I guess you're right. It is basically a scope-definition thing. But at the same time, it's a much easier sell to upper management to say, "I'm offloading that responsibility on Microsoft, a stable, financially healthy, readily available corporation."
Stable and financially healthy? Ok, I'll give you that (some people would debate that, due to assorted evidence of varying levels of book-cooking). Readily available? That I'd argue with. Have you dealt with Microsoft support? Did you see the comparison someone did of MS phone support versus the Psychic Friends Network? I actually had to call them at a previous job (company head's laptop was having nothing but problems, IBM said "call Microsoft"). The Psychic Friends Network would have been about as helpful.
And even if you _do_ find a legitimate issue? Unless you can make it financially worth their while (i.e., either go public, if it's serious enough, or be a very large customer), Microsoft will just tell you "hm. doesn't work. tough, don't do that." This is not what I'd call responsible.
Too easy to use? As others have said in the past, Windows makes easy tasks easy, but anything reasonably complicated? Better get ready to shell out the bucks. Windows breaks? Hm. How much does your company spend on MS each year? It better be a pretty big sum for MS to care. Don't like the fact that Windows has to talk to MS home base, or the fact that Windows and Office XP have to re-authorize on anything but the smallest hardware change? Tough.
Unix is easy to use too - the whole "easy to use" thing is a misnomer. There's a higher learning curve with Unix than Windows, yes, but once you've learned it, there's a whole lot more you can do with it too. Most Windows users can't deal with even a minor computer problem - they don't learn how to use the machine, but "if you click here, here and here, this will happen". That's not really learning, just simple application of patterns. They don't get an extensible skillset they can adapt to new tasks, but have to be retrained every time a new Office revision comes out, because "I can't click in the same places now!" or "everything looks different! help!". How is that better?
How does that do anything? Applications can still capture Ctrl-Alt-Del. Real secure systems have what is called a secure attention key, or SAK - it's a key combo that, no matter where a session is, if it is pressed, the open session will be _immediately_ terminated and a system login prompt will be displayed. That is not the case with Windows' poor attempt at replicating that feature.
If you consider MSDN either useful or good, yours certainly are as well. I've tried using MSDN in the past - their search feature is abysmal (as are all searches on MS's sites, unfortunately), and their API docs are woefully inadequate and vague. At least with Linux, if there's a doubt about what some function does, I can dive into the source code and figure it out. Win32 programming on the other hand is an exercise in copy-and-paste programming, because nobody seems to understand those 10-argument functions that MS programmers seem so fond of.
DRI also seems to break security. I have to block permissions of/dev/dri/ for any untrusted user, which means the user can't access it for 3D programs.
Well, if you block access to those devices, that is what will happen. You can setup a group of users, and control the owning group and mode of the/dev/dri/* devices, through XF86Config, which could help security some though...
Domain sockets are _not_ slow. They're in fact way, way faster than using TCP sockets. This is in no small part because the overhead of setting up and tearing down TCP sockets is avoided, plus pushing data into and pulling it out of the networking stack for stuff that's not going to the network anyway.
If you're going to talk like you know something about the subject, it'd be nice if you actually had some experience first.
The original "Alice in Wonderland" was not quite as light-hearted as the Disney movie based upon the book. American McGee definitely has a way or darkening the tone quite a lot, but to call the original book "light-hearted" is a bit inaccurate.
For example, recursively searching through a subtree for some text in a file.
Gee, it wasn't all that long ago that there WAS NO WAY to do that in Windows or DOS. At all. At least, certainly not without third party tools. Only recently did Windows' Find/Files and Folders get that ability. 'grep' has been around since I started using Unix/Linux, and it just does its job - that being of the many jobs that this deceptively simple tool can do
Nano is not the "updated" edition of pico. It's a GPL'd clone of pico. (pico is under the UW license, which isn't compatible with true free-software licenses or the DFSG definition).
Actually, Microsoft bought VB (it was originally a product called Ruby - no, different Ruby). As with pretty much all of Microsoft's other products, it was bought from someone else, stamped with Microsoft's name, and released as their own creation. About the only product that was of their own invention completely was MS Bob - and we all know how that turned out, don't we...
What do IRQ conflicts have to do with DirectX? That was because of PCI and ISA-PnP (less of an issue now, since ISA is just about completely dead). Especially PCI, since it got rid of the stupidity of edge-triggered interrupts, replacing them with level-triggered interrupts. That allowed IRQ channels to be shareable. And that's an advantage whether you're running Windows or Linux, definitely not an OS feature.
