Re:Already got a beta version....
on
Linux 3.0
·
· Score: 1
o in 2.5.31 Support insane number of processes (Linus Torvalds)
Just how insane would this be? Anyone know?
More than 65536. Hardly anyone needs more than 65536, but a few people with very large boxes are asking for it. Linus put the limit at something like 2^31 (~2x10^9) but that will most likely be reduced somewhat before 2.6 actually ships.
LWN has a good writeup explaining how this issue came up, though it was written before the issue was completely resolved.
What exactly do you mean by that? Surely not Uniform Modling Language.:P
User-Mode Linux. Running the Linux kernel as a regular userspace program in another copy of the Linux kernel (recurse as desired, not that you would want to). This way you can have many virtual Linux boxes housed in one physical box - just like VM for S/390, only different. Useful for hosting multiple customers on one application server and giving each customer "root" on their own virtual machine, that sort of thing. Also good for layered security even if you don't give out root.
Yes, the acronym is a bit unfortunate. Within the kernel itself, UML is referred to simply as um, just another architecture port alongside about 16 others: i386, ppc, etc.
As a device driver developer, do you think Linux's interal APIs change too frequently compared to other Unix operating systems? Would there be much value in creating a more stable, possibly cross-platform driver API for Linux? How much cross-platform driver code can be shared across Solaris, HP-UX, and Linux?
I'm not a driver developer, so I see the question wasn't aimed at me, but this has been done already - the infamous UDI (uniform driver interface) project, brainchild of Intel three or four years ago if I remember correctly. It was actually a uniform ABI, not just API. And it was endorsed by most of the commercial x86 Unix vendors. I guess they were hoping they could leverage Linux's comparatively broad driver support. (:
A few in the Linux kernel community, such as david parsons, supported UDI ("let's start harvesting Unixware drivers", I believe he said), but most of the others, including Linus, thought it was an idiotic waste of their time to support a third-party stable API (not to mention ABI) that would benefit all the other Unix vendors a great deal more than it would benefit Linux. Linus really doesn't like having his hands tied by in-kernel API compatibility. And, unlike most other OSes, Linux doesn't have much of a community of third-party driver developers who don't contribute directly to the master source, so Linus has the rare luxury of being able to change the API whenever a technical reason for doing so comes up. Pity the poor developers of AIX, Solaris or Windows, who have to hold on to any badly-designed APIs until the end of time so third-party video drivers (Windows) or those Oracle-specific kernel features (Unix) don't suddenly break.
(Note that the userspace API is different and is held near-sacred - nobody wants to break applications for no reason. Certain specialised apps like "ps" and "lsmod" can expect a rough ride, API-wise, but for the most part, you can still run 10-year-old Linux code.)
I think UDI eventually died the slow death of apathy - apparently without all those Linux drivers to leverage, there wasn't much point. But I could be wrong; perhaps Unixware and Solaris/x86 still support it.
LBX is basically dead, it didn't work very well when it was alive, and even with it X is still almost unusable over a slow
connection:(
I read an indirect reference to this somewhere, didn't know if it was fact or fiction. Anyway, the intent was to optimise bandwith usage. (: Thanks for the tip - now I have to go track down that post-mortem, which sounds interesting. (When TOG invented LBX I thought the theory sounded sound.)
As far as I know, the SHM extension is only used for pixmaps, other stuff like events still travel over a socket.
I'm sure you're right. Think about it, though - the biggest reason for the overhead of socket connections is their synchronous nature - requiring context switches. For events, you often can't really avoid that cost anyway, if you want a responsive UI..
Note however that
this is a UNIX domain socket, so is basically a direct line to another process via the kernel. It's far more efficient than network
sockets, I think people often get confused about this.
True. By "network socket" I really meant the "socket() bind() listen() accept() and associated stack" business. You're right that people get confused by this. I remember reading somewhere (how's that for a specific reference?) that unless your messages are quite large, sockets tend to be a faster IPC primitive than shared memory, in Linux at least. (For other Unices - who knows?)
XRender allows for hardware accelerated alpha-blending it is true, but I think the problem is that there isn't any way of getting a
pixmap of what is underneath a window, making it not enough for semi-transparent windows.
I bow to your knowledge of XRender, having basically none of my own except "that's how antialiased fonts are implemented".
Why can't we do the same with X? It's going to get harder and harder to grow with X, so lets lay some groundwork now for a window
system we can grow with for the next decade or more.
