The next version of the heart of the Linux operating system is expected by June, project founder and leader Linus Torvalds predicted on Thursday. “We’re pretty close to done with what will be 2.6…We’re actually looking at the second quarter 2003 for the real 2.6 release,” Torvalds told a group of Linux aficionados aboard a “Geek Cruise” in the Caribbean. News.com reviewed a recording of Torvalds’ talk.
Torvalds on 2.6 Release, 3.0 Name Debate, MacOSX, IA-64, BitKeeper
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Basically a bunch of quotes of what Linus said. But given that I can’t find a transcript, not a bad resource if you ignore the “commentary.” There is a page that sums up some points he made in his speech: http://blog.linuxjournal.com/mt/doc/
As for kernel 2.6, try out the development kernels. I know it sounds like they are a long ways from release, but over here, they’re pretty stable. It’s just that a lot of cleanup up remains and a lot of drivers don’t yet work with all the code changes. If you can get the thing to compile (a lot of stuff hasn’t yet been updated, like I said) then you should be okay. But of course, I rsync everynight to a backup server, so I can afford to be a little reckless
“I never much liked Macs. All the interesting stuff is hidden away,” he said, adding that he does like Apple hardware. …”
Heeheee! That’s funny!
Well I actually use to like Macs, I gave up Macs when NeXT took over Apple. (still have a 7600 132 and a IIci along with an ImageWriter II
I must say that I do like Apple hardware. They are of very good quality and design.
I don’t like OS X! I alway hated all the different APIs. After reading Maarten Hekkelman’s comments here: http://www.osnews.com/story.php?news_id=1659 I really hate it. Ever wonder why Quark can port their stuff over?
I am really starting to like Linux. I only wish the GUI (GTK 2.0) was easier to program. Something like the BeOS toolkits would be great! I wonder if Gobe could release their BeOS API libraries for Linux! Heck I think many people would pay big bucks for it.
Oh well, I look forward to the next Linux kernel.
ciao
yc
but linus’ pictures remind of commander data
Hmmm, he does look like Commander Data!
He also looks a bit like a Penguin with the nose.
Doesn’t his wife like Penguins? Thus the Linux mascot…
Well, there you have it!
ciao
yc
Linus has some good points about OSX. It would be better if it was released under the GPL or BSD license…all of it. OpenDarwin is one thing, but from looking at it, it appears that OpenDarwin’s purpose is two-fold: get good code from open source/free programmers, and make good Unix apps (X, OpenOffice) work under OSX. I like OSX (am using it right now, as a matter of fact), but I do sometimes question Apple’s dedication to free software. It doesn’t make much sense, considering…this has been pointed out *many* times on this board…that Apple is a hardware company. They don’t make their money from the excellent software they develop.
Whether or not Apple’s hardware is good is totally subjective. To me, today, it looks not much different than standard PC hardware.
As for his comments on Mach….I wonder if that’s just another pot shot at the “hurd people.”
OS X would be better for _you_ if it was released under the GPL or BSD license but _not_ for Apple. I don’t think anything else needs to be said about this.
I question how much Torvalds knowd about the Darwin kernel — It is _not_ simply mach, but rather based on mach — plus it’s not a true microkernel implimentation.
Oh well.
“Doesn’t his wife like Penguins? Thus the Linux mascot…”
Actually, it is Linus who likes the Penguins. His wife is the one who suggested that he use the Penguin.
The first MacOS X, that was release, aka, MacOS X Server, had a very good UI. I was basically a ported version of NeXT w/ a Mac Interface. Everything was still postscript and boy, was it fast. Now, personally, that is what they should have done. Stuck with the time proven postscript + X11.
Not sure if it makes any difference but isn’t Mac OS X based off of OpenStep, rather than directly descended from NeXTSTEP?
yc:
I don’t mean to start a flame war of which is better or anything like that, but from my experience QT is quite a nice GUI to program with. I’ve never used the BeOS stuff but using QT is loads nicer IMHO than using gtk+ (1 or 2)
The point isn’t really what engineering decisions were made or the mucky details of this or that. The point is that it works, it’s easy to use, it works, it’s beautiful, and it works. Did I mention it works?
what the heck…you “did not like all the APIs” ??!!
