You will not see consumer systems with more then 4GB any time soon
The point about mmap(2) is to let the system (the VM subsystem of the kernel) manage the caching for the userland processes using it, avoiding extra copies to/from buffers in userland and eliminating in several cases the need for custom caching code (processes don't have to worry about data being available in RAM: the kernel automatically takes care of that when needed).
You don't need gobs of memory to do this, but in order to work on large amount of data you need a large address space, which is what 64 bit architectures provide. Of course, the more physical memory you have, the less the kernel has to swap pages in and out, but the main point is not that.
A little example to clarify: in order to keep things simple (instead of needing two intermixed caching systems, one for the VM and one for disk accesses), the Hurd just mmaps the whole partition. This means that the maximum size of a partition has an upper limit given by the size of the addressing space, which is 4GB on 32bit architectures (actually less, since in that address space you have to keep also the code that uses the mmapped data, so it's more like 2GB/3GB). A 64bit architecture comes very handy here, given the current size of hard disks.
I ripped is Pink Floyd's Dark Side of the Moon, Capitol's catalog # CDP 7 46001 2, DIDX 226.
There's something strange going on here: I repeated your test just for the sake of it using a CD that according to the codes you gave should be the very same, and for the same version of cdparanoia.
# cdparanoia 1 pf2.wav cdparanoia III release 9.8 (March 23, 2001) (C) 2001 Monty <monty@xiph.org> and Xiphophorus
Report bugs to paranoia@xiph.org http://www.xiph.org/paranoia/
Ripping from sector 0 (track 1 [0:00.00]) to sector 17834 (track 1 [3:57.59]) ...
Please note that the ending sector is not the same as yours (tried also with option -T with no effect). As I said before, the CD bears the same codes as the ones you specified, and is clearly labelled "Made in USA" (I live in Europe). I also tried on another CD-reader, and the results are the same (ending sector is 17834). I'd expected that at least the TOC entries would be the same...
For further checking, here there is the output from cd-discid, if you want to check with yours:
For a big project like a book, the programmability of direct LaTeX coding is actually a big advantage. You can define your own macros, and then change them globally if you feel like it later.
Well, LyX allows you to use inline TeX/LaTeX if you need to.
The last time I tried Sketch, it was extremely lame and buggy
You have to use the stable release, 0.6.x (the Tkinter-based one), possibly not the versions pre-packaged by RedHat (which had serious problems, probably due to mistakes in the packaging of the Python Imaging Library). Never experienced problems with SuSE or Debian pre-packaged versions.
Sodipodi OTOH is coming out really well and has improved a lot lately, but it still misses some basic features like layers, styles, the ability to combine shapes while linearly transforming them, the ability to export as EPS (you need Batik or Sketch) and more advanced ones (scripting) which have been in Sketch for years, so I can't really recommend it for now.
I agree that proprietary tools are a better choice as of now, but using Illustrator just to retouch computer-generated diagrams is almost surely overkill.
Save yourself the burden of typing LaTeX directly, and get a good front-end like LyX. Btw, it can also import existing (not-too-convoluted) LaTeX, just in case you already started writing. TeXmacs could be another option if your book is on a mathematical subject.
For vectorial diagrams and images, get Sketch and Dia and forget everything else (except perhaps Xfig, which comes handy sometimes). Sketch does a decent job at importing simple PostScript by itself (so you can retouch it), and of course it exports PS and EPS. For importing complex PostScript you may also use it together with pstoedit, which supports the Sketch format natively.
For graphs and trees have a look at Graphviz, which can generate beautiful outputs (both EPS or bitmapped) from simple textual descriptions of nodes and arcs (and it saves you
From the very same study (see paragraph "the burden of expertise"):
It thus appears that extremely competent individuals suffer a burden as well. Although they perform competently, they fail to realize that their proficiency is not necessarily shared by their peers.
So it seems there could be two problems here, not just one.
The best solution I can think of is to print out two paper receipts for each vote, one to go to the election commission (for manual recounts) and one to go to the voter.
Ok for the receipt to the commission, but I'm not completely sure about the receipt to the voter: let's say that some days before the elections someone comes to you telling how you should vote, "or else". And he requires that after the elections, you show him a proof that you actually voted as you were told.