<eliza>Why do you think that it is going to get harder and harder to grow with X?</eliza>
(For that matter, why do you think X is bloated, or ugly, or slow, or obsolete, or inefficient?)
I suggest you check out the X extension mechanism. You know, the thing that makes possible such things as the RandR extension, which is what we're ostensibly talking about here. The original X11 design did not make any provisions for:
non-rectangular windows (see the SHAPE extension)
double buffering (see the DOUBLE-BUFFER extension)
3D hardware acceleration (see the GLX extension)
miscellaneous input devices (see the XInputExtension extension)
idle timeout events (see the MIT-SCREEN-SAVER extension)
optimising the protocol for low-bandwidth transports (see the LBX extension)
eliminating the network socket overhead entirely (see the MIT-SHM extension)
setting the output resolution/refresh dynamically (see the XFree86-VidModeExtension extension, and now the RandR extension)
transparency and alpha blending, including font antialiasing (see the RENDER extension)
video framebuffer acceleration (see the XVideo extension)
and much, much more. Yet these things all work fine today, without loss of any backward compatibility to older applications.
I agree with Jim Gettys on this one: people who say X is bloated and/or outdated are usually underinformed - meaning they have not really evaluated the alternatives or tried to design their own alternatives. Either that, or they are looking for something with a lot less functionality, like a simple framebuffer for a PDA.
(One thing I do think XF86 could still use is a widget set extension. I'm thinking individual toolkits like gtk2 or motif2.1, when installed on an X server, would register sub-extensions with it. The client-side libraries would use them if present, so most widget UI interaction would happen entirely on the server side. I think this would be very good for perceived UI latency. This is something I'd start investigating and hacking on if I had a lot more time than I do, but alas....)
Another important application is web design. A well-designed web site should look good at any resolution; however, since one can't
intuitively know how a page will look at every resolution, when designing a page it is convenient to change your resolution on the
fly and take a look at it.
What?!? You seriously think it's easier to change the video mode and then maximise your browser than it is just to resize your browser within its current mode?
And why would you only test widths of 800, 1024 and 1280? I almost never run a graphical browser window maximised, so my horizontal resolution (for the purpose of web design) is quite unpredictable. (For that matter, on a micro scale, the amount of border crap added by window managers and browsers varies considerably, and in many cases can vary within a single application.) To say nothing of font / font sizes... or are you one of those morons who think your exact font preferences are more important than mine?
(I won't even get into issues about layout for my preferred browsing environment, whose "resolution" is a very easy-on-the-eyes 144x72 text characters.)
Bottom line: your web page must lay out well along a 2D continuum of font sizes and horizontal widths. I don't see how the ability to switch between common screen resolutions helps much in testing this.
There is no mention anywhere of an ability to change colour depth.
Hmmm, I don't remember if this is in the article (I read it a couple weeks ago so I didn't reread it just now)... but the RandR effort does include color depth changing. I believe the old color depth is emulated if necessary, for applications that can't adjust on their own, but I don't remember if emulation is client-side or server-side.
I heard somewhere a theory that certain genetic diseases (they might have
been talking about autism; I wish I remembered) were not weeded out because it increased the survival of relatives of people who
had it.
Here's a related one. I once asked a biologist how it was possible for a genetic disposition for homosexuality to survive Darwinism. It seemed obvious to me that homosexuals would have fewer biological children than heterosexuals. He agreed with me, adding that according to computer models, even if homosexuals had 99% as many children as heterosexuals did, any hypothetical "gay gene" would be completely wiped out in a few thousand years.
But, he added, some biologist had recently published a clever theory: Apparently small tribes of prehistoric people would generally have a shaman, or medicine man. Furthermore, he said, for various reasons homosexuals were particularly well-suited to such a vocation. Therefore, a tribe whose gene pool was "just gay enough" had better survival odds than a "straight tribe", since its shamans would be more competent, and suitable shaman protegés easier to find.
In other words, although the shaman had no sons, he would have more nephews. And that could theoretically have been enough to tip the balance of natural selection.
(Note that I've been putting "gay gene" in quotes because I believe that, while there may exist congenital factors, homosexuality is a conclusion drawn from many non-congenital factors as well.)
In fact, I believe the etymology of "lame" as a description of disability actually stems from the other meanings, as opposed to vice-versa.
Go ahead and believe that, then. It doesn't appear to be true, according to Merriam-Webster. Their convention is to sort their definitions by age. And their first definition is indeed "disabled".)
