having used both to varying degrees, you should use gnome(kde) if you have an aesthetic or functional preference of gnome(kde) over kde(gnome).
no other reason. (assuming, as you said, that you don't get into the whole license issue) both of them do more or less the same thing, although each has little strengths that the other lacks. for example, i prefer the gnoe panel to the kde panel, but i refuse to use the gnome filemanager. kfm on the otherhand make a good basic file/web browser when i don't need javascript or https. so i use the gnome panel with kfm. or sometimes i get sick of one or both of them and go back to using a plain vanilla windowmanager (if you can call enlightenment plain) until the next version comes out. so if youve tried both, and decide that one works better for you, stick to that one. but if you have the time, you should still check out new releases form "the other side" every now and then. i gave up on the gnome panel a long time ago for example, but now with their "tasty yellow banana" release, i decided to tried it out again and was very impressed.
for all you people whining about learning from RISC architecture, and complex hardware, realize something. this is two grad students and their advisor working on a project to get their masters degree. how many people are there out there with a masters degree in cs? how many of these peoples projects actually turned into somthing used by a major hardware company? probably not many. but their research can still turn up a lot of useful information. that's why it's research. for example, how many people have read linus' comments about the academic community and micro kernels? he basically slams on people who advocate micro kernel technology, but he turns around and says that most of the optimizations that can be used to make micro kernels faster can also be used to make monolithic kernels (ie linux) faster as well. so maybe you think that the study these guys are doing is useless, but what they learn in this project may turn out to be very useful information. and they may do other things that could be very useful. (for example, i may be doing some work with one of these guys on the implementations of threading and garbage collection in kaffe. maybe that's not exactly related, but im sure the studies he wants us to read on garbage collection came from his work on this project)
oh andas far as comments about their html or english, neither of those is particularly relevant to their research. i've had both of these guys as TA's for cs classes here at iit, and i had dr. chang as a teacher. all three of them are not native english speakers and have very heavy accents. but while they are difficult to understand, they are all extremely knowledgable about hardware and hardware design, and are some of the brightest people i have met in college.
actually, you've got it backwards. gtk+ license would be much more favorable to them than qt. gtk+ is LGPL. OTOH, in order to release a qt app without releasing the source, they would have to pay a licensing fee to troll-tech, since qt is only free if you release the source to your qt apps. anyway, motif is still the most popular toolkit for commercial unix apps (was last i checked anyway) and it wouldn't be all that unlikely for them to write their own, if they made its API similar enough to the windows gui API. that would minimize their porting, and put most of the work in developing the tk.
iirc, this is why the"export" keyword was added to the language. however, last i checked, no compilers (at least no compiler that i used or had been told about) supported this keyword.
I have heard a number of times on/. that the content can be copied bit by bit (which seems to make a lot of sense), I'd like to know if anyone has actually done this, if so then Valenti's arguments are completely bogus.
for the average home pc pirate who rips mp3s or uses a cdrom burner to copy his friends cds, this is impractical for two reasons. one, writeable dvds have a smaller capacity than the pressed dvds. two, writeable dvds supposedly have the area on the disc where the key is stored blanked out so nothing can be written there. so if i go out and buy a writeable dvd-ram drive for my pc, i will not be able to make bitwise copies. then again, i wouldn't be able to make copies by decrypting them either, because of the disc capacity issue. it is possible (likely?) that one or both of these restrictions will no longer be an issue in the future. in any case, the average "home pirate" can likely find numerous other ways to get perfect copies of a dvd that doesn't even deal with encryption. one possible trick that comes to mind would be using the digital output that most set-top dvd players have, or doing some low level reading of the data stream off of the bus, much like microsofts secure music media was cracked within an hour by reading data as it went to the soundcard.
on the other hand, for largescale pirates who make pirated copies of movies in the thousands or more, it probably would not be too huge of a stretch to come across the hardware for pressing thir own dvds, in which case bit for bit copies would be perfectly indistinguishable (to the player at least) from the original.
i haven't read all of the comments yet, so somebody may have pointed this out already. still, i think it's worth saying
So what constitutes fair use of a DVD in your eyes besides simply buying a DVD and using one of the MPAA's authorized players? Any use by which you buy it at a price.
so, if i install linux on my computer that has a dvd-rom drive, which i paid for, and put one of my dozen or so dvd's, which i also paid for, into my dvd-rom drive on my linux box, then by his own words, i am perfectly within my rights to use DeCSS to view those movies on my linux box. whether this is practical given the current state of video support in linux and whatever else, i do not know or do not care. the point is that he has blatently said that this is allowed.
