so special about NT, with or without win32, that is so good? Is there a single piece that no one else has?
No other platform is so widely installed. There is little wrong with the platform (at least at the kernel level). The platform supports a wider variety of low-cost peripheral devices than any other. These factors are enough to make it the best platform for many applications. As long as your application doesn't require top end features that it doesn't implement (e.g. processor partitioning) or doesn't require it to be secure in a public-facing environment, it's a good choice. Even the latter problem can be mitigated substantially by installing a slimmed down version (e.g. XP Embedded).
Anyway, people split RARs for good reason. It's so that if during a massive download you have a small bit that doesn't stand up to CRC, all you need to do is redownload the segment that went bad instead of maybe 4.7gb all over again.
And then they distribute it via bittorrent, which does SHA-1 hash checks on each 1Mb chunk, and redownloads automatically until it gets it right.
3. Uninstaller want to reboot computer. WHAT? Reboot necessary for this? For God's sake - it is a simple compression utility. How much damage did the installation do to my system since a reboot is necessary to revert to my old setup?
This usually means that you're running on a Win9x, NT4 or 2K system and the installer upgraded a system DLL when you installed it. The uninstaller wants to downgrade it, but one of the applications you're running is using it, so the file is locked. The uninstaller writes instructions on what needs to be done into the registry, and it happens at next reboot.
It's basically because Windows' file locking is fsked up. Thankfully, they fixed it in XP.
Video game creators can retain control of their IP just the same as a musician can. But they routinely GIVE IT UP for profitability. If they want to do that, that's their perogative.
I don't think they want to do that. I think they'd rather have a well known publisher distribute their games without taking control of the copyright. There's no real reason why this can't happen; it's just that all the big players at the moment happen to work that way. But look at other fields: you don't get this kind of problem in the book publishing industry because book publishers realised a long time ago that allowing the authors to keep their copyright was a good way of attracting the better authors to them. It's only because the game (& music) industries are sewn up between a few small cartels that they can get away with treating their content creators like this.
The person who pays for the work deserves to own the work.
It depends on the way the work is done. I have no experience of working in the games industry, but it sounds to me like what this guy is talking about is a shift toward independent developers producing games and then persuading a publisher to distribute them: that is, a move to freelance game design. In this scenario, the person who designed the game should own it, because they did the work without assurance of earning any money from it.
This is the same idiotic logic where we have photographers owning the rights to YOUR wedding pics, even though you paid for them.
Unless they negotiated a contract with you in advance, that's bullshit. They can say what they want about industry standards (and they've tried this on me before, when my company had some photos produced for the web site of our client... an IP law firm), unless there's a contract in place, the person who asked for the work to be done (i.e. you) own the rights to the work.
Does the carpenter own the rights to your kitchen just because he builds the cabinets?
No, but he owns the rights to the templates he used to build it and can make an exact copy of it to sell to someone else, if he so desires. Unless you have a contract in place saying he can't.
I think that they will only embrace new technology if they can count the number of watchers and know stuff about them. P2P doesn't seem to be very conducive to this.
Neither is broadcasting the program over a radio signal. In both cases, the only real way to tell who's watching is to perform large scale surveys where you monitor people's activity and find out what they are watching. The infrastructure for this already exists, it just needs to be extended to cover (legally downloaded) P2P content.
That creates artefacts, though. You can't pull information out of thin air.
Typically to convert a 25fps source to a 29.97fps output, you'd slow the video down by about 4% (to 24fps, or "NTSC film"), and then perform a telecine interlacing operation to get an NTSC TV output. This is easily achievable with off-the-shelf software (even free software; VirtualDub can do this kind of thing with ease), and is exactly the same process that NTSC TV stations would use to broadcast programs from a PAL source (PAL stations usually receive their programming in NTSC film format, as this is how most programming is originally produced, and simply speed up by 4%).
If I have signed a licencing agreement with X over a (presumably popular, though I've never listened to it myself) show, wouldn't X be rightfully annoyed if I then start giving it away ?