But as long as you are arguing for what a word originally meant... then I would put forth that when I use the term gay to mean "stupid", I am actually just making fun of the Gay Nineties.[*] Can you prove or demonstrate otherwise?
By the way, posting as an AC is gay. (:
[*] For you cultural illiterates out there, the Gay Nineties refers to the last decade of the 1800s, when "gay" still meant "merry".
I do think it makes a lot more sense the way XHTML has it.
I, on the other hand, think this whole XML thing is a mistake. The whole point of computers is to automate work. The syntactical difference between SGML and XML (or between HTML and XHTML, if you'd rather frame it that way) is chiefly to make the document easier to parse at the expense of making it harder to hand-write / hand-edit.
Sure, I often use </p> tags - as I have done since 1995 or so - but that's because I'm an A-R pedant. Requiring them is silly. It just makes busy-work. Unless you think everyone should switch from vi to Dreamweaver just to edit a stupid HTML file.
There's a reason most people don't feel the need to add </li> tags to their list elements, or </td> to their table elements. (And these tags, unlike <p>, were always considered containers, as opposed to separators.)
SGML parsing was a solved problem. Sure it had its annoyances but it could be done. XML may be a step forward in terms of validation without a DTD, but for well-known, slow-moving standards like HTML, what's the point?
And I could easilly carry enough fuel
for several hundred hours of operation (of a notebook, say) in a small canister.
Oh, sure, that's convenient. And safe, I'm sure. You don't want little aerosol-size cans spewing high-pressure methane all over your laptop bag, eh?
All in all, it probably isn't going to be any more difficult to find a supply of fuel than it is to find a 110v 60hz outlet in
Nepal.
Requiring 110V 60Hz is soooo '80s. I had a Toshiba laptop (or was it merely "portable" back then?) in 1989 rated for 110-240V, 50-60Hz. Working with multiple power standards is a solved problem.
Anyway, in Nepal (or any other location without ready AC power) you are much more likely to be using a 12V lead-acid battery via the cigarette lighter plug in your truck. But that's beside my point:
Saying that availability of fuel cell refill tanks is not a problem because you can name at least one place where availability of grid power is a problem -- isn't exactly sound reasoning....
Please don't use "gay" to mean "stupid." It doesn't.
Didn't used to, you mean. As in, five years ago "gay" didn't mean "stupid". But then, twenty years ago, "gay" didn't mean "lesbian". Fifty years ago, "gay" didn't even mean "male homosexual".
But perhaps I mistake you. Perhaps your point is that "gay" can only mean "happy" or "lighthearted". In that case, while we disagree about the nature of language evolution, I take back my implication that your position is two-faced.
Well, believe it or not, I actually make the same mistake (FCC vs FAA) in conversation from time to time. I don't know why. I probably have less excuse than you, working as I do at an institute that gets a lot of FAA funding. (And no FCC funding.)
What a customer is allowed to do in terms of filtering is very different then what a 3rd party can do in a commercial setting.
Let me get this straight:
(a) Customer installs filtering proxy server which replaces banner ads with an unobtrusive GIF saying "Ad blocked by Junkbuster". Customer downloaded said proxy from some website.
(b) Customer installs software which filters amazon.com purchases to add referral credit for a specific third party (supplier of said software). This is not the primary user-visible purpose of the software, but during the installation, customer was asked whether he wants to support the third party in this manner, and he clicked [OK]. Due to either a bug or an intentional misfeature (hey, depends on whom you ask), software does not only add a redirect tag to independent purchases, but changes the redirect tag on purchases which would otherwise credit another referrer.
In both cases, the customer is running software obtained from a third party. In other words, it's running under local control - this isn't a virus or a remote-controlled DDOS agent or anything. Now, I think we agree that case (b) is a shady practice, not at all to be expected from a legitimate business whose primary interest is to help people commit copyright infringement, but how exactly are the two cases legally different?
Since Kazaa used the website without permission as part
of an intent to defaud that paragraph is illegal.
So you believe all filtering proxies are illegal? Else how do you define "used the website without permission" without including ad blockers, pr0n filters, or realtime download virus scanners?
Look, buddy -- I use P2P clients because I'm too damn cheap to buy a $15 dollar CD. What makes you think I'm going to spend
hundreds on virtual pc or vmware?
