The question is how customized: gentoo can make packages when it builds them, and after you set up a binary distribution point, there is no problem telling all the other machines to emerge -k package. So you get control, without the compilation on 80 machines. I do something like this, and have a central repository of config files, which are global & get passed to the machines, while anything that is custom gets a +i attribute, so even if I do something stupid it won't allow me to overwrite a file (unless the i attribute is removed, so etc-update runs into absolutely no problems, even if I allow it to overright files. As for install: once you have a running system, use a program to replicate it (numerous ghost clones, or just copy it the old fasioned way)
I like your idea, and if you want to work on it, gentoo already has some incomplete work done to allow it to create rpm packages (dependancy stuff is the problem...)
You can sort of do the above (kde 3 stuff from ebuilds) via portage building rpm and use alien to convert it to deb. Though as the rpm will not have the right deps, the deb will most certianly not. I would more than FHS like to see distros figure out how to use common depenancies. (are we doing package name, the.sos in it, etc) (most packages are unique enough that they don't conflict, and an idea of gentoo's virtual (virtual/motif for example) packages. (much cleaner than.so's when combined with a slot)
You are not required to accept the GPL to use GPL licensed software. Use is outside it's scope. Neither are you required to accept the Microsoft EULA, since it is just a shrink wrap
license and isn't legally binding
Actually on the box it says that you have to accept the enclosed licence agreement, and that seems to take care of the legal issues EULAs have had in the past: Modification of contract after sale. Since the term is now known at the time of sale, it is valid (as it can be). Now if it didn't (which some software companies still don't) then it would be bound by whatever you can do via fair use & whatever the box says (an example listed in the features: (which some might argue with) 8 player lan games, and no external requirement to accept an enclosed EULA, means you can use it for 8 player lan games, hence why any proprietary software company with a legal department has that printed on the outside of the box (which means most of the popular software).)
If you want it to be invalid you actually have to challenge the EULA itself, because they did that, a much tougher proposition.
I mentioned it AND that it doesn't do EVERYTHING on the GPU. Some things like geometry transforms (as I recall) are done on the CPU. Even using hardware accelerated, the CPU has traditionally had to say: do this, do this, do this for everything the grapics card does.
And if you think EVERY Mac (considering the number of mac classic & OS X (less than 10.2). Now had I restricted myself to the domain of new Macs then I would be incorrect. Windows will be starting to when Longhorn comes out, and we all know that Windows (the deployed ones) will not have the same display capablities for a very long time (aren't only 26% of Windows users are using XP by now?, MacOS likely has a higer ratio of 10.2+ than XP over earlier versions, but as a whole, is just starting to.)
You're a Microsoft apologist, aren't you? I thought your kind was extinct by now, especially here.
Wow: name calling & Inaccurate. I lying about things doesn't help the credibility of the speaker. (I have no Microsoft Operating systems in my house. Nor do I have MacOS. I do have various Linuxes, an IBM OS which is not OS/2, multiple sun (sparc32/64) boxes & Digital/Compaq/HP Alphas.)
Now my own rant, name calling & impoliteness to the presumed mac zealot (skip if you want to): oh, look at the UltraSparc... 64-bit desktop computing, new? Oh and that's an alpha, which until compaq bought Digital was beating the crap out of their precious Macs, and the low end were cheaper compared to high-end Macs (for much better performance) Again, Apple's Feces to Gold Machine that they have some people believing that anything they create is the 3rd coming, Steve being the 2nd. (not that all Apple products are bad, actually the G5 is not bad, nor is the ipod (I will argue that it is overpriced))
Actually what I'm thrilled about (even if others say its horribly inefficient) is the 3D accelerated desktop that is supposed to be in Longhorn, and doing away with 2D acceleration. The Mac has it, why can't we?
Umm... the Macs don't. The macs use display pdf which can be scaled much like vector graphics that longhorn will include. However Longhorn will do almost all of that on the card (Which macs are starting to do (Quarz Extreme which still does some things in software (CPU)).
Nor will longhorn be a '3d' desktop for the most part, instead it will be more like doing 2D acceleration in 3D spaces, which most cards cannot do efficiently. They mostly flush the render buffer for every switch & the 3D part is still seperate from the 2D portion with the end 3D buffer being blitted to the 2D buffer when it isn't full screen. The main benefit is: vector grapics (which can be done in 2D easily, (example: kde's crystal svg icons) but all 3D apis provide this accelerated if the hardware does it.) which allow smaller sized icons which can scale up & down better than bitmaps, and is useful for high resolution windows so that even if you need large things (poor eyesight) it can only look better running at higher resolution (by having the computer calculate how to display something at 200dpi to a monitor which was at 100dpi (it isn't hard, and if you don't use vector graphics, it is essentially just pixel quadrupling, however with vector grapics & aa, it looks better)
And for anyone who doesn't think cards need a lot of ram: my current desktop is using more than 12MB, and that's only going to go up.