Coaca for New OS X apps, Carbon for old ones that need to move into OS X quickly while the company makes a coaca version.
hmmm…. so you moved to Linux…we have GTK+, QT, TK, Motif, <enter API here>
Linux has just as many APIs as windows……
you are aware that you can use coaca with C right…C is the base set of Obj C…C is to Obj C as Structured programming in C++ is to OOP in C++.
and since you like GTK+, I assume you can program in C.
One API on Apple YC, One. that is coaca…carbon is being dumped next year..it was a transitional API….and you also have Python bindings for coaca, java bindings, heck I bet some company will make C++ bindings for it soon.
I would say 2.6. When Linux came 2.0 – there was a pretty good darn reason. But for 2.5, while there is a lot of features, can’t account for a 3.0. Otherwise, 2.2 would be 3.0, 2.4 as 4.0 and now 2.5 would be 5.0 🙂
Please, don’t let the Sun virus get into you…
On the Itanium issue, it is VERY different from the Opteron. The Itanium was meant to compete with POWERs and SPARCs, not Athlons and Opterons. Hammer is meant to compete with Pentiums and Xeons, not Itanium (or POWER or SPARC).
Itanium is targeting a profitable high end niche. Their processor is quite fast actually – what it lacks is a good compiler. Something HP has, but is keeping to themselves. But hey! The architecture is very new.
Until the distant future, it is unlikely Opterons and Itaniums would go head to head. When that happens, Itanium would probably have a large market share with a lot of high end applications, while Opteron no.
Plus, it is not like the current 32-bit market needs 64-bit.
yc: Well I actually use to like Macs, I gave up Macs when NeXT took over Apple. (still have a 7600 132 and a IIci along with an ImageWriter II
Well, I know many people gave up on OpenStep when Apple bought NeXT. Like it or not, NeXT had really good stuff in its OS – stuff that Apple later on messed up.
Well, on Mac OS X’s new APIs – they are quite good at first glance. Carbon and its predesscors were crap, pulling together 18 years of useless bloat. But then programming in Obj. C isn’t what many like to do….
yc: I only wish the GUI (GTK 2.0) was easier to program. Something like the BeOS toolkits would be great!
Try QT. It is very similar to BeAPI.
IFightMIBs: and make good Unix apps (X, OpenOffice) work under OSX.
Their purpose of OpenDarwin is to properly compete with Linux and FreeBSD by running every imaginable UNIX app available. With any changes. For X11 and OpenOffice.org, a lot of changes was done to work properly with Quartz.
IFightMIBs: considering…this has been pointed out *many* times on this board…that Apple is a hardware company. They don’t make their money from the excellent software they develop.
No. Look at the hardware. What do they charge for it? It is insane. How do they sell it? With the software OF COURSE. Now, would I purchase a Mac when all the features of it was on Linux or FreeBSD? Nope, never. The software is used to sell hardware.
Matthew Gardiner: Everything was still postscript and boy, was it fast. Now, personally, that is what they should have done.
This is because eye candy haven’t creeped in. If I was to implement all the eye candy on, say, X11, it would be many times slower than Mac OS X. Quartz is a remarkably fast thing, unfortunately Apple decided to use useless eye candy to spoil its name.
I doubt the software implementation of Quartz is faster than the hardware accelerated implementation in X11. I’ve written high-performance graphics code before, and even for very simple primitives (which traditionally hurt hardware because of the higher per-operation overhead) you can’t touch hardware, not these days.
you are aware that you can use coaca with C right…C is the base set of Obj C…C is to Obj C as Structured programming in C++ is to OOP in C++.
One API on Apple YC, One. that is coaca…carbon is being dumped next year..it was a transitional API….and you also have Python bindings for coaca, java bindings, heck I bet some company will make C++ bindings for it soon.
You can call C and C++ code from an Objective-C Cocoa program, but I don’t believe you can actually code in Cocoa without Objective-C. And that’s Apple’s big problem. Carbon, as crappy as it it, can never go away as long as Objective-C is the only other choice.
