And start appending '/1', '/2' and so on if you have multiple burners.
Mac OS X has kernel support for sending generic SCSI commands to a particular device. For non-Mac OS X, imagine if you could open a regular/dev/sr device node and do an ioctl to access the generic SCSI interface, but only to that device. (And there's no reason why an ATAPI device couldn't give you the same thing; after all, ATAPI is basically SCSI over ATA.)
That would also mean you wouldn't have to have your CD burning program running as root and able to access every device on the bus, just because one of them is a CD burner.
Actually all POSIX-compatible systems come with a Korn-compatible shell. In/bin/sh.
Sun and Linux are notably different on this one. Linux isn't UNIX, and if you don't like it, return your downloaded ISO for a full refund. Solaris puts the POSIX shell in/usr/xpg4/bin/sh, which really sucks when you're trying to write a portable shell script. (No/bin/ksh by default on linux,/bin/sh isn't always POSIX, and only Solaris (IME) has the POSIX programs totally separate from the OS programs. Yippee!)
This would seem to suggest that they aren't as responsible for the problems as many seem to think they are;
What Microsoft is lacking, aside from a "forbid first and allow later" approach to security, is sufficient analysis of their software's failure modes.
I'm sure whoever came up with that thing in the registry where you can cause a program to be run automatically before every.exe file is launched thought that it was very elegant and symmetrical. After all, that's how all non-.exe files are handled. In fact, they probably didn't even do anything special for.exe, it just fell out of the code that allows you to CreateProcess("foo.doc") or CreateProcess("movie.avi") and have the system magically load the right program first. (Kind of like #! on UNIX, but with control in the registry instead of in the file.)
Only problem is, CreateProcess("operating_system.exe") should really not be hijacked to run a virus before running the actual program.
Neither Mac OS Classic nor any of the UNIX-alikes (nor AmigaDOS, nor even MS-DOS actually) have anything like this. I'm pretty sure OS/2 doesn't either. It's unique to Windows, and it is the sort of thing that gives Symantec a large market for AV software.
How about "Windows Scripting Host"? I've been uninstalling that wherever I find it. (OK, 3 machines I have access to so far.) Haven't found anything that doesn't work without it. But you know what? It's been used for worms and viruses.
And Outlook's penchant for downloading things from e-mail messages. Apple's Mail.app has a setting for "do not download"--if the image is IN the message, it will display, if it has to be downloaded, no go. That download can be used to confirm an e-mail address--you all know this.
Then there's "active content" in e-mail. ActiveX and VBScript just expose too much of the system, through too few safeguards.
And how about the "always listening" ports on a client-only host? Why does a Windows 98 box on a home computer come up with any listening ports? They aren't needed unless your in a corporate LAN; the other desktop OS of the day had 0 (zero) listening ports unless the user turned on a server program.
Suffice to say, flaws in the RPC protocol which is connected to those always listening ports have been used to propagate malware recently. But can you even turn that software off? Try "net stop server"; you lose the Computer Browser too; and there's STILL an RPC port listening!
Surely the debate has little if anything to do with the protocol, it has everything to do with Yahoo's server base that sits behind their IM client and the business model that they have in place that sustains those servers.
Having recently had to switch to the truly abysmal Yahoo! Mac OS X client, I can't figure out which business model you could possibly be referring to.
The official Mac OS client, while being a good example of a bad UI:
Does not display advertisements.
Does not automatically open a web browser to any of Yahoo!'s sites (it will if you click on certain things).
Does not require any sort of payment to use.
Does play distorted sounds at 150% relative system volume (not adjustable), with a very noticable pop at the start and finish (bad sample editing--non-zero at ends?)
Does look like someone converted a Windows program to Mac OS Classic and then to Mac OS X and didn't understand any of the UI conventions on the 3 systems.
If that's a business model, I have no sympathies for them. I do have friends on Yahoo!, so I'll be using something other than Yahoo!'s awful client to connect. (Hopefully Proteus 3, but Fire will do nicely if Proteus doesn't make me happy.)
RedHat and IBM both 'patch by subsystem', instead of updating the entire OS. So does Sun.