This went so far in some areas of Italy that on the last (regional) elections the usage of photocameras and videophones were explicitly forbidden in the voting booth. And yes, someone actually tried anyways and was discovered (and his vote invalidated).
So, in some way, being unable to prove to someone else how you voted is not entirely a bad idea.
(of course it can be objected that the nasty guys could come after you anyways if the result of the elections is not the expected one, regardless of how you actually voted...).
Under the GPL they would have to release any changes they make to the kernel back into the public domain.
If and only if they distribute such derived works. There is no obligation in the GPL to redistribute derived works (there can't be, being the GPL a permission to distribute copyrighted material at certain conditions, permission which you normally don't have).
Of course, if they don't redistribute the modified sources, they also get the burden of mantaining them. That sounds fair to me.
You know when you download the kernel and you see those files that have the md5sum's and they are signed.... ever notice when you unpack the kernel you don't get Linus's private key, and you don't get GPG, nor md5.
Signed and md5summed tarballs are fine, they don't keep me from getting the kernel sources, modify them, recompile and install.
OTOH, apparently I need some non-free tools and secret data (which I'm assuming that are not part of the normal distribution) in order to install a modified version of this gnuboy port.
Now, it can be argued that I don't really need them, it's just that without them the installation becomes higly inconvenient. But if it is so, why require the installation scripts at all in the source?
...For an executable work, complete source code means all the source code for all modules it contains, plus any associated interface definition files, plus the scripts used to control compilation
and installation of the executable...
Just wondering if the signing tool could fall in the highlighted category.
OTOH, as I was forecasting in a comment of mine of some time ago (sorry, can't find it), I'm already seeing websites where if you follow a link, you are presented with an intermediate adv. page (not a popup, but a full page), which then forwards you to the real link target in 10 seconds or so. Just like ad pages in the middle of magazine articles.
Your suggestion of an amoral law-abiding social norm actually does exist already
Perhaphs you misunderstood me, or I'm misunderstanding you: mine is not a suggestion, but I was stating a fact: people choose to obey law because it serves best their long-term interests, and not because they fear the police.
If in order to punish those who were not obeying the law, also the ones obeying the law are punished... well, it's like saying "you fool!" to those who obeyed the law in the first place, and this is the thing I can't tolerate in the first place.
What really scares me is the bottomline message (from this and other recent measures): people obeying the law always end in paying also for people not obeying the law. In other words: if you obey the law, you are a (paying) moron.
Morals aside, someone please tell me a reason why one should choose to obey the law at all with such premises. People are lawful mainly because they know it's a good way to protect their long-term interests (economical, health and peace). Take away this and people should be lawful just because... they fear the police?
Because 'acca` gnisuno e` fesso (here, nobody is a moron).
Hint: I see OpenOffice for Solaris on Sparc, but I don't see OpenOffice for HP-UX on PA-RISC. Why?
I'd guess that (some) people at Sun believe that one day Solaris will make it to the non-techs desktops at large, while people at HPAQ basically don't.
In order to make it to the desktop, Sun needs (badly) something to replace the CDE, which is almost wasted disk space by today's standards (and IMHO also by yesterday's standards: NeXTStep provided a infinitely more useful desktop than CDE, and that was before the CDE was born in early '90).
Linux struggles with more than 4 CPUs and more than 4G of memory;
Please, put it on comparable hardware. Or let's talk about the very same problems popping up on Solaris for x86.
And even if Linux could do this, it doesn't run on hardware than can be dynamically partitioned, unless it runs as a guest on z/OS, and in that case it's z/OS doing the work.
Like somebody is going to care if it's doing it all by itself, just for the sake of it?
Tight integration with the underlying hardware is another advantage for Sun; they know precisely every component in every system that Solaris runs on, because they designed and built it, so there are never compatibility issues
That's the dumbest thing you could say: it's like saying that a good reason to choose a certain car is because there are no suppliers of spare parts for it that sells directly to you (but instead you have necessarily to go through the car "manifacturer"). It's not like you are going to replace the engine with a third-party one, but what about the windshield-wipers or such? Please don't tell that since you are spending X grands this doesn't matter, otherwise nobody would ever consider using something like Linux in the industry.