Ask Adobe and Macromedia – there’s no way in hell that they’re going to write Photoshop, Flash, etc. once in C++ for Windows and again in Objective-C for MacOS. It’s one thing to write a wrapper layer around some platform-specific APIs, but to be forced into a specific language for an ultra-minority platform isn’t worth it.
As elegant as Objective-C and Cocoa might be, nobody but Mac-only software houses are interested.
Carbon ain’t going anywhere. Not until we can program for MacOS in C++.
I am really starting to like Linux. I only wish the GUI (GTK 2.0) was easier to program. Something like the BeOS toolkits would be great! I wonder if Gobe could release their BeOS API libraries for Linux! Heck I think many people would pay big bucks for it.
—
Now, I am assuming that you have Windows programming experience, if so, have you looked at wxWindows or qt? apparently both are quite kind to Windows programmers.
As for programming with GTK2/X11, like Windows you can either talk directly to GDI/GDI+, aka X11, or via Win32 API, aka, widget set. Both have their strengths and weaknesses.
Openstep was the API. For a number of years SUN teamed up with NeXT to produce a cross platform API, which was called OpenStep. OpenStep ran ontop of NeXT. There was also an OpenStep API for Windows too. The rumour shop has it that Microsoft was starting to get scared when they saw this development.
A couple of years back there was a debate between Linus and Tananbam, basically, the same spelling as the German word for Christmas Tree.
Anyway, the argument by Linus was that Mach kernels were overly complex and when up weighed up by the so-called benefits, the Mach kernel either is only marginally better in some cases, however, in most, worst in performance over the monolithic kernel structure Linus prefers.
Tananbam came back stating that Mach kernels were more efficient and logical, there for better (in a nut shell). So, for the next couple of months there was a grudge match between the two programmers.
If you are in the “Linux community”, you will know what I mean.
Personally? neither of them are great when taken to the logical extreme. Personally, take the positives from both and create a balanced kernel that isn’t too big, but isn’t too small to the point of being stupid.
Regards to the Opteron comments. There is backwards compatibility, HOWEVER, alot, IIRC, all the really old crusty stuff has been removed to make way for new 3D Now! instructions and so-forth.
As for Itanium and its so-called niche. It was not only meant to complete in the server, but in the workstation space with SGI and SUN where workstations can cost up to $20,000 a piece.
The fact remains that there are still no workstation applications, that is, CAD engineering tools to take advantage of the 64bit processor, the current compilers, in all due respects are crap, and to top it all off, no one is going to develop for it when it costs a kings ransom to buy!
If they charged slight more, say $100 more than the Xeon, I’ll put money on it, they could EASILY make money plus a whole new market will develop creating high end applications for it. Engineering, scientific, programming, image manipulation.
Personally, they should have used the Alpha intellectual property they obtained, and simply gave it an overhaul. It would cost a hell of alot less to develop + would have remained compatible with the compilers etc that had already been developed and matured.
As for AMD, I wouldn’t mind them selling MIPS 64bit processors. IIRC, they have just recently got a license to allow them to do that. Previously they only had a license to produce 32bit MIPS chips. If AMD produced a MIPS processor, 300Mhz 2MB of cache, and sold for the price of a Xeon, I would be more than willing to pay for it.
About the only thing I am disappointed in is the fact they, AMD, have decided to keep with the old crusty BIOS. This should have been their opportunity to either replace it with OpenBoot, which is my preference, or license the PXE that Intel uses in their Itanium.
I would like to see a Triple threat match between Steve Jobs, Bill Gates, and Linus Torvalds with Jean Louis Gassee as special guest referee
Personally, I would have to disagree with Linus on the subject of the Itanium machines. Itanium1 was indeed slow, but Itanium2 deserves most of its price. Check the SPEC results.
As for being so incompatible and having a new instruction set and philosophy, I am all for it, for the removal of the legacy stuff. I give Intel and HP props for that.
I think Linus doesn’t like it because he has to code too differently for it. ;D
note: it was minix not the mach set…
For those of you a little confused:
Darwin is not strictly a micro kernel, it’s based on mach 2.5 (I think). Mach was “not really a microkernel” till 3.0. But 2.5 is pretty microkernel-esque and there are still performance issues.