An IBM AIX 'system update' is just a bunch of previously-released patches all grouped together. Sun would call that a 'patch cluster'. Sometimes IBM will repackage the patches so they have a coordinated version number. Since RedHat's package versions are all over the place, there's no point for them.
ELF isn't even enough to specify an ABI. ELF simply gives you the linker and loader format.
You still have to deal with minor issues like: - Which way does the stack grow? - Which register is the stack pointer (not always dictated by the hardware, especially on RISC chips) - Which register is used for globals? How is global data accessed? (TOC and GOT are two techniques; load-time address mapping is another one.) - How are structure members laid out in memory? Padding and alignment requirements are influenced by the hardware, but that doesn't always mean the ABI is the most obvious interpretation of the hardware specs. - What function arguments are in registers, which are on the stack? How are "ellipsis" functions handled? How are K&R argument promotions handled? How are aggregates passed? Are small aggregates (such as char[4]) are passed in a register, on the stack, or by pointer. Same with large floats, is a quad float passed by address or value? - Setjmp/longjmp, how do they work? - How does a stack frame look? If this isn't standard, exception handling can't unwind the stack, debuggers can't do a backtrace, and so on. - Where's the heap? Register pointer, fixed segment, what?
While it is possible to have an ABI that is common across operating systems on the same CPU architecture, it is impossible to have the same ABI across CPU architectures. You just can't use R31 as a stack pointer on IA32; there isn't one. You can't use SPARC register windows on PowerPC. And so on.
I've used an OLD Apple USB keyboard on a Windows machine quite successfully. And on Linux, and on my PS2. The gotchas are already mentioned; with Meta and Alt being where they would be on any UNIX machine, and backwards from the Windows location.
The OTHER gotcha is that the old USB keyboards, Apple or otherwise, did not generate separate keysyms for left and right qualifiers. You could, therefore, not distinguish L_ALT from R_ALT, L_SHIFT from R_SHIFT, and so on. Sucks for games.
New keyboards distinguish left and right qualifiers. Other than that, the poster who said USB HIDs are USB HIDs has it; the protocol doesn't care where the buttons are.
My reading of the Act in question lead me to the important conclusion:
You may make copies of audio recordings you own (and give them away), as long as you receive nothing in return. No swapping, nothing. Just the goodness of your heart.
I couldn't find anything that would prevent you from making the copy on blank media given to you by someone else.
As for P2P, I suppose one could argue that the sending computer is copying the work as it sends it... but is a computer acting on behalf of its owner good enough for this? You wouldn't be able to make stuff you downloaded available for upload, unless you also owned a copy of the work in question.
The law says specifically audio recordings, but in Canada, we have a history of broadening laws to cover similar situations--such as copying DVDs. (The proposed tariff includes DVDs; it hasn't been approved--yet.) I would be interested in finding out what a judge would decide on a DVD movie copied following the same restrictions as an audio program. Especially since the prerecorded discs cost about the same.
Yeah and they could call it "USB". Why hasn't someone thought of this?
That's just the physical interconnect. How about the actual bits sent over the wire?
Oh wait, that's called PostScript. And Adobe charged so much it never made it on the low end.
Plus, a lot of hardware makers hide problems with their hardware in the driver software. At least, I can't think of any other reason why they have so many different drivers for the same task; unless they're all just control freaks and won't work with NIH tech.
register variables may not have their address taken, which does give the compiler a hint that can improve optimization. It can also then diagnose a later programmer changing the routine to take the address of that variable; once the address is taken of a variable, several easy optimizations can't be done. (Most of which involve keeping the value in a... register.)
In my copy of 2.4.20, the first slide is arch/ia64/sn/io/ate_utils.c; and the start of the file, after the GPL, is "Copyright (C) 1992 - 1997, 2000-2002 Silicon Graphics, Inc. All righs reserved." The second slide appears to be Later That Same Function.
And if SCO is a Utah-based company, why are they programming in Greek?
So I suppose we'll see SGI getting sued next.
Also, someone else said that code is from BSD. Could it not also be from a "Getting Started with your New Itanium" kit from Intel?
Wait... Itanium... SCO... IBM... that was Montery, right? Wasn't that the project to make SCO, AIX, and Linux all more compatible, and to run on the Itanium?