And who exactly, in this scenario, is X? AFAIK (and I have read the FAQ that the GP linked to), the Archers is purely a BBC production.
The only "rights issue" I see here is that assuming they're on a fairly standard contract, the voice actors could possible demand repeat fees for every download. Essentially, the BBC just need to renegotiate those contracts, with internet availability in mind. I'm sure the actors wouldn't really expect something for nothing...?
I strongly recommend you hunt down both it, and its sequel Yes Prime Minister, then. They are without doubt among the best political satires ever produced. TorrentSpy.com seem to have at least some of them in their "TV / Misc - British Comedy" category. It isn't particularly necessary to watch them in order, although there are some references back to earlier episodes.
European phone carriers have researched the possibility of a dual mode phone that switch from GSM to DECT (digital cordless) depending on what's available, but they never commercialized it.
so he may be under the impression that you need a special kind of CD to make music CDs
If you have one of the Sony standalone CD recording devices, you do. I suspect this is the case for many amateur musicians, but probably not many others... so they're taxing exactly the wrong people!
I'm still prompted to reboot my Windows machine when I install userland software.
You are aware that there's a *lot* of software that tells you to reboot just because, well, it's traditional to do so, aren't you? They also do this if they install anything to run at startup/login time, or if they install but don't start a service (try start/run services.msc and start anything they seem to have installed).
Next time it happens, try running it without rebooting. I think you'll find that almost everything works if you do this.
Gee, I remember something called the Apple II doing this long before microsoft was the force it was. What a maroon.
While the Apple II was revolutionary, there were only a million installed at its peek. The PC (with MS supplied software) surpassed this within 3 years of its launch, and after the same length of time the Apple II took to reach that figure, had over ten times as many installed users. It didn't reach figures that I'd call "mass acceptance" until around '99 (another order of magnitude on top of that), which is the point at which I think it's safe to say that few people in the technologically advanced areas of the world did not have access to a personal computer of some kind (although it's still only one computer per... what, 10 people in america, europe and the pacific rim?).
However, the compiled headers option in Visual Studio is a "bug", IMO.
If you're developing on a machine older than about 4 years old and using any reasonably complex C++ library (e.g. MFC, Boost, etc.) it's a necessary evil. If you want your binary today, that is.
Or is Internet Explorer a very successful third-party trojan?
Yep. Look at the copyright notice:
Based on NCSA Mosaic. NCSA Mosaic(TM); was developed at the National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign.
Distributed under a licensing agreement with Spyglass, Inc. Contains security software licensed from RSA Data Security Inc. Portions of this software are based in part on the work of the Independent JPEG Group. Multimedia software components, including Indeo(R); video, Indeo(R) audio, and Web Design Effects are provided by Intel Corp.
So we can blame NCSA, Spyglass, RSA, IJG *and* Intel.;)
DJB has a mail system that does what he wants, and does it securely. He's not willing to accept patches into the mail distro because it taints its quality. That's his choice.
Fine. But unfortunately, it has resulted in large numbers of people running his mail server unmodified without fully understanding the implications of doing so. While I don't blame him for not accepting patches to qmail, my understanding is that it has been repeatedly pointed out to him that qmail's habit of sending bounce messages itself when it receives an undeliverable e-mail rather than sending a 550 response to instruct the sending MTA to do so causes problems for just about every e-mail user on the internet, and so far he has not fixed the issue. I'm sick to death of receiving qmail's bounce messages from thousands of different servers that I've never sent e-mail to, just because some fucking tiny minded spammer is using my address as the envelope sender address, and I hold DJB responsible for this. He has popularised an approach that is inferior to the method that was common to the one used before, which is not really a productive practice.
He wrote the code, he puts it up there. If you like it, use it. If you don't, avoid it.
I don't use it. But I still can't avoid the consequences of the fact that others around me do, which truly sucks.
As well, it keeps those who don't take the time to learn the basics of administration from operating misconfigured qmail installations.