NTFS _is_ HPFS. Or was, anyway - that was one of the things that Microsoft took away from the storied OS/2 split. The original NTFS in Windows NT 3.1 literally _was_ HPFS, however - you stuck an HPFS volume in an NT system, booted it, it'd mount up and be readable just like an NTFS volume because IT WAS THE SAME CODE. Recent versions have diverged from HPFS feature-wise, but the first NTFS wasn't just a cheap rip of HPFS - it was the same thing.
"commercialized FreeBSD"? What the hell is that moron blathering about now? SunOS 4 was BSD-derived, and it was derived from BSD lineage WAY before FreeBSD happened. Solaris as it exists today is SysV all the way. Sun isn't commercializing FreeBSD.
Someone needs to take away Ballmer's crack pipe. He's fried what little brains he still had.
Dunno about Ed Furlong, but I do know that Linda Hamilton expressed no interest in being involved. I think just about the _only_ person to carry over from T2 is Ahh-nold. Amazing he's still up for playing the Terminator again at his age, really.
She actually wrote a completely separate title, "Essential Windows NT System Administration". I think that speaks volumes about what a different animal sysadmining on Windows "servers" is - all these disparate UNIX and UNIX-like platforms can be covered in one (good, big) tome, but Windows needs a separate text altogether.
I can't entirely agree with the "it should have lasted another season" sentiment. Not that I don't like Cowboy Bebop - I do. I own all 6 DVDs. I _love_ it. But I think that's, if anything, one good thing about Japanese anime serials - they say "it's going to be this long", and it's that long. Too many American TV shows get stretched, and stretched, and stretched, until the plot is so thin, and the characters so worn out, that we can't stand the story anymore. I didn't entirely like the way the series ended, but I'm glad they didn't just keep making more and more episodes.
Cooperative multitasking isn't worthless. Actually, many OSes use it internally - but that's the difference. It's one thing to use it in a group of well-reviewed internal components, where it can be reasonably trusted that each will give up the CPU to the others and allow the system to continue functioning. It's another when all programs use it, and you have code from a wide variety of sources that must be trusted to properly concede access to the CPU (like with Win16 apps).
Well, so did Windows 95. Except that pretty much the entire GDI (as well as part of the USER subsystem) remained 16-bit code, so they had to run in that 16-bit cooperatively-multitasked subsystem. So if a 16-bit app blocked - oops! There goes your entire GDI.
I personally do that all the time, and ignored how annoying it is because I got used to it. With Safari, it works like this: 1. Click the link.
Gee, just the way it works when I'm reading my email in Evolution, and I click a link to have a page open in Phoenix.
Now, about tabbed browsers. In general, I hate tabbed interfaces for the following reasons: 1. Too much screen real estate. I have a 17" wide-screen, but I STILL don't want to waste it.
You can turn tabs off in the browsers that support them.
2. Tabs don't scale well. Beyond the discussion in the article, regarding where new tabs should appear, what happens when there are too many tabs for one row? Multiple row tabs take up even more screen real estate, and present a perplexing interface, since the rows must rearrange themselves as tabs are added and deleted. (The only alternative is horizontal scrolling of the tabs -- hideous!)Imagine twenty tabs called "Slashdot..." Which one did you want to pick? Your only choice will be to click at random, and so it's no better than Apple-~ to cycle through all open windows
If there are too many tabs, open another window. And as for twenty Slashdot tabs, um, that's what hover text is for - hover the mouse over the tab you want to open, and you get a tooltip with the full title of the page attached to that tab.
XTI/TLI/SysV STREAMS is for all intents and purposes dead. There is an occasional reason to use an implementation of it to port legacy software that's built around STREAMS, but there are commercial implementations for that purpose, that are implemented fully in userspace. And really, the Berkeley Sockets API isn't sufficiently difficult or convoluted, IMO, nor is XTI/TLI/STREAMS a sufficient simplification, that replacing the Berkeley APIs ever really became worth it.
And Windows 2000 for the Alpha _never_ shipped. It was killed before the release. So they may have beta'd it, but it was never a product you could buy. (Not that this makes me feel bad in the slightest.:) Besides the fact that apps that even ran natively in Windows NT on Alpha were few and far between - BackOffice was one of the few, and practically no third party Windows software did - the very reason why FX!32 was so prominently mentioned in any Windows NT for Alpha sales pitch. (The place I was working back in '96-'97 was looking into buying an AlphaServer system, and one company we approached was very into Windows NT on Alpha, and FX!32 was frequently mentioned. We never bought from them.)
Slashdot Mirror
I guess you're right. It is basically a scope-definition thing. But at the same time, it's a much easier sell to upper management to say, "I'm offloading that responsibility on Microsoft, a stable, financially healthy, readily available corporation."