Anybody know if vanilla Netscape is/was skinnable? (i.e., no Mozilla)
if you're talking about windows, you really can't do much to it outside of changing the colors ala windows themes. in which case, of course, that doesn't really count, because you're changing all of your winddows, not just netscape.
in linux, you can do a lot of funky stuff to your netscape windows via motif and Xresources. it's not easy, i've tried it, but i have seen other people who have done amazingly cool stuff to their netscape windows with this. run editres, and import the widget tree from your running copy of netscape navigator/communicator. it should give you a whole list of widgets, and you can change the color, shodow, border width, text font, maybe set a background image...
anyway, it takes a lot of patience, and your average user probably wouldn't consider it too worthwhile. the other problem is that you can't really pack up your customizations as an easily installable "skin". you have to manually add the changes to your.Xdefaults file.
> If you decide the 20th Century ends at the end of 1999, then which past century loses the year?
it would obviously have to be this century, because the end of the 19th century was celebrated on dec 31st of the year 1900. 1901 was therefore officially recognized as the beginning of the 20th century. of course now that we've decided that the 20th century gets the shaft, we have to pick which decade of the 20th century only has 9 years....
well, im not going to go into it too deep, but at least in reference to this situation:
obscurity in this case is giving the client the positions of all of the players in the game, and trusting the client to only display to the player the other players that can be seen from his current viewpoint, leaving the rest of the players obscured. minimum disclosure would be only telling the client where are the other players that the player can see from his current viewpoint. this way, the client/player can't cheat because the cient doesn't know any more about the positions of the other players than the current player is supposed to.
in general, obscurity means leaving extra data that could be used to cheat hidden somewhere in the protocol. the programmers are trusting that they are hiding the information well enough that the players won't be able to find it. of course once the source is released, this "hiding" is no longer good enough, because cheaters can just look at the source and see where it is hidden. minimum disclosure means only giving out as much information as anyone needs to know. there is no extra information hidden in the protocol for cheaters to find or use to their advantage.
like lynx, except it supports tables, frames (it renders them as tables), cookies, https, and something else that i am forgetting. (note cookies and ssl can be disabled if you don't want them)
it takes a little longer to render pages, because it does two pass rendering, meaning that it has to download the entire page before it renders it, but i generally dont find that very annoying.
ati cards have fairly decent hardware 3d in windows. however, until very recently, they refused to release any of the 3d specs to their cards to open source driver authors. as such, it will be a while before you are able to get any kind of hardware 3d out of your ati card in linux. you will be able to compile mesa and run it just fine, but you won't get any hardware acceleration. in windows, it would probably do ok, although i never used mine for 3d much. when i traded in my ati 3d rage pro, they had just released gl drivers that allowed you to play glquake at an almost reasonable frame rate.
they are nominees for this years award, but the award is not for what they've done this year. it's kind of a lifetime achievement award. linus is not listed because he already one and so is not eligible to recieve it again.
he actually says that in the insterview: It might make more sense to have a family of algorithms, each tuned to a particular application,
he continues: but there will be only one AES. And when AES becomes a standard, customers will want their encryption products to be "buzzword compliant." They'll demand it in hardware, in desktop computer software, on smart cards, in electronic-commerce terminals, and other places we never thought it would be used.
so, no, we don't need one algorithm to do everything. but there can only be one standard, and when there is, everyone will want to use that standard, even if it's not the best tool for the job. so they want to choose the standard that is as close as possible to being the best tool for all of these jobs.
well, i believe his point was that, since lib-apt was licensed under the gpl, there is something you could not do (link against it wit qt code) that you would be able to do if lib-apt was released under some other licence, such as bsd or lgpl. meaning that releasing lib-apt under the gpl makes it lessfree than if it were released under one of the other licences.
he's very right about the advantage of open source being the right to modify. but his opening examples don't go along with that argument.