As far as I'm concerned, all unpatched qmail installations are misconfigured.
I also hold him responsible for the fact that the results of his security hole discovery course last year were posted onto slashdot only days after the notifications of the vulnerabilities his students had discovered were sent to the maintainers of those projects. Best practice for vulnerability disclosure is to allow the maintainer at least 28 days to respond to your discovery before disclosing it publically, and you can't get much more public than the front page of slashdot.
Hmmm. Interesting, except that to exploit an XSS vulnerability, you would typically find a user of the site, and then persuade them to click on a crafted link with your exploit code embedded in it. Therefore you would normally target a user and check the sites they regularly visit. A honeypot would achieve nothing.
that's still too much, but half of the memory usage gone from a simple minimize/maximize!:)
All that does is swap the memory out. If there's anything actually being used on those pages (and due to fragmentation you'll frequently find a few useful items on a page that's mostly filled with leaked garbage) then they'll get paged back in pretty quickly.
of course, IE is a lot worse on this and many more fields, and while FF is actively developed, IE is not, even though micro$oft says it is.
Err.. if it isn't, how come they've just released a beta of a new version?
Slashdot Mirror
Let's just say that I admire how much resources it takes under NT to spawn one new process. In fact I'm positively astonished. A good thing? I think not.
Can I ask: in your tests of resource usage for process spawning in NT, which API did you use? There are multiple process spawning APIs available, and to get good performance you will need to avoid using the backwards compatible "CreateProcess" or "CreateProcessEx" APIs and go directly to the kernel level variants.
The CreateProcess family of calls are not designed for the specific case of forking an existing process, but instead are intended as a 'fork & exec' kind of call. Some systems (e.g. cygwin) have been known to use them to emulate fork, but this is not really a good way of doing it.
I'm also in awe of the way the NT kernel is virtualized and compartimentalized. Wait, it's not. You do know, don't you, that a Sun E15k with an arbitrary number of CPUs under Solaris can be split any which way (dynamically even) as virtual computers?
OK, so Solaris supports a feature that NT doesn't. I'm afraid I don't get your point. Different systems have different feature levels, and different price points. You can't expect total equivalence of such things. (Out of interest: can Solaris do this on x86 machines?)
Is it the TCP/IP stack that you admire? Hmm, where was that taken from again?
The Windows TCP/IP stack is not taken from BSD code, at least not the version in present editions of windows.
In fact what's so special about NT, with or without win32, that is so good? Is there a single piece that no one else has?
No other platform is so widely installed. There is little wrong with the platform (at least at the kernel level). The platform supports a wider variety of low-cost peripheral devices than any other. These factors are enough to make it the best platform for many applications. As long as your application doesn't require top end features that it doesn't implement (e.g. processor partitioning) or doesn't require it to be secure in a public-facing environment, it's a good choice. Even the latter problem can be mitigated substantially by installing a slimmed down version (e.g. XP Embedded).
Anyway, people split RARs for good reason. It's so that if during a massive download you have a small bit that doesn't stand up to CRC, all you need to do is redownload the segment that went bad instead of maybe 4.7gb all over again.
And then they distribute it via bittorrent, which does SHA-1 hash checks on each 1Mb chunk, and redownloads automatically until it gets it right.
There's no point in doing the same thing twice...
3. Uninstaller want to reboot computer.
WHAT? Reboot necessary for this? For God's sake - it is a simple compression utility. How much damage did the installation do to my system since a reboot is necessary to revert to my old setup?
This usually means that you're running on a Win9x, NT4 or 2K system and the installer upgraded a system DLL when you installed it. The uninstaller wants to downgrade it, but one of the applications you're running is using it, so the file is locked. The uninstaller writes instructions on what needs to be done into the registry, and it happens at next reboot.
It's basically because Windows' file locking is fsked up. Thankfully, they fixed it in XP.
Vista allows you to:
* Play Minesweeper.
Be fair. Vista allows you to play minesweeper and scale its window up to full screen.