Performance is one of the main reasons Hurd is moving from mach to L4. Apparently, L4 outperforms mach by a pretty large amount.
Perhaps apple should have gone with L4 instead of mach?
Isn’t BeOS’s kernel based on like mach 1.0 or something? Don’t flame for this, I am sincerely curious.
>Isn’t BeOS’s kernel based on like mach 1.0 or something?
No. BeOS has a unique design, however, back in the early days it had some influences from XINU and it was “somewhat” microkernel. But the “modern” BeOS, 1998 and later, does not resemble any other kernel/OS design, it is really unique. And no, BeOS is not a Unix. The fact that it has a bash terminal doesn’t make it a Unix (WindowsNT/2k/XP are also Posix compliant I think, but that wouldn’t make it a unix
.
As I started reading the commentaries about this particular news, as I often do at this time of the night after a hard week at work, I stumbled upon yet another moronic post by our good friend YC. The guy’s been around for as long as I can remember (that probably means… hmmm.. since the happy days of BeNews, as far as I’m concerned), and for some reason, I never really paid attention to him, because hey, we’re all reasonable, polite people (aren’t we?), we care about the topic, we express our views in a mature, elegant manner and, all in all, we try to make the world a better place. For some weird reason, tonight I snapped. Who are you, YC? Why do are you still here? What keeps you alive? Fuck, I can’t find the words to express how much I despise you.
aaaaaaaaaaaaahhhhhhhhhhhhhhhh
I feel much better.
Thank you.
You can program Cocoa in C. The Objective C runtime is exposed by a set of C libraries. You can create classes, call methods, create objects, lookup selectors, the works from C. But it’s just so much easier to do this in Objective C so there is really no need to do this in C.
NT/2k/XP’s posix layer isnt compliant. Very very broken is a better description =).
BeOS POSIX layer was not 100% compatible either… If I had to give it a.. number, I would say it was close to 80-85% compliant.
(AtheOS is not fully compliant either, but it is more compliant than BeOS ever was. I think QNX has the best certification for POSIX compliancy).
What do you mean? I simply stated that the whole Mach issue had be bashed away in a mailing list a year or so ago, so, why the heck did you bring up minix. I didn’t even mention!
“A couple of years back there was a debate between Linus and Tananbam, basically, the same spelling as the German word for Christmas Tree.
Anyway, the argument by Linus was that Mach kernels were overly complex and when up weighed up by the so-called benefits, the Mach kernel either is only marginally better in some cases, however, in most, worst in performance over the monolithic kernel structure Linus prefers.
Tananbam came back stating that Mach kernels were more efficient and logical, there for better (in a nut shell). So, for the next couple of months there was a grudge match between the two programmers.
If you are in the “Linux community”, you will know what I mean.
Personally? neither of them are great when taken to the logical extreme. Personally, take the positives from both and create a balanced kernel that isn’t too big, but isn’t too small to the point of being stupid.”
it wasnt mach it was minix…that linus and andrew argued about…
Yeah, I can program in C/C++ on any OS. Thanks for the info on QT and other libs. I had read about QT before and I think it’s a nice package although pricey. I never liked the idea of making an entire app totally dependent on a 3rd party product even if it does offer great cross platform advantages.
I prefer to use what the OS comes with. Or perhaps a lib from a major compiler vendors like Metrowerks or Borland. Especially when it’s free
ciao
yc
>>what a moron!! Macosx is for consumers mostly and they dont give
>>a shit about command lines and geeky unix!!
>>he simply does not get it
Hmmm, well…
How many consumers buy Rack mountable servers?
I think Linus does get it!
Apple has lost much of the consumer market to Winblows.
They have lost a lot of the education market to Dell as well.
The consumer market is no longer enough, Apple is trying to go for the enterprise with BSD underpinnings etc… I think Linux will have more success in the enterprise than OS X. Especially with IBM, HP/Compaq and now Sun Microsystems behind it.