Stable and financially healthy? Ok, I'll give you that (some people would debate that, due to assorted evidence of varying levels of book-cooking). Readily available? That I'd argue with. Have you dealt with Microsoft support? Did you see the comparison someone did of MS phone support versus the Psychic Friends Network? I actually had to call them at a previous job (company head's laptop was having nothing but problems, IBM said "call Microsoft"). The Psychic Friends Network would have been about as helpful.
And even if you _do_ find a legitimate issue? Unless you can make it financially worth their while (i.e., either go public, if it's serious enough, or be a very large customer), Microsoft will just tell you "hm. doesn't work. tough, don't do that." This is not what I'd call responsible.
Too easy to use? As others have said in the past, Windows makes easy tasks easy, but anything reasonably complicated? Better get ready to shell out the bucks. Windows breaks? Hm. How much does your company spend on MS each year? It better be a pretty big sum for MS to care. Don't like the fact that Windows has to talk to MS home base, or the fact that Windows and Office XP have to re-authorize on anything but the smallest hardware change? Tough.
Unix is easy to use too - the whole "easy to use" thing is a misnomer. There's a higher learning curve with Unix than Windows, yes, but once you've learned it, there's a whole lot more you can do with it too. Most Windows users can't deal with even a minor computer problem - they don't learn how to use the machine, but "if you click here, here and here, this will happen". That's not really learning, just simple application of patterns. They don't get an extensible skillset they can adapt to new tasks, but have to be retrained every time a new Office revision comes out, because "I can't click in the same places now!" or "everything looks different! help!". How is that better?
How does that do anything? Applications can still capture Ctrl-Alt-Del. Real secure systems have what is called a secure attention key, or SAK - it's a key combo that, no matter where a session is, if it is pressed, the open session will be _immediately_ terminated and a system login prompt will be displayed. That is not the case with Windows' poor attempt at replicating that feature.
If you consider MSDN either useful or good, yours certainly are as well. I've tried using MSDN in the past - their search feature is abysmal (as are all searches on MS's sites, unfortunately), and their API docs are woefully inadequate and vague. At least with Linux, if there's a doubt about what some function does, I can dive into the source code and figure it out. Win32 programming on the other hand is an exercise in copy-and-paste programming, because nobody seems to understand those 10-argument functions that MS programmers seem so fond of.
The error correction on audio CDs is done in hardware, not software. CD audio is linear PCM audio. The math works like this:
44100 (samples/sec) * 16 (bits/sample) * 2 (channels) / 8 (bits/byte) / 75 (frames/sec) = 2532 (bytes/frame)
And the frame (or block, if you will) size of a CD when speaking in Mode2 terms is 2352 (also, the block size of an MPEG system stream, you'll note).
DRI also seems to break security. I have to block permissions of /dev/dri/ for any untrusted user, which means the user can't access it for 3D programs.
/dev/dri/* devices, through XF86Config, which could help security some though...
Well, if you block access to those devices, that is what will happen. You can setup a group of users, and control the owning group and mode of the
Domain sockets are _not_ slow. They're in fact way, way faster than using TCP sockets. This is in no small part because the overhead of setting up and tearing down TCP sockets is avoided, plus pushing data into and pulling it out of the networking stack for stuff that's not going to the network anyway.
If you're going to talk like you know something about the subject, it'd be nice if you actually had some experience first.
The original "Alice in Wonderland" was not quite as light-hearted as the Disney movie based upon the book. American McGee definitely has a way or darkening the tone quite a lot, but to call the original book "light-hearted" is a bit inaccurate.
For example, recursively searching through a subtree for some text in a file.
Gee, it wasn't all that long ago that there WAS NO WAY to do that in Windows or DOS. At all. At least, certainly not without third party tools. Only recently did Windows' Find/Files and Folders get that ability. 'grep' has been around since I started using Unix/Linux, and it just does its job - that being of the many jobs that this deceptively simple tool can do
Nano is not the "updated" edition of pico. It's a GPL'd clone of pico. (pico is under the UW license, which isn't compatible with true free-software licenses or the DFSG definition).
Dennis Ritchie is one of the original developers of Unix. What part of his "anti-foreword" doesn't make sense?
Actually, Microsoft bought VB (it was originally a product called Ruby - no, different Ruby). As with pretty much all of Microsoft's other products, it was bought from someone else, stamped with Microsoft's name, and released as their own creation. About the only product that was of their own invention completely was MS Bob - and we all know how that turned out, don't we...
What do IRQ conflicts have to do with DirectX? That was because of PCI and ISA-PnP (less of an issue now, since ISA is just about completely dead). Especially PCI, since it got rid of the stupidity of edge-triggered interrupts, replacing them with level-triggered interrupts. That allowed IRQ channels to be shareable. And that's an advantage whether you're running Windows or Linux, definitely not an OS feature.