That quote does not say that you may make only one additional copy. In fact, it implies that many archival copies may be made: "...and that all archival copies are destroyed...".
Although much of the text is singular, without an explicit clause saying something like, "may make or authorize to be made no more than once...".
So you are saying source code covered by the GPL has no value? Why else would you want it then?
Consideration doesn't have to be payment, "i.e." was wrong, it should be "e.g.".
Further, under the GPL, you do not give up any legal rights. You obtain additional rights; and that could be considered "consideration" all on its own. They are not as far-reaching as rights given under other licenses, but they are still in addition to those provided by law--without the GPL, you do not have the right to reproduce that copyrighted work.
Just a point of clarification, AIX is actually SVR3-based, not SVR4. And there's been a LOT done do it since AT&T handed the code over, as anyone who's programmed for or adminned AIX can tell you.
Like it or loathe it, you know it is 'different' from both SVR4 and BSD.
You realize most EULAs cover distribution as the major point to the license, right?
Of course they do, if you distribute it, you are not an END-USER, and therefore you are reading the wrong license agreement!
And they don't have to restrict distribution, because distributing the software without a distribution license would be a copyright violation anyway.
Just because something is in a license doesn't mean that it needs to be there.
There's a lot about software licensing that is completely unnecessary under post-1990s copyright law.
And have you ever run one of those brain-dead InstallShield installers for GPL software on Windows that says, "You MUST accept the license agreement to continue"? When one of the terms is, "You do not have to accept this license...."?
Just because a lot of people do something doesn't mean they're right.
Anybody that could afford the API document could implement. Years ago when I wanted to know what POSIX really said, I just couldn't afford a copy. And I couldn't justify it to get the company to spring for it either. So I got an O'Reilly book instead.
Mind you, this is exactly why a famous "POSIX work-alike" system is able to do exactly that. Get a copy of the spec, start coding.
Part of what made this so easy is that the shredder doesn't seperate the strands after shredding. They just kind of fall on the pile more or less in linear order.
I've heard that bi-directional shredders are better, I haven't gotten around to buying one yet.
One of the reasons I got a cross-cut shredder is because I had borrowed someone's linear shredder. It was rated for 3 sheets, which isn't a very strong shredder. But I noticed if I ran 2 or 3 sheets through it at once, it would often fail to cut through the paper--leaving a document nearly entirely intact. Sometimes there were just creases in the paper where the shredding wheels had bent the page, but didn't cut it.
When it did cut through, all the strips were side by side like you describe.
So far, the crosscut machine does a good job, I haven't seen it fail to cut anything that doesn't jam the motor first.
Then you can header-match in your MUA all you want--or not. (I run it all through procmail, but that's because I want all the filtering done before it hits my IMAP server.)
I can't check from here (no Mac), but look up "Mike Oldfield", "Tubular Bells". The original; it's two tracks, on each side of the record--Part 1 and Part 2. Both rather longer than 7 minutes.
So is that a $9.99 album, or a $1.98 album, or a not-available album?
Those features have also been licensed by Panasonic; I've got two S-VHS decks with them.
The default setting for CA is manual, but a quick trip to the menus sets it to automatically skip on playback.
It isn't 100% IME, but it is most likely to miss a commercial than to skip during the show.
Movie Advance has been useless. When I rent tapes, they're seldom Hollywood. So the thing scans for a bit, then winds all the way back to the beginning because there wasn't anything to skip.
Re:Saw this on Google News a while back
on
OSI vs SCO
·
· Score: 2, Interesting
AIX as a kernel is the Mach microkernel,
I know that AIX (along with every other OS at IBM) was to be hosted under Mach back in the mid-90s, but every system house except OS/2 abandoned that project. And OS/2... well, we know about what happened there. This was all part of the much-hyped and little-designed Workplace project.
AIX does have a rather unusual kernel for UNIX; it has quite a lot of dynamic configuration and device management, and it has had it for a very long time--back when it was still standard practice to reboot a Sun to make a config change, you almost never needed to reboot an RS/6000. AIX has, by Sun and Linux standards, an unusual shared object system as well. But it worked back when everyone was saying, "always build static on Solaris" (ca. 1993). That it hasn't changed much (prior to AIX 5.2) kind of hurts C++, but oh well.
