You failed politics 101, didn't you ?:-)
In theory, the citizens of democratic countries decide where the line is drawn. This is typically done through elected representatives (the legislature). The justice system role is to interpret those laws and determine sanctions (still according to the limits fixed by the legislature). The executive system (what most people refer to as "government") can, among other roles, ask the legislature to prepare new laws and has to provide means for the justice to do its job (by catching criminals and applying the sanctions). Thankfully governments do not get to draw the line and enforce it. This is a good recipe for a dictature. The practice may be a bit more complex obviously.
ob. RMS: is he acting like a dictator in handling the GPL and the surrounding FSF propaganda ? I have my answer, I'll let you decide for yourself.
BTW I love emacs (gcc a lot less). Respecting RMS coding abilities (and shortcomings) should not lead to believing his rhetoric. In the same way that I can respect ESR technical contributions despite his (insert a derogatory adjective) political positions.
Oracle DBA instructors defined (2 years ago) a big database to be > 1TB.
How many such databases does Postgres manage worldwide ?
As far as fast is concerned, how is Postgres performance affected with > 1000 simultaneous connections ? Last I heard Oracle 8.0 was demonstrated responding under, what, 30k simultaneous connections (or was it 100k?). The 8.1.5 JVM was tested with 10k simultaneous JVM instances (and the 8.1.5 JVM was its 1.0 release and is now _old_).
Heck, if you want pure performance for low levels of service you'd be better off with Sybase or SQL/Server.
I have heard that for one special customer Oracle did sign a contract guaranteeing (sp?) 5 9's availability (less than 5min downtime a year). Would any open-source/cheap DB user even bet on such availability ?
This is the kind of performance you will only find in "enterprise" DBs. Those features are hard to explain and are not check-box items.
Of course it's mostly when you reach those levels of availability/scalability that most other "features" become important. Online rebuilding of indexes is not that important if shutdown your DB every evening for backup (you would typically rebuild at startup then), etc.
The Infiniband 1.0 standard has been published, we may see the first products available by the end of the year (most likely mid to end 2002 with the tech in PCs by 2004 or so).
IB is endorsed by every company in a position to promote such technology (IBM/Intel/HP/Sun/Q/Cisco/MSFT/Oracle/...). Thanks to such backers, IB is almost guaranteed to become prevalent in server rooms in such volumes that will lead the technology down the food chain.
I am betting that IB will deal FC a not so quick, not so painless, death.
The only technology that can stall IB is TCP/IP-based SANs. However IB has been designed to be almost completely handled by hardware, and even taking TCP offloading engines in the picture there is no way SANs will ever be as efficient as IB. Moreover even if IB were to lose the remote disks war it will still be used for local interconnects as a PCI/PCI-X replacement, or for clusters as a fast message passing interconnect (against Myrinet,...)
One thing I am looking forward to is Oracle's SQL/Net running directly on IB with no networking stack, no context switches along send/receives. Mmm, talk about fast response times...
One thing I am wondering though is whether Intel will use IB as their next graphic-card standard post AGP 8x. IMO they would be stupid not too, but IB may be a little late to catch this opportunity.
Moron, learn how to read first. I merely pointed how ridiculous it was to diss OS-X lack of freedom, while OS-X will offer as much freedom in the choice of shells as any decent unix-derived system.
If I were you I'd refrain from posting your opinion on technical subject before you learn the basics of the field.
I have never touched a Linux or BSD system, I do not have OS-X yet, my desktop is actually running MS software. Yet I would probably be able to install a new shell under Linux in download time + 30 seconds. Same or very similar in OS-X I'd bet.
This may not be basic knowledge and 99% of OS-X users will never have to install a new shell. Yet any motivated "geek" should be able to do this. I expect people to try get such knowledge before they offer their opinion on Linux superiority...
The guy makes an unfounded/uninformed criticism of OS X based on the "fact" (among others) that Apple won't let him choose his shell.
If he wants to stick to Linux because he does not know how to install a new shell in a BSD derivative, then his technical comments won't elicit much respect from me.
So when shown wrong you refer to marketing arguments ?
Boy you do not deserve to run a Unix-derived system !
Who said that Apple has to ship 3 gazillions shells for you to have a choice. They will probably ship tcsh and zsh by default. But then what stops you from installing your own ?
Let's see: su, vi/etc/shells (or the equivalent) will do just fine. Because you can too. Pfff
Please forgive me if I am wrong, but isn't Apple not going to allow you to use whatever configuration of GUI or shell that you want?
Others commented about GUI customizations. Why do you even mention shell ? To me "shell" = "sh/bash/bsh/ksh/zsh/csh/tcsh/..". And Apple will not limit your choice here.
Personnaly I'll try to stick to zsh in my OS X environment. I like bash features but zsh is at least as good and I like BSD.