No company now would commit to a closed hardware strategy. It would cost them more than using commodity components first of all, what does he mean by "closed hardware" strategy? sun sparcs? ibm mainframes? i know there are plenty of companies that have committed themselves to both. and there still will be for some time. besides, who really has the "right to modify" when it comes to hardware? i don't. at least not with any hardware i know of. which actually means there are no "open hardware" strategies. at least not in the sense that he talks about with software. with hardware, the only choice you have is to go with something else if your hardware vendor isn't meeting your demands.
well, surprise! you also have that option with software. if sybase isn't living up to your expectations, you can switch to oracle. or vice versa. if you have a beef with solaris for your oracle servers, use irix. or nt. or linux. or.... oracle will run on any of these.
while the right to modify is a big advantage for open ource, it is not the only criteria for choosing software. oracle has a huge market right now for databases, and no one has the right to modify it. and as long as they continue to offer other advantages that outweigh the right to modifiy, they will continue to have a huge market share. if people decide that mySQL does everything that they need, and the right to modify is important to them, they will use it. but there are several commercial databases that have huge advantages in both performance and power over mySQL that, for most companies, greatly outweigh the disadvantage of not being able to modify them to fix their problems.
Re:It's not Pro-NT, it's a Linux Bug Report. :)
on
John Carmack Answers
·
· Score: 1
> Just remember Redmond, we'll always be one step behind!
Is it just me, or does that not sound like something we should really be bragging about?
was probably the guy we just interviewed (and the rest of id software)
for at least a year or two, the biggest criticisms i heard about os/2 was "it doesn't run doom!!!" while this probably wasn't an issue for commercial settings, i imagine that hurt it as much as anything else as an alternative on the home desktop....
the downside to having to maintain that many packages, of ocurse, is that in the ~2 years that have have been using linux, there have only been two versions of debian released, while this is redhat's 5th (not counting 4.2 which was only a few months old when i first used linux)
also, the one time that i did decide to install debian, i gave up after having to sort through that list of all several thousand packages to choose which ones i wanted and which ones i didn't, and then have it prompt me for configuration info for a quarter of them as it installed them. the system administrator where i used to work, who has probably installed the same version of debian at least 20 or thirty times, once told me that he has gotten to the point where he can get through that whole process in about a half hour on a fast enough computer. i can take a version of redhat i've never seen before, and go through all of it's installation questions in under 15 minutes (much less if i choose their predefined groups rather than individual packages), and then let it sit for anywhere from 15 minutes to several hours (depending on whether i am doing an install from hard disk, cdrom, or ftp) while it installs all of those packages without any intervention from me.
most browsers do this, but not all. i worked for an e-commerce company where we had to disable some of our tracking stuff that was based on http-referer because a few widespread browsers don't implement it. MSIE for the mac was the most notable one, but i believe there were one or two other (reasonably) popular browsers that do the same thing.
and since dave taylor works there, maybe they're building this thing out of pinky processors....
does anybody have a copy of the pinky processor spec that he had posted on his page at crack dot com before they went bankrupt? i had it bookmarked, but it went away, and i have never been able to find another copy of it....
being primarily a c programmer, i was initially dissappointed that this book seemed to focus almost entirely on mod_perl. (while i like perl for quick and dirty wirk, it never fails to infuriate me when i try to use it for anything significant) however, after reading through the book, i still found it to be extremely useful. the middle chapters give a very good explanation of how the apache api can be used to do what you want. Basically, i skimmed through the rest of the book to get the basic concepts down, and since then i have lived in chapters 10 and 11, the c api reference. even without the rest of the book to explain how to use it, these two chapters are by far the most useful reference for the apache api that i have found.
according to the website for the voodoo3 Xserver: Why does it use so much memory? Ps and top report the amount of memory being used by the number of pages mapped to your process. Because the Banshee has 16MB of framebuffer and 32MB of control registers the process size will be at least 48MB. Of course, this is using almost no RAM in your system since it is real memory or registers on the board.
i don't know if this is generally true or not, but if the numbers you have for memory usage by your x server seem unreasonably high, this may be the reason...
I'm missing your point, you could have just as easily s/OS/compiler/g in that comment. If I'm not going to pay a few hundred for a OS, and the OS I get come packaged with a compiler that works (well enough...) I'm going to go get a $500 compiler? Not likely.
His point was that in a setting where you have say 20 computers working on big projects, a person or organization could spend $500 on on one copy of the compiler, and distribute the generated executables accross all of the machines, as oppsed to buying 20 copies of the OS. another issue is that if speed is critical for you applications, but you are still on a low budget, a $500 dollar compiler may seem like a very good deal as far as price/performance, but a several tousand dollar OS probably will not, so it is not neccessarily true that someone who is willing to spend the money for extra speed will go buy their OS rather than just the compiler
Slashdot Mirror
having used both to varying degrees, you should use gnome(kde) if you have an aesthetic or functional preference of gnome(kde) over kde(gnome).