With Windows XP professional and Home edition Microsoft is really squezing Apple! Just think, Apple no longer has the advantages it use to have:
1. No more GUI advantage.
2. No more Audio/Video Advantage.
3. No speed advantage (not that Apple ever had that)
4. No more desktop publishing apps advantage.
5. BSD underpinnings? XP & Linux have been there, done that…
$200 Linux desktops will get better and better.
.Net will soon come out with advanced “managed apps” technologies.
All Apple has at this point is nice hardware design and fanatics.
OS X does not give Apple any major advantage. Especially with proprietary PC class hardware.
ciao
yc
Rayiner: I doubt the software implementation of Quartz is faster than the hardware accelerated implementation in X11.
You are comparing *software* implementation of Quartz vs. *hardware* implementation of X11. X11 can be fast, but it is limiting. Now, if say Quartz was something open and you and me could easily take it, expose it to plain vanilla hardware perhaps via GGI/KGI or DirectFB – well, I’m quite darn sure it would be faster.
Anonymous: Ask Adobe and Macromedia – there’s no way in hell that they’re going to write Photoshop, Flash, etc. once in C++ for Windows and again in Objective-C for MacOS.
It is actually quite easy to convert an C++ application to a Objective C++ application.
Matthew Gardiner: The rumour shop has it that Microsoft was starting to get scared when they saw this development.
Yes, there was a rumour that Microsoft delibrately killed Kurt to end AtheOS development because they were scared when they saw this development.
Matthew Gardiner: As for Itanium and its so-called niche. It was not only meant to complete in the server, but in the workstation space with SGI and SUN where workstations can cost up to $20,000 a piece.
I never actually meant it for servers only. High end workstations are also a big market for Itanium. in fact, probably bigger than the high end server itself. 🙂
But one thing for sure, it sure is *profitable*.
Matthew Gardiner: The fact remains that there are still no workstation applications, that is, CAD engineering tools to take advantage of the 64bit processor, the current compilers, in all due respects are crap, and to top it all off, no one is going to develop for it when it costs a kings ransom to buy!
Well, the only thing I actually agree is that compilers are crap. It fact, that is an understatment. But for CAD engineering tools – there isn’t much needs for 64-bit as for now (there is an high end niche for the CAD market, controled by SGI…. mainly for national defense purposes…).
As for the price, the price is way cheaper than of SPARCs (of similar capablities, of course) and POWERs and MIPSs. Notice two major players of this niche, HP and SGI had moved or made plans to move to Itanium. I don’t see how the price is hindering them. The big challege for ISVs is not optaining a Itanium chip, but porting their software to Linux.
Matthew Gardiner: If AMD produced a MIPS processor, 300Mhz 2MB of cache, and sold for the price of a Xeon, I would be more than willing to pay for it.
Unfortunately, not many would.
Travis: I would like to see a Triple threat match between Steve Jobs, Bill Gates, and Linus Torvalds with Jean Louis Gassee as special guest referee
Bad referee if you ask me. I mean, come on, he hates Jobs’ guts from his time with him at Apple. Then again, he also hates Bill Gates. Andd Linus Torvalds. On second thoughts, maybe placing him as a referee might be a good idea.
Hiryu: Perhaps apple should have gone with L4 instead of mach?
Too late now, moving to L4 would break binary compatiblity, and in variying cases source compatiblity.
Eugenia: (WindowsNT/2k/XP are also Posix compliant I think, but that wouldn’t make it a unix
.
Posix doesn’t make it UNIX. bash does. Does Windows have bash? No. sure, it get bashed a lot, but hey, not counted. 🙂 (No, I’m kidding).
Eugenia: BeOS POSIX layer was not 100% compatible either… If I had to give it a.. number, I would say it was close to 80-85% compliant.
But Windows NT’s was quite broken. 50-60% compliancy, at best.
armin: what a moron!! Macosx is for consumers mostly and they dont give a shit about command lines and geeky unix!!
Which is he not. He couldn’t care less about the Genie effects or the fact that there is such things as blue jelly beans. He thinks the architecture sucks. So?