Video game creators can retain control of their IP just the same as a musician can. But they routinely GIVE IT UP for profitability. If they want to do that, that's their perogative.
I don't think they want to do that. I think they'd rather have a well known publisher distribute their games without taking control of the copyright. There's no real reason why this can't happen; it's just that all the big players at the moment happen to work that way. But look at other fields: you don't get this kind of problem in the book publishing industry because book publishers realised a long time ago that allowing the authors to keep their copyright was a good way of attracting the better authors to them. It's only because the game (& music) industries are sewn up between a few small cartels that they can get away with treating their content creators like this.
Agreed that it's no excuse for piracy, though.
The person who pays for the work deserves to own the work.
It depends on the way the work is done. I have no experience of working in the games industry, but it sounds to me like what this guy is talking about is a shift toward independent developers producing games and then persuading a publisher to distribute them: that is, a move to freelance game design. In this scenario, the person who designed the game should own it, because they did the work without assurance of earning any money from it.
This is the same idiotic logic where we have photographers owning the rights to YOUR wedding pics, even though you paid for them.
Unless they negotiated a contract with you in advance, that's bullshit. They can say what they want about industry standards (and they've tried this on me before, when my company had some photos produced for the web site of our client... an IP law firm), unless there's a contract in place, the person who asked for the work to be done (i.e. you) own the rights to the work.
Does the carpenter own the rights to your kitchen just because he builds the cabinets?
No, but he owns the rights to the templates he used to build it and can make an exact copy of it to sell to someone else, if he so desires. Unless you have a contract in place saying he can't.
No... you sample the habits of the people your advertisers care about.
I think that they will only embrace new technology if they can count the number of watchers and know stuff about them. P2P doesn't seem to be very conducive to this.
Neither is broadcasting the program over a radio signal. In both cases, the only real way to tell who's watching is to perform large scale surveys where you monitor people's activity and find out what they are watching. The infrastructure for this already exists, it just needs to be extended to cover (legally downloaded) P2P content.
That creates artefacts, though. You can't pull information out of thin air.
Typically to convert a 25fps source to a 29.97fps output, you'd slow the video down by about 4% (to 24fps, or "NTSC film"), and then perform a telecine interlacing operation to get an NTSC TV output. This is easily achievable with off-the-shelf software (even free software; VirtualDub can do this kind of thing with ease), and is exactly the same process that NTSC TV stations would use to broadcast programs from a PAL source (PAL stations usually receive their programming in NTSC film format, as this is how most programming is originally produced, and simply speed up by 4%).
If I have signed a licencing agreement with X over a (presumably popular, though I've never listened to it myself) show, wouldn't X be rightfully annoyed if I then start giving it away ?
And who exactly, in this scenario, is X? AFAIK (and I have read the FAQ that the GP linked to), the Archers is purely a BBC production.
The only "rights issue" I see here is that assuming they're on a fairly standard contract, the voice actors could possible demand repeat fees for every download. Essentially, the BBC just need to renegotiate those contracts, with internet availability in mind. I'm sure the actors wouldn't really expect something for nothing...?
At least with digital TV broadcasting, they didn't have much choice -- there's a legally mandated switchover scheduled not too far in the future...
BBC are already doing this for many of their radio programs, and have been for some time. See here for details.
I never managed to see [...] Yes Minister
I strongly recommend you hunt down both it, and its sequel Yes Prime Minister, then. They are without doubt among the best political satires ever produced. TorrentSpy.com seem to have at least some of them in their "TV / Misc - British Comedy" category. It isn't particularly necessary to watch them in order, although there are some references back to earlier episodes.
European phone carriers have researched the possibility of a dual mode phone that switch from GSM to DECT (digital cordless) depending on what's available, but they never commercialized it.
Ahem
I recall visiting IBM's offices in Warwick, UK around '97 or '98; they had something similar then. It certainly isn't a new idea.
so he may be under the impression that you need a special kind of CD to make music CDs
If you have one of the Sony standalone CD recording devices, you do. I suspect this is the case for many amateur musicians, but probably not many others... so they're taxing exactly the wrong people!