Yes, but they are actually supporting someone else's standard. And lets face it, they wouldn't have unless a lot of the major OEMs hadn't twisted their arms.
Nah, It's how well you can make them come back to life!
Seriously, I have had 3 motherboards die (as in will not work period) & become resurected (full operational status) & stable 6 months later. (motherboards were: 1 socket A (Asus A7V), 1 K6-2 Super7, 1 dual socket 370 (serverworks chipset)
They might be able to be, but it might be better to just extend DRI a bit more, and have X able to talk to a specialized interface (as it does for DRI)
Basing it on the EULA is not valid, UNLESS it states on the outside of the box (visibly) that you have to agree to an (enclosed) eula. If it doesn't then it breaks down to modification of contract after sale according to at least one Federal Court (if you bought the box of software & they are the copyright holder). You could potentially get into a lot of trouble if you sold a linux distro & didn't mention that the software inside was licenced under various terms (see inside). Because without the Licence: GPL, LGPL... etc, you are in violation because you are selling something you don't have the rights to (if you don't accept the GPL (or other licence), it's a copyright violation, if you do, just print it visibly.
Please note that the use of it is now covered by copyright, and fair use applies (which also implies that it can be installed to a computer to use it). Which means that copying it & selling it (etc) are not legal, but things that take away fair use in the EULA (or any good permissions that aren't fair use are taken away with the usually bad restrictions).
Microsoft, Red Hat, and any software company with a legal department worth 0.02USD who wants to be legal (Linux distros, see above) or enforce the EULA (Microsoft for example) has it on the outside of the box, in which case it is mostly valid (a person would have to challenge it for what it says, such as one software company that put in your firstborn (pretty open & shut case of that part or whole of that EULA being invalid.))
With regards to the Intel thing, one might argue the 'extensions' they keep adding on seem to be attempts to make binary software somewhat incompatible with other x86 & now ARM cpu manufacturers.
MMX was certainly a reaction to WinChip developing 3dnow! which wasn't perfect, but better than MMX (and certianly was much more open than MMX (remember the fit Intel threw when AMD for example added them?)) One of my questions is if intel has finally added 3dnow! to it's x86-64 compatible processors.
Intel mostly supports intel made standards (though admittedly it can support other people's standards), but for the arguement you are making, intel is a VERY bad example.
Actually it has this right now, if you apply the grsecurity patches. Unfortunately certain programs (XFree86) will not run if this is enabled, according to the kernel help for it.
In the 'big iron' enterprise market against RISC where Itanium is beating everything handily (check out the latest TPC-C list Top 10 where Itanium holds spots 1,3,4,7,10
Yeah, spot #1 is held by IA-64 with 64 processors, and #2 by Power4 with 32 processors, as are all except #10 on that list (where every power4 is a 32 processor box)
Not to mention TPC-C is something for which vendors tweak heavily, and it is a fairly exclusive and expensive club to get into.
The only TPC comparison between Itanium & Opteron can be found in the 300GB TPC-H with a 2GHz 16-way opteron cluster (13,194) vs a 1GHz Itanium 2 (4,774) SMP box. Unless a 1.5GHz Itanium2 has a significant core change it isn't going to deal with the almost 4x lead, assuming the benchmarks are good, which I have some doubt of.
A week of on and off compiling, while I was using it on a daily basis, so it was mostly only working in the evenings, and sometimes durring days, assuming it was usable (some compiles are of course more processor hungry than others) And if FreeBSD can compile the c library, the c compiler, kernel, xfree86, qt, kde, python, perl, and many other packages in 2 hours on a p2-300 w/160MB RAM and a slow hard drive, then I would be amazed.
Your link is broken, so I can't see it without going and hunting on freebsd's site, but I suspect you were rebuilding a fraction of what I was, and had I been compiling them in one single session it would likely not have taken as long.
I use gentoo on a p2-300 (ca 1997-8) & frankly it makes it HIGHLY usable.
Most of the usablity improvements come from KDE 3.2 & linux 2.6 (though it was fast enough with 3.1.2 & 2.4) In fact, it is much more responsive than Windows (any version with a bit of security (not 9x, ME)), Fedora Core 1 (fresh install on a 1GHz P3-celeron, 256MB RAM & 10K rpm SCSI drive) Now admittedly the fedora installed very quickly. The recent updating included building most of the system over again (for about 6months only security related packages had been updated) which took about a week, but the system was still often usable while it was compiling (slow, but usable).