no other reason. (assuming, as you said, that you don't get into the whole license issue) both of them do more or less the same thing, although each has little strengths that the other lacks. for example, i prefer the gnoe panel to the kde panel, but i refuse to use the gnome filemanager. kfm on the otherhand make a good basic file/web browser when i don't need javascript or https. so i use the gnome panel with kfm. or sometimes i get sick of one or both of them and go back to using a plain vanilla windowmanager (if you can call enlightenment plain) until the next version comes out. so if youve tried both, and decide that one works better for you, stick to that one. but if you have the time, you should still check out new releases form "the other side" every now and then. i gave up on the gnome panel a long time ago for example, but now with their "tasty yellow banana" release, i decided to tried it out again and was very impressed.
for all you people whining about learning from RISC architecture, and complex hardware, realize something. this is two grad students and their advisor working on a project to get their masters degree. how many people are there out there with a masters degree in cs? how many of these peoples projects actually turned into somthing used by a major hardware company? probably not many. but their research can still turn up a lot of useful information. that's why it's research. for example, how many people have read linus' comments about the academic community and micro kernels? he basically slams on people who advocate micro kernel technology, but he turns around and says that most of the optimizations that can be used to make micro kernels faster can also be used to make monolithic kernels (ie linux) faster as well. so maybe you think that the study these guys are doing is useless, but what they learn in this project may turn out to be very useful information. and they may do other things that could be very useful. (for example, i may be doing some work with one of these guys on the implementations of threading and garbage collection in kaffe. maybe that's not exactly related, but im sure the studies he wants us to read on garbage collection came from his work on this project)
oh andas far as comments about their html or english, neither of those is particularly relevant to their research. i've had both of these guys as TA's for cs classes here at iit, and i had dr. chang as a teacher. all three of them are not native english speakers and have very heavy accents. but while they are difficult to understand, they are all extremely knowledgable about hardware and hardware design, and are some of the brightest people i have met in college.
actually, you've got it backwards. gtk+ license would be much more favorable to them than qt. gtk+ is LGPL. OTOH, in order to release a qt app without releasing the source, they would have to pay a licensing fee to troll-tech, since qt is only free if you release the source to your qt apps. anyway, motif is still the most popular toolkit for commercial unix apps (was last i checked anyway) and it wouldn't be all that unlikely for them to write their own, if they made its API similar enough to the windows gui API. that would minimize their porting, and put most of the work in developing the tk.
iirc, this is why the"export" keyword was added to the language. however, last i checked, no compilers (at least no compiler that i used or had been told about) supported this keyword.
I have heard a number of times on /. that the content can be copied bit by bit (which seems to make a lot of sense), I'd like to know if anyone has actually done this, if so then Valenti's arguments are completely bogus.
for the average home pc pirate who rips mp3s or uses a cdrom burner to copy his friends cds, this is impractical for two reasons. one, writeable dvds have a smaller capacity than the pressed dvds. two, writeable dvds supposedly have the area on the disc where the key is stored blanked out so nothing can be written there. so if i go out and buy a writeable dvd-ram drive for my pc, i will not be able to make bitwise copies. then again, i wouldn't be able to make copies by decrypting them either, because of the disc capacity issue. it is possible (likely?) that one or both of these restrictions will no longer be an issue in the future. in any case, the average "home pirate" can likely find numerous other ways to get perfect copies of a dvd that doesn't even deal with encryption. one possible trick that comes to mind would be using the digital output that most set-top dvd players have, or doing some low level reading of the data stream off of the bus, much like microsofts secure music media was cracked within an hour by reading data as it went to the soundcard.
on the other hand, for largescale pirates who make pirated copies of movies in the thousands or more, it probably would not be too huge of a stretch to come across the hardware for pressing thir own dvds, in which case bit for bit copies would be perfectly indistinguishable (to the player at least) from the original.
i haven't read all of the comments yet, so somebody may have pointed this out already.
still, i think it's worth saying
So what constitutes fair use of a DVD in your eyes besides simply buying a DVD and using one of the MPAA's authorized players?
Any use by which you buy it at a price.
so, if i install linux on my computer that has a dvd-rom drive, which i paid for, and put one of my dozen or so dvd's, which i also paid for, into my dvd-rom drive on my linux box, then by his own words, i am perfectly within my rights to use DeCSS to view those movies on my linux box.
whether this is practical given the current state of video support in linux and whatever else, i do not know or do not care. the point is that he has blatently said that this is allowed.