I'm still prompted to reboot my Windows machine when I install userland software.
You are aware that there's a *lot* of software that tells you to reboot just because, well, it's traditional to do so, aren't you? They also do this if they install anything to run at startup/login time, or if they install but don't start a service (try start/run services.msc and start anything they seem to have installed).
Next time it happens, try running it without rebooting. I think you'll find that almost everything works if you do this.
Gee, I remember something called the Apple II doing this long before microsoft was the force it was. What a maroon.
While the Apple II was revolutionary, there were only a million installed at its peek. The PC (with MS supplied software) surpassed this within 3 years of its launch, and after the same length of time the Apple II took to reach that figure, had over ten times as many installed users. It didn't reach figures that I'd call "mass acceptance" until around '99 (another order of magnitude on top of that), which is the point at which I think it's safe to say that few people in the technologically advanced areas of the world did not have access to a personal computer of some kind (although it's still only one computer per... what, 10 people in america, europe and the pacific rim?).
[Source]
However, the compiled headers option in Visual Studio is a "bug", IMO.
If you're developing on a machine older than about 4 years old and using any reasonably complex C++ library (e.g. MFC, Boost, etc.) it's a necessary evil. If you want your binary today, that is.
Yep. Look at the copyright notice:
So we can blame NCSA, Spyglass, RSA, IJG *and* Intel.
DJB has a mail system that does what he wants, and does it securely. He's not willing to accept patches into the mail distro because it taints its quality. That's his choice.
Fine. But unfortunately, it has resulted in large numbers of people running his mail server unmodified without fully understanding the implications of doing so. While I don't blame him for not accepting patches to qmail, my understanding is that it has been repeatedly pointed out to him that qmail's habit of sending bounce messages itself when it receives an undeliverable e-mail rather than sending a 550 response to instruct the sending MTA to do so causes problems for just about every e-mail user on the internet, and so far he has not fixed the issue. I'm sick to death of receiving qmail's bounce messages from thousands of different servers that I've never sent e-mail to, just because some fucking tiny minded spammer is using my address as the envelope sender address, and I hold DJB responsible for this. He has popularised an approach that is inferior to the method that was common to the one used before, which is not really a productive practice.
He wrote the code, he puts it up there. If you like it, use it. If you don't, avoid it.
I don't use it. But I still can't avoid the consequences of the fact that others around me do, which truly sucks.
As well, it keeps those who don't take the time to learn the basics of administration from operating misconfigured qmail installations.
As far as I'm concerned, all unpatched qmail installations are misconfigured.
I also hold him responsible for the fact that the results of his security hole discovery course last year were posted onto slashdot only days after the notifications of the vulnerabilities his students had discovered were sent to the maintainers of those projects. Best practice for vulnerability disclosure is to allow the maintainer at least 28 days to respond to your discovery before disclosing it publically, and you can't get much more public than the front page of slashdot.
Hmmm. Interesting, except that to exploit an XSS vulnerability, you would typically find a user of the site, and then persuade them to click on a crafted link with your exploit code embedded in it. Therefore you would normally target a user and check the sites they regularly visit. A honeypot would achieve nothing.
Hmmm. I propose that the next discovered satellite should be named "Traycee".
[Yes, I know where the name Charon comes from.)
that's still too much, but half of the memory usage gone from a simple minimize/maximize! :)
All that does is swap the memory out. If there's anything actually being used on those pages (and due to fragmentation you'll frequently find a few useful items on a page that's mostly filled with leaked garbage) then they'll get paged back in pretty quickly.
of course, IE is a lot worse on this and many more fields, and while FF is actively developed, IE is not, even though micro$oft says it is.
Err.. if it isn't, how come they've just released a beta of a new version?
[Firefox] doesn't garbage collect in a timely manner.
It doesn't use GC at all -- firefox uses the mozilla code base which uses standard platform malloc & free implementations. Source for the mozilla memory allocater here.