Some people have argued that in a particular library/program compiling for -march= with gcc (as opposed to most of the binary distros -mcpu=) only leads to a 2-3% improvement in speed per program/libary. If true (and only that much), that adds up, X is 3% faster, Qt is 3% faster, kdelibs is 3% faster, konqueror is 3% faster= 112.6% faster, which is better than the change between any of the Pentium 4 Extremely expensive edition, and Athlon 64 (of which I have not seen ONE mainstream benchmark site running in 64-bit mode...Windows (beta) or Linux) Many over clockers don't get that much performance increase stablily.
If you want to there is the GRP (Gentoo Reference Platform) which will install with binaries, so you don't have to compile everything.
SCO UNIX is actually more encumbered than Linux, because they are being sued, for violating 4 of IBM's patents (well 4 they chose to sue about). Guess what? They have dropped all thier copyright stuff against IBM. Where is Linux Legally Encumbered?
Having been on the internet since 1989 & never having used anti-virus that much (other people's computers might have it, or computers at labs). Nor has my account ever gotten a virus (detected them in other people's accounts, and killed them, but not personally had them)
Most important:
1. Don't be an idiot.
2. Kill Outlook Express or make sure it never gets run... (could be part of #1)
3. See parent poster
I recieve virii all the bloody time in email from idiots (for me it's worse than spam, yet still a very small amount...) YET they never run, because you had better believe that #1 covers running unsigned binary attachments from people you don't know...
Patent-pending? When were you first on the internet? (Comments about thinking all the money grubbing IP.commers were either broke or at amazon.com....or SCO)
The internal storage on a zarus is similar to the bios/harddrive on a PC, so no, unless you manage to corrupt the file system (can happen, but who sane is going to dd to the flash?)
Sharp's original Zaurus (5000D/5500) had a rom (& due to hw limits, continues to act like this) which acted like one and stored all user created docs/installed programs in ram (32MB was storage, 32MB real ram)(with a certain setup so that unless you corrupt it (unlikely) or run out of batteries (much more likely) that it would die.
This is the reason that the later Zarii have more flash, and less ROM, Also, because of it being linux, there are full distros www.openzaurus.org which make the flash writable, and use no RAM as storage, so that docs/programs can install to it & aside from data corruption, you wont lose it. Additionally, All Zaurus models can store data/install programs to CF/SD, and backup the config to those.
I have never had my Zaurus lose data, except when I decided to try an experemental openzaurus (flash the rom) and forgot to backup, which was my fault.
Short answer: not unless you are REALLY unlucky/stupid(ex: dd).
I suspect that the Athlon 64 "A64" is a much better 32-bit chip than an AXP (more commonly called an Alpha)
Too bad Compaq & HP killed off the Alpha (the 364 is still a 264 (and was untill last year still kicking ass.) and scales better than almost anything) the 464 was to be the first new core, and was up to a 5X performance increase because of the 1st major addition to the alpha arch (Brief overview) It has been speculated that the reason HP doesn't release Alpha 364 scores, it that it doesn't want to embarass intel (& cause more ummm... stuff to fall on the pro IA-64 group of which HP is the only major OEM), which may or may not be true.
Historically, its been Intel introducing the extensions, and AMD quietly implementing them in response. The x86-64 extensions are shaping up to be the first example of an AMD-pioneered extension that Intel will implement!
1: no, it was never out of beta (nontheless, it could be obtained)
2: I don't recall it being a full 64-bit, just the it's 64-bit, but we will have it act like 32-bit that was done for NT.
Slashdot Mirror
Not to mention a PDA *already* runs Basilisk 2 http://www.killefiz.de/zaurus/showdetail.php?app=6 15
I like your idea, and if you want to work on it, gentoo already has some incomplete work done to allow it to create rpm packages (dependancy stuff is the problem...)
You can sort of do the above (kde 3 stuff from ebuilds) via portage building rpm and use alien to convert it to deb. Though as the rpm will not have the right deps, the deb will most certianly not. I would more than FHS like to see distros figure out how to use common depenancies. (are we doing package name, the .sos in it, etc) (most packages are unique enough that they don't conflict, and an idea of gentoo's virtual (virtual/motif for example) packages. (much cleaner than .so's when combined with a slot)
Actually on the box it says that you have to accept the enclosed licence agreement, and that seems to take care of the legal issues EULAs have had in the past: Modification of contract after sale. Since the term is now known at the time of sale, it is valid (as it can be). Now if it didn't (which some software companies still don't) then it would be bound by whatever you can do via fair use & whatever the box says (an example listed in the features: (which some might argue with) 8 player lan games, and no external requirement to accept an enclosed EULA, means you can use it for 8 player lan games, hence why any proprietary software company with a legal department has that printed on the outside of the box (which means most of the popular software).)