Solaris' high-performance, high-reliability filesystems are proven, not just betas (yes XFS is also proven, but in IRIX not Linux).
Yes, but where high-performance is a concern (video), people are moving to Linux, not Solaris. SGI has been wise enough to port XFS, because it knows Irix has no long-term future within the video industry.
Solaris has ACLs, whereas Linux just has the relatively crude user-group model.
And thanks to the No Such Agency, SELinux provides mandatory access control policies on more than just files. And they actually use it.
Slashdot Mirror
Well, it exists (it's called UNO), but quite obviously it is not the same.
At least it comes with nice bindings for Java, C++, Perl and Python.
A UNO-CORBA bridge was in the works, but I believe there is little interest in this.
With 3 poles, you need (2 ^ discs)-1 moves to solve the problem. With 64 disks, and one move per second it's more like 584 billion years.
Uhm, wasn't that just for trademarks?
The point about mmap(2) is to let the system (the VM subsystem of the kernel) manage the caching for the userland processes using it, avoiding extra copies to/from buffers in userland and eliminating in several cases the need for custom caching code (processes don't have to worry about data being available in RAM: the kernel automatically takes care of that when needed).
You don't need gobs of memory to do this, but in order to work on large amount of data you need a large address space, which is what 64 bit architectures provide. Of course, the more physical memory you have, the less the kernel has to swap pages in and out, but the main point is not that.
A little example to clarify: in order to keep things simple (instead of needing two intermixed caching systems, one for the VM and one for disk accesses), the Hurd just mmaps the whole partition. This means that the maximum size of a partition has an upper limit given by the size of the addressing space, which is 4GB on 32bit architectures (actually less, since in that address space you have to keep also the code that uses the mmapped data, so it's more like 2GB/3GB). A 64bit architecture comes very handy here, given the current size of hard disks.
There's something strange going on here: I repeated your test just for the sake of it using a CD that according to the codes you gave should be the very same, and for the same version of cdparanoia.
Please note that the ending sector is not the same as yours (tried also with option -T with no effect). As I said before, the CD bears the same codes as the ones you specified, and is clearly labelled "Made in USA" (I live in Europe). I also tried on another CD-reader, and the results are the same (ending sector is 17834). I'd expected that at least the TOC entries would be the same...
For further checking, here there is the output from cd-discid, if you want to check with yours:
Uhm, I believe it was the CP/M guys saying something along the lines of "64KB should be enough".
That quote actually is by Andrew Tanenbaum (from "Computer Networks"):
"Never underestimate the bandwidth of a station wagon full of tapes hurtling down the highway."
Instead, this one has been attributed to Linus:
Real Men don't make backups. They upload it via ftp and let the world mirror it.
Well, LyX allows you to use inline TeX/LaTeX if you need to.
The last time I tried Sketch, it was extremely lame and buggy
You have to use the stable release, 0.6.x (the Tkinter-based one), possibly not the versions pre-packaged by RedHat (which had serious problems, probably due to mistakes in the packaging of the Python Imaging Library). Never experienced problems with SuSE or Debian pre-packaged versions.
Sodipodi OTOH is coming out really well and has improved a lot lately, but it still misses some basic features like layers, styles, the ability to combine shapes while linearly transforming them, the ability to export as EPS (you need Batik or Sketch) and more advanced ones (scripting) which have been in Sketch for years, so I can't really recommend it for now.
I agree that proprietary tools are a better choice as of now, but using Illustrator just to retouch computer-generated diagrams is almost surely overkill.
It thus appears that extremely competent individuals suffer a burden as well. Although they perform competently, they fail to realize that their proficiency is not necessarily shared by their peers.
So it seems there could be two problems here, not just one.
Ok for the receipt to the commission, but I'm not completely sure about the receipt to the voter: let's say that some days before the elections someone comes to you telling how you should vote, "or else". And he requires that after the elections, you show him a proof that you actually voted as you were told.
This went so far in some areas of Italy that on the last (regional) elections the usage of photocameras and videophones were explicitly forbidden in the voting booth. And yes, someone actually tried anyways and was discovered (and his vote invalidated).
So, in some way, being unable to prove to someone else how you voted is not entirely a bad idea.
(of course it can be objected that the nasty guys could come after you anyways if the result of the elections is not the expected one, regardless of how you actually voted...).