And I can't find any trace of Mach on IBM's marketing pages for AIX; companies who use Mach seem to be quite happy to add it to the list of features they're marketing.
IBM also outright denies that its AIX flavor of Unix is a modification of SCO's licensed Unix.
This one is easy to prove; any non-AIX UNIX administrator will happily proclaim that AIX bears no relation to Solaris. "ODM? What's an ODM? Why doesn't this machine work the way I want! Arrrrgh!"
Slashdot Mirror
You'll like Mac OS X then:
/dev/sr device node and do an ioctl to access the generic SCSI interface, but only to that device. (And there's no reason why an ATAPI device couldn't give you the same thing; after all, ATAPI is basically SCSI over ATA.)
dev=IOCompactDiscServices
dev=IODVDServices
And start appending '/1', '/2' and so on if you have multiple burners.
Mac OS X has kernel support for sending generic SCSI commands to a particular device. For non-Mac OS X, imagine if you could open a regular
That would also mean you wouldn't have to have your CD burning program running as root and able to access every device on the bus, just because one of them is a CD burner.
Sun and Linux are notably different on this one. Linux isn't UNIX, and if you don't like it, return your downloaded ISO for a full refund. Solaris puts the POSIX shell in /usr/xpg4/bin/sh, which really sucks when you're trying to write a portable shell script. (No /bin/ksh by default on linux, /bin/sh isn't always POSIX, and only Solaris (IME) has the POSIX programs totally separate from the OS programs. Yippee!)
What Microsoft is lacking, aside from a "forbid first and allow later" approach to security, is sufficient analysis of their software's failure modes.
I'm sure whoever came up with that thing in the registry where you can cause a program to be run automatically before every .exe file is launched thought that it was very elegant and symmetrical. After all, that's how all non-.exe files are handled. In fact, they probably didn't even do anything special for .exe, it just fell out of the code that allows you to CreateProcess("foo.doc") or CreateProcess("movie.avi") and have the system magically load the right program first. (Kind of like #! on UNIX, but with control in the registry instead of in the file.)
Only problem is, CreateProcess("operating_system.exe") should really not be hijacked to run a virus before running the actual program.
Neither Mac OS Classic nor any of the UNIX-alikes (nor AmigaDOS, nor even MS-DOS actually) have anything like this. I'm pretty sure OS/2 doesn't either. It's unique to Windows, and it is the sort of thing that gives Symantec a large market for AV software.
How about "Windows Scripting Host"? I've been uninstalling that wherever I find it. (OK, 3 machines I have access to so far.) Haven't found anything that doesn't work without it. But you know what? It's been used for worms and viruses.
And Outlook's penchant for downloading things from e-mail messages. Apple's Mail.app has a setting for "do not download"--if the image is IN the message, it will display, if it has to be downloaded, no go. That download can be used to confirm an e-mail address--you all know this.
Then there's "active content" in e-mail. ActiveX and VBScript just expose too much of the system, through too few safeguards.
And how about the "always listening" ports on a client-only host? Why does a Windows 98 box on a home computer come up with any listening ports? They aren't needed unless your in a corporate LAN; the other desktop OS of the day had 0 (zero) listening ports unless the user turned on a server program.
Suffice to say, flaws in the RPC protocol which is connected to those always listening ports have been used to propagate malware recently. But can you even turn that software off? Try "net stop server"; you lose the Computer Browser too; and there's STILL an RPC port listening!
Yeesh.
Not all software requires a license.
Having recently had to switch to the truly abysmal Yahoo! Mac OS X client, I can't figure out which business model you could possibly be referring to.
The official Mac OS client, while being a good example of a bad UI:
If that's a business model, I have no sympathies for them. I do have friends on Yahoo!, so I'll be using something other than Yahoo!'s awful client to connect. (Hopefully Proteus 3, but Fire will do nicely if Proteus doesn't make me happy.)
RedHat and IBM both 'patch by subsystem', instead of updating the entire OS. So does Sun.
An IBM AIX 'system update' is just a bunch of previously-released patches all grouped together. Sun would call that a 'patch cluster'. Sometimes IBM will repackage the patches so they have a coordinated version number. Since RedHat's package versions are all over the place, there's no point for them.