NTFS _is_ HPFS. Or was, anyway - that was one of the things that Microsoft took away from the storied OS/2 split. The original NTFS in Windows NT 3.1 literally _was_ HPFS, however - you stuck an HPFS volume in an NT system, booted it, it'd mount up and be readable just like an NTFS volume because IT WAS THE SAME CODE. Recent versions have diverged from HPFS feature-wise, but the first NTFS wasn't just a cheap rip of HPFS - it was the same thing.
"commercialized FreeBSD"? What the hell is that moron blathering about now? SunOS 4 was BSD-derived, and it was derived from BSD lineage WAY before FreeBSD happened. Solaris as it exists today is SysV all the way. Sun isn't commercializing FreeBSD.
Someone needs to take away Ballmer's crack pipe. He's fried what little brains he still had.
Dunno about Ed Furlong, but I do know that Linda Hamilton expressed no interest in being involved. I think just about the _only_ person to carry over from T2 is Ahh-nold. Amazing he's still up for playing the Terminator again at his age, really.
Funny, considering I just played the trailer in Xine, and I got video and audio playback.
She actually wrote a completely separate title, "Essential Windows NT System Administration". I think that speaks volumes about what a different animal sysadmining on Windows "servers" is - all these disparate UNIX and UNIX-like platforms can be covered in one (good, big) tome, but Windows needs a separate text altogether.
It's just evolution at work, dontcha know.
I can't entirely agree with the "it should have lasted another season" sentiment. Not that I don't like Cowboy Bebop - I do. I own all 6 DVDs. I _love_ it. But I think that's, if anything, one good thing about Japanese anime serials - they say "it's going to be this long", and it's that long. Too many American TV shows get stretched, and stretched, and stretched, until the plot is so thin, and the characters so worn out, that we can't stand the story anymore. I didn't entirely like the way the series ended, but I'm glad they didn't just keep making more and more episodes.
Cooperative multitasking isn't worthless. Actually, many OSes use it internally - but that's the difference. It's one thing to use it in a group of well-reviewed internal components, where it can be reasonably trusted that each will give up the CPU to the others and allow the system to continue functioning. It's another when all programs use it, and you have code from a wide variety of sources that must be trusted to properly concede access to the CPU (like with Win16 apps).
Well, so did Windows 95. Except that pretty much the entire GDI (as well as part of the USER subsystem) remained 16-bit code, so they had to run in that 16-bit cooperatively-multitasked subsystem. So if a 16-bit app blocked - oops! There goes your entire GDI.
I personally do that all the time, and ignored how annoying it is because I got used to it. With Safari, it works like this:
1. Click the link.
Gee, just the way it works when I'm reading my email in Evolution, and I click a link to have a page open in Phoenix.
Now, about tabbed browsers. In general, I hate tabbed interfaces for the following reasons:
1. Too much screen real estate. I have a 17" wide-screen, but I STILL don't want to waste it.
You can turn tabs off in the browsers that support them.
2. Tabs don't scale well. Beyond the discussion in the article, regarding where new tabs should appear, what happens when there are too many tabs for one row? Multiple row tabs take up even more screen real estate, and present a perplexing interface, since the rows must rearrange themselves as tabs are added and deleted. (The only alternative is horizontal scrolling of the tabs -- hideous!)Imagine twenty tabs called "Slashdot..." Which one did you want to pick? Your only choice will be to click at random, and so it's no better than Apple-~ to cycle through all open windows
If there are too many tabs, open another window. And as for twenty Slashdot tabs, um, that's what hover text is for - hover the mouse over the tab you want to open, and you get a tooltip with the full title of the page attached to that tab.
XTI/TLI/SysV STREAMS is for all intents and purposes dead. There is an occasional reason to use an implementation of it to port legacy software that's built around STREAMS, but there are commercial implementations for that purpose, that are implemented fully in userspace. And really, the Berkeley Sockets API isn't sufficiently difficult or convoluted, IMO, nor is XTI/TLI/STREAMS a sufficient simplification, that replacing the Berkeley APIs ever really became worth it.
And Windows 2000 for the Alpha _never_ shipped. It was killed before the release. So they may have beta'd it, but it was never a product you could buy. (Not that this makes me feel bad in the slightest. :) Besides the fact that apps that even ran natively in Windows NT on Alpha were few and far between - BackOffice was one of the few, and practically no third party Windows software did - the very reason why FX!32 was so prominently mentioned in any Windows NT for Alpha sales pitch. (The place I was working back in '96-'97 was looking into buying an AlphaServer system, and one company we approached was very into Windows NT on Alpha, and FX!32 was frequently mentioned. We never bought from them.)