The “consumer” OS is also trying to get the low end server market. And with a slow OS (by all means, Darwin is certainly slower than Linux) on slow hardware (don’t get me started!)….
1. No more GUI advantage.
Actually, they do still have that advantage. Mac OS X, for me, a Linux (well, ex-Linux until I save up enough money to buy a new PC) and Windows user, has the best UI among modern (read: not something that is no longer supported, nor not coming out for the next 10 years *hint* *hint* Amiga OS 4).
Sure, that advantage is diminishing at the rate Apple is destroying that advantage with Aqua and Microsoft spending some time on their UI….
But hey, Apple after it all, might just file an antitrust suite against Microsoft for improving 🙂
yc: 4. No more desktop publishing apps advantage.
Actually, they do have an advantage on that. Quartz supports native CMYK which overall improves speed tremedously. Plus the fact that most plugins/filters for QuarkExpress and Illustrator is Mac-only.
yc: Especially with proprietary PC class hardware.
PC class hardware? Puh-leze, that is an insult to the highest degree for modern PCs 🙂
what the ###? afaik you are somehow “doomed” if you use the development kernel because in case of data loss – you should be blamed ? sorry i am not such a big fan of beta testing a operating system kernel just because i want my graphic tablet driver to work. even if it works i have to patch it 165 times paint a pentagram on the floor and pray for my x-server to be of the right x.x.x.x subversion cause today if you dont use redhat or suse you are doomed too.
so what development kernel version is usable under normal production circumstances? i see no chart listing what works and what not – in what patch etc. for example redhat applies dozens of patches to their kernels, suse and mandrake too. so if you download the kernel.org you are really on your own. only thing you got is a changelog about some functions that have been changed – a mailinglist where noone knows what you mean ( who wonders – i dont understand the changelog either ) and readme’s last changed in year 98.
Just wanted to point that out.
Mach can be used to build monolithic kernels, “microkernels”, and kernels somewhere between the two extreams. Mach is a starting point to abstract hardware. It wasn’t explicitly made to just be a microkernel. OSF (now known as True64) UNIX uses mach to build thier kernel. It has some elements of microkernel design using message passing though mach, but the service modules are not in user space. The modules are also not 100% independent of eachother, which I think is a good thing from a performance standpoint. This also allows for one to keep the fork()… exec() model. I rather hate the windows spawn() model.
If Linux is so great, then why hasn’t it made it to the desk top? I mean, it’s been arround since 1991 (as long as Windows 3x). It’s open source, it has a pool of millions of developers.
Cry all you like about Apple and Mordorsoft, but they can harness the power of thousands to move in one direction, which is why their OSes are Consumer/Business usable. A decade of geek fiddling is just now giving us Linux flavors that are somewhat desktop useable. Linux has been a decade long beta, and nobody ever conquered the world in beta.
Apple hardware is pricey. It lasts. You pay more for an Acura RSX than a Ford Focus. Used Apple systems can be had very cheaply and can be easily upgraded to run OSX. And if you don’t like OS X, then you can evilBay it and recoup your investment.
And finally, why should Apple make all of their OS open to the public? They spent Millions to develop it and make it USEABLE (and beautiful). It’s a business, not a charity.
You have their Kernel. You have what they started from. You have the tools for coding programs.
For goodness’ sake, stop mewling and puking that Apple baked a cake but won’t give it away to you so that you can slather on the icing of your choice. Apple can’t stop you from taking the same the eggs, flour, milk and sugar they started with. If you can bake a better OS with the same ingredients, then grab the cocoa (pun intended) and get to it.
“Linux has been a decade long beta, and nobody ever conquered the world in
beta. ”
Alexander the Great did.
One of the reasons linux hasn made as much strides on the desktop is because kde and gnome have only been around for 4 years. I believe linux early focus was as a server. you really dont need a gui for servers for joe consumer.
Rome wasn’t built in a day, and Microsoft started with MS-DOS.
Because UNIX is server operating system!
Look: http://news.com.com/2100-1001-963285.html?tag=cd_mh
Let’s see…
SUN Solaris – never made big on the desktop.
HP UX – never made big on the desktop.
IBM AIX – never made big on the desktop.