Screw that, he wanted to assign TOAST to SCO and get it the hell over with.
ELF isn't even enough to specify an ABI. ELF simply gives you the linker and loader format.
You still have to deal with minor issues like:
- Which way does the stack grow?
- Which register is the stack pointer (not always dictated by the hardware, especially on RISC chips)
- Which register is used for globals? How is global data accessed? (TOC and GOT are two techniques; load-time address mapping is another one.)
- How are structure members laid out in memory? Padding and alignment requirements are influenced by the hardware, but that doesn't always mean the ABI is the most obvious interpretation of the hardware specs.
- What function arguments are in registers, which are on the stack? How are "ellipsis" functions handled? How are K&R argument promotions handled? How are aggregates passed? Are small aggregates (such as char[4]) are passed in a register, on the stack, or by pointer. Same with large floats, is a quad float passed by address or value?
- Setjmp/longjmp, how do they work?
- How does a stack frame look? If this isn't standard, exception handling can't unwind the stack, debuggers can't do a backtrace, and so on.
- Where's the heap? Register pointer, fixed segment, what?
While it is possible to have an ABI that is common across operating systems on the same CPU architecture, it is impossible to have the same ABI across CPU architectures. You just can't use R31 as a stack pointer on IA32; there isn't one. You can't use SPARC register windows on PowerPC. And so on.
I've used an OLD Apple USB keyboard on a Windows machine quite successfully. And on Linux, and on my PS2. The gotchas are already mentioned; with Meta and Alt being where they would be on any UNIX machine, and backwards from the Windows location.
The OTHER gotcha is that the old USB keyboards, Apple or otherwise, did not generate separate keysyms for left and right qualifiers. You could, therefore, not distinguish L_ALT from R_ALT, L_SHIFT from R_SHIFT, and so on. Sucks for games.
New keyboards distinguish left and right qualifiers. Other than that, the poster who said USB HIDs are USB HIDs has it; the protocol doesn't care where the buttons are.
My reading of the Act in question lead me to the important conclusion:
You may make copies of audio recordings you own (and give them away), as long as you receive nothing in return. No swapping, nothing. Just the goodness of your heart.
I couldn't find anything that would prevent you from making the copy on blank media given to you by someone else.
As for P2P, I suppose one could argue that the sending computer is copying the work as it sends it... but is a computer acting on behalf of its owner good enough for this? You wouldn't be able to make stuff you downloaded available for upload, unless you also owned a copy of the work in question.
The law says specifically audio recordings, but in Canada, we have a history of broadening laws to cover similar situations--such as copying DVDs. (The proposed tariff includes DVDs; it hasn't been approved--yet.) I would be interested in finding out what a judge would decide on a DVD movie copied following the same restrictions as an audio program. Especially since the prerecorded discs cost about the same.
That's just the physical interconnect. How about the actual bits sent over the wire?
Oh wait, that's called PostScript. And Adobe charged so much it never made it on the low end.
Plus, a lot of hardware makers hide problems with their hardware in the driver software. At least, I can't think of any other reason why they have so many different drivers for the same task; unless they're all just control freaks and won't work with NIH tech.
register variables may not have their address taken, which does give the compiler a hint that can improve optimization. It can also then diagnose a later programmer changing the routine to take the address of that variable; once the address is taken of a variable, several easy optimizations can't be done. (Most of which involve keeping the value in a... register.)
In my copy of 2.4.20, the first slide is arch/ia64/sn/io/ate_utils.c; and the start of the file, after the GPL, is "Copyright (C) 1992 - 1997, 2000-2002 Silicon Graphics, Inc. All righs reserved." The second slide appears to be Later That Same Function.
And if SCO is a Utah-based company, why are they programming in Greek?
So I suppose we'll see SGI getting sued next.
Also, someone else said that code is from BSD. Could it not also be from a "Getting Started with your New Itanium" kit from Intel?
Wait... Itanium... SCO... IBM... that was Montery, right? Wasn't that the project to make SCO, AIX, and Linux all more compatible, and to run on the Itanium?
Get rid of SCO. They're just wasting electricity.
That quote does not say that you may make only one additional copy. In fact, it implies that many archival copies may be made: "...and that all archival copies are destroyed...".