Heh, that's what User-Mode Linux is for. Of course, that does cost you that initial $0 to acquire Linux if you don't already have it. (:
So, to touch on your fear that "sharing may not be legal" and that "the RIAA could[n't] care less what the bands want," in this
case, the sharing is legal, and the RIAA can't do nuthin' about it. The RIAA does not have copyright on a live recording, simple as
that. The copyright for that show belongs solely to the band that played
Are you sure about this? What about whoever wrote the music the band is playing? I think in most countries you need a license from the copyright holder of the music in order to put on a public performance of it. This is what ASCAP and BMI are all about, right?
So unless the band you are tape-trading (a) writes all their own music and lyrics and (b) didn't sign away those copyright interests in their label contract, I don't think the situation is as simple as you say.
When you download the album, Best Buy still has it.
What they have is only worth about $1 in manufacturing costs, though, right? And another $2 or so in distribution costs? Isn't that what the pro-napster crowd loves to say? That CDs are massively overpriced?
So if all you care to think about is the actual cost of product to Best Buy, you must think it perfectly OK to shoplift an album from their store, leaving behind a $3 tip and saving yourself $9. After all, it's not like Best Buy was going to get $12 from you anyway, since you didn't think the album was worth that much, so really they haven't lost anything by your shoplifting, right?
Or how about, you buy an album from Best Buy, then set up shop in their parking lot selling CD-Rs with photocopied cover art for $3? Is that OK too? After all, you aren't depriving Best Buy of anything directly, right?
Slashdot Mirror
More than 65536. Hardly anyone needs more than 65536, but a few people with very large boxes are asking for it. Linus put the limit at something like 2^31 (~2x10^9) but that will most likely be reduced somewhat before 2.6 actually ships.
LWN has a good writeup explaining how this issue came up, though it was written before the issue was completely resolved.
User-Mode Linux. Running the Linux kernel as a regular userspace program in another copy of the Linux kernel (recurse as desired, not that you would want to). This way you can have many virtual Linux boxes housed in one physical box - just like VM for S/390, only different. Useful for hosting multiple customers on one application server and giving each customer "root" on their own virtual machine, that sort of thing. Also good for layered security even if you don't give out root.
Yes, the acronym is a bit unfortunate. Within the kernel itself, UML is referred to simply as um, just another architecture port alongside about 16 others: i386, ppc, etc.
We did, a few months before Windows 95 came out. Read the release announcement for kernel 1.2.0. Funny stuff.
I'm not a driver developer, so I see the question wasn't aimed at me, but this has been done already - the infamous UDI (uniform driver interface) project, brainchild of Intel three or four years ago if I remember correctly. It was actually a uniform ABI, not just API. And it was endorsed by most of the commercial x86 Unix vendors. I guess they were hoping they could leverage Linux's comparatively broad driver support. (:
A few in the Linux kernel community, such as david parsons, supported UDI ("let's start harvesting Unixware drivers", I believe he said), but most of the others, including Linus, thought it was an idiotic waste of their time to support a third-party stable API (not to mention ABI) that would benefit all the other Unix vendors a great deal more than it would benefit Linux. Linus really doesn't like having his hands tied by in-kernel API compatibility. And, unlike most other OSes, Linux doesn't have much of a community of third-party driver developers who don't contribute directly to the master source, so Linus has the rare luxury of being able to change the API whenever a technical reason for doing so comes up. Pity the poor developers of AIX, Solaris or Windows, who have to hold on to any badly-designed APIs until the end of time so third-party video drivers (Windows) or those Oracle-specific kernel features (Unix) don't suddenly break.
(Note that the userspace API is different and is held near-sacred - nobody wants to break applications for no reason. Certain specialised apps like "ps" and "lsmod" can expect a rough ride, API-wise, but for the most part, you can still run 10-year-old Linux code.)
I think UDI eventually died the slow death of apathy - apparently without all those Linux drivers to leverage, there wasn't much point. But I could be wrong; perhaps Unixware and Solaris/x86 still support it.
I read an indirect reference to this somewhere, didn't know if it was fact or fiction. Anyway, the intent was to optimise bandwith usage. (: Thanks for the tip - now I have to go track down that post-mortem, which sounds interesting. (When TOG invented LBX I thought the theory sounded sound.)
I'm sure you're right. Think about it, though - the biggest reason for the overhead of socket connections is their synchronous nature - requiring context switches. For events, you often can't really avoid that cost anyway, if you want a responsive UI..