If you want it to be invalid you actually have to challenge the EULA itself, because they did that, a much tougher proposition.
And I don't have them with me & haven't used them as Personal Computers. :)
And if you think EVERY Mac (considering the number of mac classic & OS X (less than 10.2). Now had I restricted myself to the domain of new Macs then I would be incorrect. Windows will be starting to when Longhorn comes out, and we all know that Windows (the deployed ones) will not have the same display capablities for a very long time (aren't only 26% of Windows users are using XP by now?, MacOS likely has a higer ratio of 10.2+ than XP over earlier versions, but as a whole, is just starting to.)
You're a Microsoft apologist, aren't you? I thought your kind was extinct by now, especially here.
Wow: name calling & Inaccurate. I lying about things doesn't help the credibility of the speaker. (I have no Microsoft Operating systems in my house. Nor do I have MacOS. I do have various Linuxes, an IBM OS which is not OS/2, multiple sun (sparc32/64) boxes & Digital/Compaq/HP Alphas.)
Now my own rant, name calling & impoliteness to the presumed mac zealot (skip if you want to): oh, look at the UltraSparc... 64-bit desktop computing, new? Oh and that's an alpha, which until compaq bought Digital was beating the crap out of their precious Macs, and the low end were cheaper compared to high-end Macs (for much better performance) Again, Apple's Feces to Gold Machine that they have some people believing that anything they create is the 3rd coming, Steve being the 2nd. (not that all Apple products are bad, actually the G5 is not bad, nor is the ipod (I will argue that it is overpriced))
also known as the way agnostics think christians are:
open (PERLYGATES) || die (trying);
They don't care about the compiler: some might realize that if the first is true, then the second doesn't need to be run, but others might run both...
And the athiest's view:
open(PERLYGATES) && die(trying);
Umm... the Macs don't. The macs use display pdf which can be scaled much like vector graphics that longhorn will include. However Longhorn will do almost all of that on the card (Which macs are starting to do (Quarz Extreme which still does some things in software (CPU)).
Nor will longhorn be a '3d' desktop for the most part, instead it will be more like doing 2D acceleration in 3D spaces, which most cards cannot do efficiently. They mostly flush the render buffer for every switch & the 3D part is still seperate from the 2D portion with the end 3D buffer being blitted to the 2D buffer when it isn't full screen. The main benefit is: vector grapics (which can be done in 2D easily, (example: kde's crystal svg icons) but all 3D apis provide this accelerated if the hardware does it.) which allow smaller sized icons which can scale up & down better than bitmaps, and is useful for high resolution windows so that even if you need large things (poor eyesight) it can only look better running at higher resolution (by having the computer calculate how to display something at 200dpi to a monitor which was at 100dpi (it isn't hard, and if you don't use vector graphics, it is essentially just pixel quadrupling, however with vector grapics & aa, it looks better)
And for anyone who doesn't think cards need a lot of ram: my current desktop is using more than 12MB, and that's only going to go up.
Yes, but they are actually supporting someone else's standard. And lets face it, they wouldn't have unless a lot of the major OEMs hadn't twisted their arms.
Seriously, I have had 3 motherboards die (as in will not work period) & become resurected (full operational status) & stable 6 months later. (motherboards were: 1 socket A (Asus A7V), 1 K6-2 Super7, 1 dual socket 370 (serverworks chipset)
They might be able to be, but it might be better to just extend DRI a bit more, and have X able to talk to a specialized interface (as it does for DRI)
Please note that the use of it is now covered by copyright, and fair use applies (which also implies that it can be installed to a computer to use it). Which means that copying it & selling it (etc) are not legal, but things that take away fair use in the EULA (or any good permissions that aren't fair use are taken away with the usually bad restrictions).
Microsoft, Red Hat, and any software company with a legal department worth 0.02USD who wants to be legal (Linux distros, see above) or enforce the EULA (Microsoft for example) has it on the outside of the box, in which case it is mostly valid (a person would have to challenge it for what it says, such as one software company that put in your firstborn (pretty open & shut case of that part or whole of that EULA being invalid.))
MMX was certainly a reaction to WinChip developing 3dnow! which wasn't perfect, but better than MMX (and certianly was much more open than MMX (remember the fit Intel threw when AMD for example added them?)) One of my questions is if intel has finally added 3dnow! to it's x86-64 compatible processors.