Sketch at least is capable of reading .cmx files.
See also boost::lambda
If and only if they distribute such derived works. There is no obligation in the GPL to redistribute derived works (there can't be, being the GPL a permission to distribute copyrighted material at certain conditions, permission which you normally don't have).
Of course, if they don't redistribute the modified sources, they also get the burden of mantaining them. That sounds fair to me.
Signed and md5summed tarballs are fine, they don't keep me from getting the kernel sources, modify them, recompile and install.
OTOH, apparently I need some non-free tools and secret data (which I'm assuming that are not part of the normal distribution) in order to install a modified version of this gnuboy port.
Now, it can be argued that I don't really need them, it's just that without them the installation becomes higly inconvenient. But if it is so, why require the installation scripts at all in the source?
From the GPL:
Just wondering if the signing tool could fall in the highlighted category.
Try using a unsigned char i; as requested. D'oh! Perfectly valid code.
Not exactly a new idea.
OTOH, as I was forecasting in a comment of mine of some time ago (sorry, can't find it), I'm already seeing websites where if you follow a link, you are presented with an intermediate adv. page (not a popup, but a full page), which then forwards you to the real link target in 10 seconds or so. Just like ad pages in the middle of magazine articles.
I'm pretty sure it would be called ViMS, and it would run mainly on old VAXen. :-)
Then, some other people implement a shell and several command line utilities in Emacs Lisp (EShell).
Then some other people implement a Scheme system inside the Linux kernel.
Secret long-term goal: create a self-sufficient Emacs (codename Emax) that boots, thus obtaining the One True OS! :-)
Perhaphs you misunderstood me, or I'm misunderstanding you: mine is not a suggestion, but I was stating a fact: people choose to obey law because it serves best their long-term interests, and not because they fear the police.
If in order to punish those who were not obeying the law, also the ones obeying the law are punished... well, it's like saying "you fool!" to those who obeyed the law in the first place, and this is the thing I can't tolerate in the first place.
Morals aside, someone please tell me a reason why one should choose to obey the law at all with such premises. People are lawful mainly because they know it's a good way to protect their long-term interests (economical, health and peace). Take away this and people should be lawful just because... they fear the police?
Because 'acca` gnisuno e` fesso (here, nobody is a moron).
Hint: I see OpenOffice for Solaris on Sparc, but I don't see OpenOffice for HP-UX on PA-RISC. Why?
I'd guess that (some) people at Sun believe that one day Solaris will make it to the non-techs desktops at large, while people at HPAQ basically don't.
In order to make it to the desktop, Sun needs (badly) something to replace the CDE, which is almost wasted disk space by today's standards (and IMHO also by yesterday's standards: NeXTStep provided a infinitely more useful desktop than CDE, and that was before the CDE was born in early '90).
Please, put it on comparable hardware. Or let's talk about the very same problems popping up on Solaris for x86.
And even if Linux could do this, it doesn't run on hardware than can be dynamically partitioned, unless it runs as a guest on z/OS, and in that case it's z/OS doing the work.
Like somebody is going to care if it's doing it all by itself, just for the sake of it?
Tight integration with the underlying hardware is another advantage for Sun; they know precisely every component in every system that Solaris runs on, because they designed and built it, so there are never compatibility issues
That's the dumbest thing you could say: it's like saying that a good reason to choose a certain car is because there are no suppliers of spare parts for it that sells directly to you (but instead you have necessarily to go through the car "manifacturer"). It's not like you are going to replace the engine with a third-party one, but what about the windshield-wipers or such? Please don't tell that since you are spending X grands this doesn't matter, otherwise nobody would ever consider using something like Linux in the industry.
Solaris' high-performance, high-reliability filesystems are proven, not just betas (yes XFS is also proven, but in IRIX not Linux).
Yes, but where high-performance is a concern (video), people are moving to Linux, not Solaris. SGI has been wise enough to port XFS, because it knows Irix has no long-term future within the video industry.
Solaris has ACLs, whereas Linux just has the relatively crude user-group model.
And thanks to the No Such Agency, SELinux provides mandatory access control policies on more than just files. And they actually use it.
no but it does imply bad programming.
And a hell in long-term maintenance as well... (learned it the hard way).