Anybody know if vanilla Netscape is/was skinnable? (i.e., no Mozilla)
.Xdefaults file.
if you're talking about windows, you really can't do much to it outside of changing the colors ala windows themes. in which case, of course, that doesn't really count, because you're changing all of your winddows, not just netscape.
in linux, you can do a lot of funky stuff to your netscape windows via motif and Xresources. it's not easy, i've tried it, but i have seen other people who have done amazingly cool stuff to their netscape windows with this. run editres, and import the widget tree from your running copy of netscape navigator/communicator. it should give you a whole list of widgets, and you can change the color, shodow, border width, text font, maybe set a background image...
anyway, it takes a lot of patience, and your average user probably wouldn't consider it too worthwhile. the other problem is that you can't really pack up your customizations as an easily installable "skin". you have to manually add the changes to your
> If you decide the 20th Century ends at the end of 1999, then which past century loses the year?
it would obviously have to be this century, because the end of the 19th century was celebrated on dec 31st of the year 1900. 1901 was therefore officially recognized as the beginning of the 20th century. of course now that we've decided that the 20th century gets the shaft, we have to pick which decade of the 20th century only has 9 years....
> and if the official version isn't available - there's surely an opensource reader....
why, there certainly is. GPL i believe... it's called (strangely enough) "xpdf"! now who would've guessed that?
anyway, you're right, pdf is probably as close as you can get to a true cross platform document format.
well, im not going to go into it too deep, but at least in reference to this situation:
obscurity in this case is giving the client the positions of all of the players in the game, and trusting the client to only display to the player the other players that can be seen from his current viewpoint, leaving the rest of the players obscured.
minimum disclosure would be only telling the client where are the other players that the player can see from his current viewpoint. this way, the client/player can't cheat because the cient doesn't know any more about the positions of the other players than the current player is supposed to.
in general, obscurity means leaving extra data that could be used to cheat hidden somewhere in the protocol. the programmers are trusting that they are hiding the information well enough that the players won't be able to find it. of course once the source is released, this "hiding" is no longer good enough, because cheaters can just look at the source and see where it is hidden.
minimum disclosure means only giving out as much information as anyone needs to know. there is no extra information hidden in the protocol for cheaters to find or use to their advantage.
check out w3m (search on freshmeat)
like lynx, except it supports tables, frames (it renders them as tables), cookies, https, and something else that i am forgetting. (note cookies and ssl can be disabled if you don't want them)
it takes a little longer to render pages, because it does two pass rendering, meaning that it has to download the entire page before it renders it, but i generally dont find that very annoying.
ati cards have fairly decent hardware 3d in windows. however, until very recently, they refused to release any of the 3d specs to their cards to open source driver authors. as such, it will be a while before you are able to get any kind of hardware 3d out of your ati card in linux. you will be able to compile mesa and run it just fine, but you won't get any hardware acceleration. in windows, it would probably do ok, although i never used mine for 3d much. when i traded in my ati 3d rage pro, they had just released gl drivers that allowed you to play glquake at an almost reasonable frame rate.
they are nominees for this years award, but the award is not for what they've done this year. it's kind of a lifetime achievement award. linus is not listed because he already one and so is not eligible to recieve it again.
he actually says that in the insterview:
It might make more sense to have a family of algorithms, each tuned to a particular application,
he continues:
but there will be only one AES. And when AES becomes a standard, customers will want their encryption products to be "buzzword compliant." They'll demand it in hardware, in desktop computer software, on smart cards, in electronic-commerce terminals, and other places we never thought it would be used.
so, no, we don't need one algorithm to do everything. but there can only be one standard, and when there is, everyone will want to use that standard, even if it's not the best tool for the job. so they want to choose the standard that is as close as possible to being the best tool for all of these jobs.
well, i believe his point was that, since lib-apt was licensed under the gpl, there is something you could not do (link against it wit qt code) that you would be able to do if lib-apt was released under some other licence, such as bsd or lgpl. meaning that releasing lib-apt under the gpl makes it lessfree than if it were released under one of the other licences.
he's very right about the advantage of open source being the right to modify. but his opening examples don't go along with that argument.