Although much of the text is singular, without an explicit clause saying something like, "may make or authorize to be made no more than once...".
Consideration doesn't have to be payment, "i.e." was wrong, it should be "e.g.".
Further, under the GPL, you do not give up any legal rights. You obtain additional rights; and that could be considered "consideration" all on its own. They are not as far-reaching as rights given under other licenses, but they are still in addition to those provided by law--without the GPL, you do not have the right to reproduce that copyrighted work.
If that was a "pre-Wave Table" SoundBlaster, it would generate noise through FM synthesis, would it not?
Just a point of clarification, AIX is actually SVR3-based, not SVR4. And there's been a LOT done do it since AT&T handed the code over, as anyone who's programmed for or adminned AIX can tell you.
Like it or loathe it, you know it is 'different' from both SVR4 and BSD.
Of course they do, if you distribute it, you are not an END-USER, and therefore you are reading the wrong license agreement!
And they don't have to restrict distribution, because distributing the software without a distribution license would be a copyright violation anyway.
Just because something is in a license doesn't mean that it needs to be there.
There's a lot about software licensing that is completely unnecessary under post-1990s copyright law.
And have you ever run one of those brain-dead InstallShield installers for GPL software on Windows that says, "You MUST accept the license agreement to continue"? When one of the terms is, "You do not have to accept this license...."?
Just because a lot of people do something doesn't mean they're right.
Anybody that could afford the API document could implement. Years ago when I wanted to know what POSIX really said, I just couldn't afford a copy. And I couldn't justify it to get the company to spring for it either. So I got an O'Reilly book instead.
Mind you, this is exactly why a famous "POSIX work-alike" system is able to do exactly that. Get a copy of the spec, start coding.
One of the reasons I got a cross-cut shredder is because I had borrowed someone's linear shredder. It was rated for 3 sheets, which isn't a very strong shredder. But I noticed if I ran 2 or 3 sheets through it at once, it would often fail to cut through the paper--leaving a document nearly entirely intact. Sometimes there were just creases in the paper where the shredding wheels had bent the page, but didn't cut it.
When it did cut through, all the strips were side by side like you describe.
So far, the crosscut machine does a good job, I haven't seen it fail to cut anything that doesn't jam the motor first.
I don't know spambayes, but bogofilter most definately can operating in a "ranking" mode:
Then you can header-match in your MUA all you want--or not. (I run it all through procmail, but that's because I want all the filtering done before it hits my IMAP server.)
I can't check from here (no Mac), but look up "Mike Oldfield", "Tubular Bells". The original; it's two tracks, on each side of the record--Part 1 and Part 2. Both rather longer than 7 minutes.
So is that a $9.99 album, or a $1.98 album, or a not-available album?
Those features have also been licensed by Panasonic; I've got two S-VHS decks with them.
The default setting for CA is manual, but a quick trip to the menus sets it to automatically skip on playback.
It isn't 100% IME, but it is most likely to miss a commercial than to skip during the show.
Movie Advance has been useless. When I rent tapes, they're seldom Hollywood. So the thing scans for a bit, then winds all the way back to the beginning because there wasn't anything to skip.
I know that AIX (along with every other OS at IBM) was to be hosted under Mach back in the mid-90s, but every system house except OS/2 abandoned that project. And OS/2... well, we know about what happened there. This was all part of the much-hyped and little-designed Workplace project.
AIX does have a rather unusual kernel for UNIX; it has quite a lot of dynamic configuration and device management, and it has had it for a very long time--back when it was still standard practice to reboot a Sun to make a config change, you almost never needed to reboot an RS/6000. AIX has, by Sun and Linux standards, an unusual shared object system as well. But it worked back when everyone was saying, "always build static on Solaris" (ca. 1993). That it hasn't changed much (prior to AIX 5.2) kind of hurts C++, but oh well.
And I can't find any trace of Mach on IBM's marketing pages for AIX; companies who use Mach seem to be quite happy to add it to the list of features they're marketing.
This one is easy to prove; any non-AIX UNIX administrator will happily proclaim that AIX bears no relation to Solaris. "ODM? What's an ODM? Why doesn't this machine work the way I want! Arrrrgh!"