DEC Unix – never made big on the desktop.
SCO Unix – never made big on the desktop.
Linux – never made big on the desktop.
Zenix – never made big on the desktop.
BSD – never made big on the desktop.
Apple OS X – never made big on the desktop.
See the pattern?
ciao
yc
If Linux is so great, then why hasn’t it made it to the desk top?
Because noone cared enough up till recently to bring it there, simple ey?
-fooks
Rome wasn’t built in a day, and Microsoft started with MS-DOS.
Good and simple quote, I like that! 🙂
As has been said before, Linux hasn’t made it to the desktop because it was only recently that anybody cared to move it to the desktop. Windows was around a long time before it made it to the desktop, too. That said, an OS doesn’t have to be a desktop OS to be good. You can make a good desktop OS out of pretty much any kernel out there, what matters is the GUI and all the userspace stuff (which has little to do with the OS, aside from the fact that it depends on certain primitives exposed by the kernel). But it takes a really good OS to handle multi-way NUMA systems and gigabytes per second of I/O from multi-terabyte files. That said, I found the RSX comment to be very funny. People pay more for the RSX because it *performs* better than a focus. The RSX is built for someone who appreciates speed. It’s an insult to the 200 ponies in the RSX to compare it with Apple hardware. Maybe you should have used the New Beetle insead? Or the Mini-Cooper?
Posix doesn’t make it UNIX. bash does. Does Windows have bash? No.
Sure it does. Just go and download it with the whole cygwin package.
Oh. I know. Now you’re saying “it does not come as part of the OS”. Of course it’s not a part of the OS, it’s not in Unix either. It ships with most Linux distros though, and therefore is on Linux as on Windows just another program like any toher.
I really would like to see your Windows Bash scripts
then. Bash under Windows is absolutely not the same as bash under Linux. Try this is a Windows Bash script:
if uname -a | grep Linux>/dev/null
then
echo good
else echo bad
fi
If it returns good then your allright:)
>Posix doesn’t make it UNIX. bash does.
Don’t be silly please. BeOS *IS NOT* a Unix. But it does have Bash and it is somewhat POSIX. In fact, the BeOS kernel was firstly developed with a copy of the XINU book side by side of the developes (XINU means “XINU Is Not Unix”). After that, it took its own, original way of doing things.
>> I only wish the GUI (GTK 2.0) was easier to program.
I haven’t read all comments so this may have already been said somewhere, but if you would like to program using GTK2 in a very easy way you may want to try Mono with Gtk#
I’ve found it to be the easiest solution yet.
—
(please don’t flame because I’m recommending C# – I also hate C#, but I really love Gtk#)
Some people were talking about it (the Torvalds vs. Tannebaum flamewar), anyone who doesn’t know about it can read it here:
http://www.oreilly.com/catalog/opensources/book/appa.html
It’s kinda long and becomes boring really fast.
Anyway the microkernel vs. monolithic kernel debate is old news (a 1992 flamewar), so if anyone wants to talk about exokernels (http://www.pdos.lcs.mit.edu/exo.html) be my guest, it could be the first exokernel flamewar 🙂
It’s Tannenbaum not Tannebaum
Sorry for the typo
Bash under Windows is absolutely not the same as bash under Linux.
Except that it’s compiled from the same sources using the same complier, right?
Try this is a Windows Bash script:
I should rather be concerned if it returned ‘bad’. Bash on Windows is the same as bash on Unix or bash on BeOS. ‘uname’ returns information about the host OS.
Linus’ comments on IA-64 makes me think how knowledgeable he is when it comes to processor design. IA-64 is slow?
IA-64 must be the fastest mass produced processor. Because it (in general “EPIC-explicitly parallel instruction computing” architecture) is designed to leave all complications related to instruction parallelism to the compiler and replace the circuitry required for this with extra computation units. And EPIC brings much better parallelism compared to VLIW by additional hinting. That’s why it’s such a design feat, and it’s one of the most amazing things in processor history. That’s why if IA-64 fails and AMD wins, we would all lose. Because all the desktops will still be using more than a decade old technology.