True. By "network socket" I really meant the "socket() bind() listen() accept() and associated stack" business. You're right that people get confused by this. I remember reading somewhere (how's that for a specific reference?) that unless your messages are quite large, sockets tend to be a faster IPC primitive than shared memory, in Linux at least. (For other Unices - who knows?)
I bow to your knowledge of XRender, having basically none of my own except "that's how antialiased fonts are implemented".
<eliza>Why do you think that it is going to get harder and harder to grow with X?</eliza>
(For that matter, why do you think X is bloated, or ugly, or slow, or obsolete, or inefficient?)
I suggest you check out the X extension mechanism. You know, the thing that makes possible such things as the RandR extension, which is what we're ostensibly talking about here. The original X11 design did not make any provisions for:
and much, much more. Yet these things all work fine today, without loss of any backward compatibility to older applications.
I agree with Jim Gettys on this one: people who say X is bloated and/or outdated are usually underinformed - meaning they have not really evaluated the alternatives or tried to design their own alternatives. Either that, or they are looking for something with a lot less functionality, like a simple framebuffer for a PDA.
(One thing I do think XF86 could still use is a widget set extension. I'm thinking individual toolkits like gtk2 or motif2.1, when installed on an X server, would register sub-extensions with it. The client-side libraries would use them if present, so most widget UI interaction would happen entirely on the server side. I think this would be very good for perceived UI latency. This is something I'd start investigating and hacking on if I had a lot more time than I do, but alas....)
What?!? You seriously think it's easier to change the video mode and then maximise your browser than it is just to resize your browser within its current mode?
And why would you only test widths of 800, 1024 and 1280? I almost never run a graphical browser window maximised, so my horizontal resolution (for the purpose of web design) is quite unpredictable. (For that matter, on a micro scale, the amount of border crap added by window managers and browsers varies considerably, and in many cases can vary within a single application.) To say nothing of font / font sizes ... or are you one of those morons who think your exact font preferences are more important than mine?
(I won't even get into issues about layout for my preferred browsing environment, whose "resolution" is a very easy-on-the-eyes 144x72 text characters.)
Bottom line: your web page must lay out well along a 2D continuum of font sizes and horizontal widths. I don't see how the ability to switch between common screen resolutions helps much in testing this.
Hmmm, I don't remember if this is in the article (I read it a couple weeks ago so I didn't reread it just now) ... but the RandR effort does include color depth changing. I believe the old color depth is emulated if necessary, for applications that can't adjust on their own, but I don't remember if emulation is client-side or server-side.
In a cut-n-paste sort of way, you mean? How can something with 0% original content be considered insightful?
(You did read the article, right?)
Here's a related one. I once asked a biologist how it was possible for a genetic disposition for homosexuality to survive Darwinism. It seemed obvious to me that homosexuals would have fewer biological children than heterosexuals. He agreed with me, adding that according to computer models, even if homosexuals had 99% as many children as heterosexuals did, any hypothetical "gay gene" would be completely wiped out in a few thousand years.
But, he added, some biologist had recently published a clever theory: Apparently small tribes of prehistoric people would generally have a shaman, or medicine man. Furthermore, he said, for various reasons homosexuals were particularly well-suited to such a vocation. Therefore, a tribe whose gene pool was "just gay enough" had better survival odds than a "straight tribe", since its shamans would be more competent, and suitable shaman protegés easier to find.
In other words, although the shaman had no sons, he would have more nephews. And that could theoretically have been enough to tip the balance of natural selection.
(Note that I've been putting "gay gene" in quotes because I believe that, while there may exist congenital factors, homosexuality is a conclusion drawn from many non-congenital factors as well.)
Go ahead and believe that, then. It doesn't appear to be true, according to Merriam-Webster. Their convention is to sort their definitions by age. And their first definition is indeed "disabled".)
But as long as you are arguing for what a word originally meant ... then I would put forth that when I use the term gay to mean "stupid", I am actually just making fun of the Gay Nineties.[*] Can you prove or demonstrate otherwise?
By the way, posting as an AC is gay. (:
[*] For you cultural illiterates out there, the Gay Nineties refers to the last decade of the 1800s, when "gay" still meant "merry".
<reply content="So in other words your parser can't handle this sentence?" />
I, on the other hand, think this whole XML thing is a mistake. The whole point of computers is to automate work. The syntactical difference between SGML and XML (or between HTML and XHTML, if you'd rather frame it that way) is chiefly to make the document easier to parse at the expense of making it harder to hand-write / hand-edit.