First, MacOS X doesn't treat directories as files, only the GUI does. Proof: you can cd into any application-"fork" in MacOS X (and NeXTstep for that
matter). Perfectly tar-able, perfectly Posix.
Thanks for helping me making my point. As soon as you drop in a shell you lose the bundle POV and only see bundles as directories. As a user this is not what I want 95% of the time.
I want to have Emacs in a bundle, complete with all the share/ files. From the command line I want to be able to cp this "file" around (no -R), I want to have my PATH include the directory containing the bundle, not a bin/ underneath. Only when I need to change the content of this bundle should I have to see this bundle as what it is, i.e. a directory.
But such semantics would not play well with the POSIX view of a file system. Therefore I am not holding my breath and will have to live with a half solution.
This isn't about narrow-mindedness, it's about sanity and interoperability. It's about not making the same mistakes
Microsoft keep doing over and over again. NTFS streams ARE a complete mess. Try to map them sanely into the Unix-world, and you'll see. Try to use tar to
backup an NTFS-volume and see how much you'll preserve...
Why is this considered a problem with the streams concept rather than with tar or Posix FS APIs?
Strict compliance with POSIX brings you only so far (i.e. about as far as today Unix FSes). Systems with richer semantics would solve quite a few problems in the computer world.
I want to see attributed file systems (file encoding using a hierarchical system expanding on mine, file type, file creator, date of last backup, md5 checksum, etc.). I'd also like to see multi-stream files, but slightly less. I wish my Python files would contain source and bytecode (of course with some security features to protect against virii propagation). I want my dynamic optimizer to be able to store its information in the same executable it improves, in a separate stream.
Now you will tell me that FS's that provided such features have not been successful. Sure, guess why: because of the POSIX smallest common denominator tyranny. MacOS forked and type-encoded files may be a nightmare in multi-OS configurations, but I put the blame on Windows and Unix FSes rather than on HFS.
I also realize that most of the multi-stream improvements could come by treating directories as files (e.g. running a python script by exec'ing the directory it is contained in). This is exactly what Apple is doing with Mac OS X application bundles. It doesn't help for scripts and is of limited use until more OSes implement the idea.
Slashdot Mirror
You failed politics 101, didn't you ? :-)
In theory, the citizens of democratic countries decide where the line is drawn. This is typically done through elected representatives (the legislature). The justice system role is to interpret those laws and determine sanctions (still according to the limits fixed by the legislature). The executive system (what most people refer to as "government") can, among other roles, ask the legislature to prepare new laws and has to provide means for the justice to do its job (by catching criminals and applying the sanctions). Thankfully governments do not get to draw the line and enforce it. This is a good recipe for a dictature. The practice may be a bit more complex obviously.
ob. RMS: is he acting like a dictator in handling the GPL and the surrounding FSF propaganda ? I have my answer, I'll let you decide for yourself.
BTW I love emacs (gcc a lot less). Respecting RMS coding abilities (and shortcomings) should not lead to believing his rhetoric. In the same way that I can respect ESR technical contributions despite his (insert a derogatory adjective) political positions.
How many such databases does Postgres manage worldwide ?
As far as fast is concerned, how is Postgres performance affected with > 1000 simultaneous connections ? Last I heard Oracle 8.0 was demonstrated responding under, what, 30k simultaneous connections (or was it 100k?). The 8.1.5 JVM was tested with 10k simultaneous JVM instances (and the 8.1.5 JVM was its 1.0 release and is now _old_).
Heck, if you want pure performance for low levels of service you'd be better off with Sybase or SQL/Server.
I have heard that for one special customer Oracle did sign a contract guaranteeing (sp?) 5 9's availability (less than 5min downtime a year). Would any open-source/cheap DB user even bet on such availability ?
This is the kind of performance you will only find in "enterprise" DBs. Those features are hard to explain and are not check-box items.
Of course it's mostly when you reach those levels of availability/scalability that most other "features" become important. Online rebuilding of indexes is not that important if shutdown your DB every evening for backup (you would typically rebuild at startup then), etc.
The Infiniband 1.0 standard has been published, we may see the first products available by the end of the year (most likely mid to end 2002 with the tech in PCs by 2004 or so).
IB is endorsed by every company in a position to promote such technology (IBM/Intel/HP/Sun/Q/Cisco/MSFT/Oracle/...). Thanks to such backers, IB is almost guaranteed to become prevalent in server rooms in such volumes that will lead the technology down the food chain.
I am betting that IB will deal FC a not so quick, not so painless, death.