He could still go as far to say “AMD’s approach is more interesting”, whereas their approach is the least innovative. They simply didn’t do anything interesting.
He is also working for Transmeta which is probably completely unfamiliar with IA-64. And they would probably wither and die much sooner than Linus expects IA-64 to.
He should probably argue how big a design mistake Transmeta is making and how their processors are crawling first before he belittles IA-64.
He also had funny ideas about the development model of OSs( can be reached through http://kerneltrap.org/node.php?id=11 ).
All these make me shudder and love BSDs even more. Their core group is lot more knowledgeable when it comes to computer science concepts.
Eugenia, it was a joke 🙂
Ah, ok.
Bash under Windows is NOT the same as the Unix version
Hence even Solaris bash is different from the Gnu bash.
Bash comes in lot of flavours….bas also
Bash under Windows is NOT the same as the Unix version
Hence even Solaris bash is different from the Gnu bash.
I don’t know what Solaris does, but bash/Windows from Cygwin is GNU bash, according to ‘version’. Care to show me the difference between GNU bash and GNU bash?
Why would someone flame you for recommending C#? so what if it was designed by Microsft, at the end of the day, its out in the wild blue yonder for the OSS community to embrace, so why no encourage its development?
I personally think a large number of people get a little over the top with the whole “Microsoft is more evil than satan” rubbish. Sure, Bill and Ballmer are complete dork features, however, at the end of the day, the people that do the research aren’t the spin doctors or managers, they are the men and women who occupy the several hundred desks at the Microsoft research lab.
You’re right, but I was justing voiding out some _possible_ flames
He (Linus) does know what he is talking about. He works for transmeta, which uses the same technology, VLIW, as what is used in Itanium. He would HAVE TO KNOW how it works as he is helping to write the morphing that runs ontop of VLIW.
As for Itanium, it will NEVER get to the desktop or even high priced servers, heck, you’re more likely to see a honest politician in your life time than an Itanium server. Sales of Itanium are so rare, they have to have a press release just to remind people that the darn thing is still alive.
If Intel want to get more people to adopt it, they should lower the price so that is replaces the Xeon in servers. If they price the same as a Xeon and increased manufuring, they would find more would be sold. I would be VERY surprised whether they have even come close to paying back they money they used for R&D. Personally, they would have been better off improving Alpha, which is built on RISC. A proven and mature chip for both the workstation and server market
As for those idiots who don’t realise, there IS A DIFFERENCE BETWEEN WORKSTATION AND DESKTOP!
Wow, why not you give us some sales statistics for Itanium? Itanium sales is picking up lately, and would continue to do so at a slow pace as more and more software is ported to Itanium. SGI owns a profitable niche with its hardware, and now it is moving to Itanium + Linux, I think sales would pick up. HP have long been pushing Itanium, and it is pushing it more and more, in fact more than Alpha now.
Why do you not write a nice bash script (no like echo helo, but somewhat more complex) under Linux and
then run it on your Windows Cygwin GNU? then write a
HP-Unix bash script and run that under Linux.
I work with HP-Unix, Solaris and Linux on my work all the time and believe me they are different not extrem but enough to make the Bash scripts fail, if its gnu Bash that you run on Windows then a lot may look the same and a lot may work but if scripts become complex problems are starting to rise..
Anyway if you believe that Windows bash suits your needs and/or is the same then Unix versions that fine.
Bas
Where are the sales stats? Why does it cost $15,000 for a low end Itanium workstation when a SUN Dual workstation costs considerably less? I also thought that the big thing people like you go on about is that the itanium is cheaper than a Ultra Sparc of PA-RISC system, WHICH THEY ARE NOT! and that they would be cheaper because of “economies of scale”, yet, again I stress, they still cost a kings ransom.
I’m waiting for the x86-64, why? because it is a realistic goal. It is affordable, and whether Intel like it or not, that will be the deciding bottom line. If Dell says, “oo, this is great + I can get a nice chipset”, they’re (Intel) are going to be toast in the server market. The Xeons won’t be able to compete and their Itaniums will be too expensive.
Conclusion: Intel has dug its own grave.