Sure, I often use </p> tags - as I have done since 1995 or so - but that's because I'm an A-R pedant. Requiring them is silly. It just makes busy-work. Unless you think everyone should switch from vi to Dreamweaver just to edit a stupid HTML file.
There's a reason most people don't feel the need to add </li> tags to their list elements, or </td> to their table elements. (And these tags, unlike <p>, were always considered containers, as opposed to separators.)
SGML parsing was a solved problem. Sure it had its annoyances but it could be done. XML may be a step forward in terms of validation without a DTD, but for well-known, slow-moving standards like HTML, what's the point?
...Says the guy in a console fanboy flamefest. Thanks - that was the funniest thing I've read all day.
Oh, sure, that's convenient. And safe, I'm sure. You don't want little aerosol-size cans spewing high-pressure methane all over your laptop bag, eh?
Requiring 110V 60Hz is soooo '80s. I had a Toshiba laptop (or was it merely "portable" back then?) in 1989 rated for 110-240V, 50-60Hz. Working with multiple power standards is a solved problem.
Anyway, in Nepal (or any other location without ready AC power) you are much more likely to be using a 12V lead-acid battery via the cigarette lighter plug in your truck. But that's beside my point:
Saying that availability of fuel cell refill tanks is not a problem because you can name at least one place where availability of grid power is a problem -- isn't exactly sound reasoning....
Didn't used to, you mean. As in, five years ago "gay" didn't mean "stupid". But then, twenty years ago, "gay" didn't mean "lesbian". Fifty years ago, "gay" didn't even mean "male homosexual".
But perhaps I mistake you. Perhaps your point is that "gay" can only mean "happy" or "lighthearted". In that case, while we disagree about the nature of language evolution, I take back my implication that your position is two-faced.
Well, believe it or not, I actually make the same mistake (FCC vs FAA) in conversation from time to time. I don't know why. I probably have less excuse than you, working as I do at an institute that gets a lot of FAA funding. (And no FCC funding.)
What? General aviation is somehow influenced by the interests of the Federal Communications Commission?
And here I thought aviation would be more the business of the FAA.
Heh. Invoke the ADA to get yourself an office with a window. Employers are required to make reasonable accommodations for disabilities, y'know.
Hey, get a sense of humor already. (:
Let me get this straight:
(a) Customer installs filtering proxy server which replaces banner ads with an unobtrusive GIF saying "Ad blocked by Junkbuster". Customer downloaded said proxy from some website.
(b) Customer installs software which filters amazon.com purchases to add referral credit for a specific third party (supplier of said software). This is not the primary user-visible purpose of the software, but during the installation, customer was asked whether he wants to support the third party in this manner, and he clicked [OK]. Due to either a bug or an intentional misfeature (hey, depends on whom you ask), software does not only add a redirect tag to independent purchases, but changes the redirect tag on purchases which would otherwise credit another referrer.
In both cases, the customer is running software obtained from a third party. In other words, it's running under local control - this isn't a virus or a remote-controlled DDOS agent or anything. Now, I think we agree that case (b) is a shady practice, not at all to be expected from a legitimate business whose primary interest is to help people commit copyright infringement, but how exactly are the two cases legally different?
So you believe all filtering proxies are illegal? Else how do you define "used the website without permission" without including ad blockers, pr0n filters, or realtime download virus scanners?
Heh, that's what User-Mode Linux is for. Of course, that does cost you that initial $0 to acquire Linux if you don't already have it. (:
Are you sure about this? What about whoever wrote the music the band is playing? I think in most countries you need a license from the copyright holder of the music in order to put on a public performance of it. This is what ASCAP and BMI are all about, right?
So unless the band you are tape-trading (a) writes all their own music and lyrics and (b) didn't sign away those copyright interests in their label contract, I don't think the situation is as simple as you say.
What they have is only worth about $1 in manufacturing costs, though, right? And another $2 or so in distribution costs? Isn't that what the pro-napster crowd loves to say? That CDs are massively overpriced?
So if all you care to think about is the actual cost of product to Best Buy, you must think it perfectly OK to shoplift an album from their store, leaving behind a $3 tip and saving yourself $9. After all, it's not like Best Buy was going to get $12 from you anyway, since you didn't think the album was worth that much, so really they haven't lost anything by your shoplifting, right?
Or how about, you buy an album from Best Buy, then set up shop in their parking lot selling CD-Rs with photocopied cover art for $3? Is that OK too? After all, you aren't depriving Best Buy of anything directly, right?