The only technology that can stall IB is TCP/IP-based SANs. However IB has been designed to be almost completely handled by hardware, and even taking TCP offloading engines in the picture there is no way SANs will ever be as efficient as IB. Moreover even if IB were to lose the remote disks war it will still be used for local interconnects as a PCI/PCI-X replacement, or for clusters as a fast message passing interconnect (against Myrinet,...)
One thing I am looking forward to is Oracle's SQL/Net running directly on IB with no networking stack, no context switches along send/receives. Mmm, talk about fast response times...
One thing I am wondering though is whether Intel will use IB as their next graphic-card standard post AGP 8x. IMO they would be stupid not too, but IB may be a little late to catch this opportunity.
While I would not bet much on Apple opening the GUI to customizations, I would put my money on them not doing anything to limit the choice of shells.
Now it may not be an excellent idea to change root's shell if some startup script rely on a special default one, but these would be bugs.
Sorry, no Monty Python quote to finish this post :-(
Moron, learn how to read first. I merely pointed how ridiculous it was to diss OS-X lack of freedom, while OS-X will offer as much freedom in the choice of shells as any decent unix-derived system.
If I were you I'd refrain from posting your opinion on technical subject before you learn the basics of the field.
I have never touched a Linux or BSD system, I do not have OS-X yet, my desktop is actually running MS software. Yet I would probably be able to install a new shell under Linux in download time + 30 seconds. Same or very similar in OS-X I'd bet.
This may not be basic knowledge and 99% of OS-X users will never have to install a new shell. Yet any motivated "geek" should be able to do this. I expect people to try get such knowledge before they offer their opinion on Linux superiority...
Again, how is Apple limiting your freedom, mmmm ?
And I do fart in your general direction :-)
The guy makes an unfounded/uninformed criticism of OS X based on the "fact" (among others) that Apple won't let him choose his shell.
If he wants to stick to Linux because he does not know how to install a new shell in a BSD derivative, then his technical comments won't elicit much respect from me.
Boy you do not deserve to run a Unix-derived system !
Who said that Apple has to ship 3 gazillions shells for you to have a choice. They will probably ship tcsh and zsh by default. But then what stops you from installing your own ?
Let's see: su, vi /etc/shells (or the equivalent) will do just fine. Because you can too. Pfff
Others commented about GUI customizations. Why do you even mention shell ? To me "shell" = "sh/bash/bsh/ksh/zsh/csh/tcsh/..". And Apple will not limit your choice here.
Personnaly I'll try to stick to zsh in my OS X environment. I like bash features but zsh is at least as good and I like BSD.
This moron is moderated UP ?!
What is happening to /. ? Fact-free, FUD-full posts get moderated up.
Taken from a darwin-dev post by Strobe Anarkhos: What can Darwin offer that *BSD and Linux can't?
1) Darwin is more organized from top to bottom. From drivers to solving the /etc/ chaos
2) Darwin is a macrokernel with less maintenance. No module recompiling.
3) Bundles bundles bundles
4) Potentially faster than FreeBSD
5) Automatic kext loading
6) NetInfo
7) IOKit driver architecture
8) More flexible BSD-like license
9) Corporate support
10) Mach-O binaries. dyld memory use efficiency. FAT binaries
Thanks for helping me making my point. As soon as you drop in a shell you lose the bundle POV and only see bundles as directories. As a user this is not what I want 95% of the time.
I want to have Emacs in a bundle, complete with all the share/ files. From the command line I want to be able to cp this "file" around (no -R), I want to have my PATH include the directory containing the bundle, not a bin/ underneath.
Only when I need to change the content of this bundle should I have to see this bundle as what it is, i.e. a directory.
But such semantics would not play well with the POSIX view of a file system. Therefore I am not holding my breath and will have to live with a half solution.
Does tar even respect the bundle bit ?
Why is this considered a problem with the streams concept rather than with tar or Posix FS APIs? Strict compliance with POSIX brings you only so far (i.e. about as far as today Unix FSes). Systems with richer semantics would solve quite a few problems in the computer world.
I want to see attributed file systems (file encoding using a hierarchical system expanding on mine, file type, file creator, date of last backup, md5 checksum, etc.). I'd also like to see multi-stream files, but slightly less. I wish my Python files would contain source and bytecode (of course with some security features to protect against virii propagation). I want my dynamic optimizer to be able to store its information in the same executable it improves, in a separate stream.
Now you will tell me that FS's that provided such features have not been successful. Sure, guess why: because of the POSIX smallest common denominator tyranny. MacOS forked and type-encoded files may be a nightmare in multi-OS configurations, but I put the blame on Windows and Unix FSes rather than on HFS.
I also realize that most of the multi-stream improvements could come by treating directories as files (e.g. running a python script by exec'ing the directory it is contained in). This is exactly what Apple is doing with Mac OS X application bundles. It doesn't help for scripts and is of limited use until more OSes implement the idea.