No company now would commit to a closed hardware strategy. It would cost them more than using commodity components
first of all, what does he mean by "closed hardware" strategy? sun sparcs? ibm mainframes? i know there are plenty of companies that have committed themselves to both. and there still will be for some time. besides, who really has the "right to modify" when it comes to hardware? i don't. at least not with any hardware i know of. which actually means there are no "open hardware" strategies. at least not in the sense that he talks about with software. with hardware, the only choice you have is to go with something else if your hardware vendor isn't meeting your demands.
well, surprise! you also have that option with software. if sybase isn't living up to your expectations, you can switch to oracle. or vice versa. if you have a beef with solaris for your oracle servers, use irix. or nt. or linux. or.... oracle will run on any of these.
while the right to modify is a big advantage for open ource, it is not the only criteria for choosing software. oracle has a huge market right now for databases, and no one has the right to modify it. and as long as they continue to offer other advantages that outweigh the right to modifiy, they will continue to have a huge market share. if people decide that mySQL does everything that they need, and the right to modify is important to them, they will use it. but there are several commercial databases that have huge advantages in both performance and power over mySQL that, for most companies, greatly outweigh the disadvantage of not being able to modify them to fix their problems.
> Just remember Redmond, we'll always be one step behind!
Is it just me, or does that not sound like something we should really be bragging about?
was probably the guy we just interviewed (and the rest of id software)
for at least a year or two, the biggest criticisms i heard about os/2 was "it doesn't run doom!!!"
while this probably wasn't an issue for commercial settings, i imagine that hurt it as much as anything else as an alternative on the home desktop....
the downside to having to maintain that many packages, of ocurse, is that in the ~2 years that have have been using linux, there have only been two versions of debian released, while this is redhat's 5th (not counting 4.2 which was only a few months old when i first used linux)
also, the one time that i did decide to install debian, i gave up after having to sort through that list of all several thousand packages to choose which ones i wanted and which ones i didn't, and then have it prompt me for configuration info for a quarter of them as it installed them. the system administrator where i used to work, who has probably installed the same version of debian at least 20 or thirty times, once told me that he has gotten to the point where he can get through that whole process in about a half hour on a fast enough computer. i can take a version of redhat i've never seen before, and go through all of it's installation questions in under 15 minutes (much less if i choose their predefined groups rather than individual packages), and then let it sit for anywhere from 15 minutes to several hours (depending on whether i am doing an install from hard disk, cdrom, or ftp) while it installs all of those packages without any intervention from me.
most browsers do this, but not all. i worked for an e-commerce company where we had to disable some of our tracking stuff that was based on http-referer because a few widespread browsers don't implement it. MSIE for the mac was the most notable one, but i believe there were one or two other (reasonably) popular browsers that do the same thing.
and since dave taylor works there, maybe they're building this thing out of pinky processors....
does anybody have a copy of the pinky processor spec that he had posted on his page at crack dot com before they went bankrupt? i had it bookmarked, but it went away, and i have never been able to find another copy of it....
yes, but what about the companies that make the bottles? you can't just change a label on a gallon jug and suddenly it's a 2 liter jug.
unless you change the label to say 3.xxxx liters......
being primarily a c programmer, i was initially dissappointed that this book seemed to focus almost entirely on mod_perl. (while i like perl for quick and dirty wirk, it never fails to infuriate me when i try to use it for anything significant) however, after reading through the book, i still found it to be extremely useful. the middle chapters give a very good explanation of how the apache api can be used to do what you want. Basically, i skimmed through the rest of the book to get the basic concepts down, and since then i have lived in chapters 10 and 11, the c api reference. even without the rest of the book to explain how to use it, these two chapters are by far the most useful reference for the apache api that i have found.
according to the website for the voodoo3 Xserver:
Why does it use so much memory? Ps and top report the amount of memory being used by the number of pages mapped to your process. Because the Banshee has 16MB of framebuffer and 32MB of control registers the process size will be at least 48MB. Of course, this is using almost no RAM in your system since it is real memory or registers on the board.
i don't know if this is generally true or not, but if the numbers you have for memory usage by your x server seem unreasonably high, this may be the reason...
I'm missing your point, you could have just as easily s/OS/compiler/g in that comment. If I'm not going to pay a few hundred for a OS, and the OS I get come packaged with a compiler that works (well enough...) I'm going to go get a $500 compiler? Not likely.
His point was that in a setting where you have say 20 computers working on big projects, a person or organization could spend $500 on on one copy of the compiler, and distribute the generated executables accross all of the machines, as oppsed to buying 20 copies of the OS.
another issue is that if speed is critical for you applications, but you are still on a low budget, a $500 dollar compiler may seem like a very good deal as far as price/performance, but a several tousand dollar OS probably will not, so it is not neccessarily true that someone who is willing to spend the money for extra speed will go buy their OS rather than just the compiler