Intel mostly supports intel made standards (though admittedly it can support other people's standards), but for the arguement you are making, intel is a VERY bad example.
Actually it has this right now, if you apply the grsecurity patches. Unfortunately certain programs (XFree86) will not run if this is enabled, according to the kernel help for it.
Linux uses it to a certain extent, but only 2-3 layers total. (1 or 2 for kernel, 1 for user space) I suspect that most others are similar.
Or leet, they look equally unintelligible at first glance.
Yeah, spot #1 is held by IA-64 with 64 processors, and #2 by Power4 with 32 processors, as are all except #10 on that list (where every power4 is a 32 processor box)
Not to mention TPC-C is something for which vendors tweak heavily, and it is a fairly exclusive and expensive club to get into.
The only TPC comparison between Itanium & Opteron can be found in the 300GB TPC-H with a 2GHz 16-way opteron cluster (13,194) vs a 1GHz Itanium 2 (4,774) SMP box. Unless a 1.5GHz Itanium2 has a significant core change it isn't going to deal with the almost 4x lead, assuming the benchmarks are good, which I have some doubt of.
Your link is broken, so I can't see it without going and hunting on freebsd's site, but I suspect you were rebuilding a fraction of what I was, and had I been compiling them in one single session it would likely not have taken as long.
Most of the usablity improvements come from KDE 3.2 & linux 2.6 (though it was fast enough with 3.1.2 & 2.4) In fact, it is much more responsive than Windows (any version with a bit of security (not 9x, ME)), Fedora Core 1 (fresh install on a 1GHz P3-celeron, 256MB RAM & 10K rpm SCSI drive) Now admittedly the fedora installed very quickly. The recent updating included building most of the system over again (for about 6months only security related packages had been updated) which took about a week, but the system was still often usable while it was compiling (slow, but usable).
Some people have argued that in a particular library/program compiling for -march= with gcc (as opposed to most of the binary distros -mcpu=) only leads to a 2-3% improvement in speed per program/libary. If true (and only that much), that adds up, X is 3% faster, Qt is 3% faster, kdelibs is 3% faster, konqueror is 3% faster= 112.6% faster, which is better than the change between any of the Pentium 4 Extremely expensive edition, and Athlon 64 (of which I have not seen ONE mainstream benchmark site running in 64-bit mode...Windows (beta) or Linux) Many over clockers don't get that much performance increase stablily.
If you want to there is the GRP (Gentoo Reference Platform) which will install with binaries, so you don't have to compile everything.
They meant Linux is Legally Unencumbered.
and SCO Unix is Legally Encumbered.
Most important:
1. Don't be an idiot.
2. Kill Outlook Express or make sure it never gets run... (could be part of #1)
3. See parent poster
I recieve virii all the bloody time in email from idiots (for me it's worse than spam, yet still a very small amount...) YET they never run, because you had better believe that #1 covers running unsigned binary attachments from people you don't know...
Patent-pending? When were you first on the internet? (Comments about thinking all the money grubbing IP .commers were either broke or at amazon.com....or SCO)
Sharp's original Zaurus (5000D/5500) had a rom (& due to hw limits, continues to act like this) which acted like one and stored all user created docs/installed programs in ram (32MB was storage, 32MB real ram)(with a certain setup so that unless you corrupt it (unlikely) or run out of batteries (much more likely) that it would die.
This is the reason that the later Zarii have more flash, and less ROM, Also, because of it being linux, there are full distros www.openzaurus.org which make the flash writable, and use no RAM as storage, so that docs/programs can install to it & aside from data corruption, you wont lose it. Additionally, All Zaurus models can store data/install programs to CF/SD, and backup the config to those.
I have never had my Zaurus lose data, except when I decided to try an experemental openzaurus (flash the rom) and forgot to backup, which was my fault.
Short answer: not unless you are REALLY unlucky/stupid(ex: dd).
Too bad Compaq & HP killed off the Alpha (the 364 is still a 264 (and was untill last year still kicking ass.) and scales better than almost anything) the 464 was to be the first new core, and was up to a 5X performance increase because of the 1st major addition to the alpha arch (Brief overview) It has been speculated that the reason HP doesn't release Alpha 364 scores, it that it doesn't want to embarass intel (& cause more ummm... stuff to fall on the pro IA-64 group of which HP is the only major OEM), which may or may not be true.
Ode to Alpha
Great... Intels will finally have 3dnow! :)
1: no, it was never out of beta (nontheless, it could be obtained)
2: I don't recall it being a full 64-bit, just the it's 64-bit, but we will have it act like 32-bit that was done for NT.