dont bother calling your MEP: the european parliament is virtually powerless. real decisions are made in the council of ministers, hence national parliaments should be contacted.
If the EP lets through the text that the Council of ministers approved in its first reading, it's over. Then that text will immediately become the directive.
a propos: the new constitution would have remedied that situation by awarding substantial powers to the european parliament.
That's wrong. It just would have brought more law making initiatives under co-decision (which is the procedure in which the EP has the most power of all procedures), but this directive is already being treated under co-decision. So it wouldn't have changed anything.
No, not many small companies have done that. The BSA (yes, the Business Software Allience) recently ordered a study trying to prove your point, but it resulted in proving exactly the reverse.
'Technical = natural science" (which is equivalent with forces of nature) on its own is not enough, you need a lot more than that. Article 2b states that by definition everything which is new is technical. Article 3 then explicitly states that all new features (technical or not) can be used to fulfill the inventive step.
This even includes business techniques and things like that, which is even worse than the EPO (they at least require you to talk about some "technical effects" of the implementation of your business method, and though this is easily fulfilled, that principle in theory at least allows for backtracking to a more sane situation).
The strongest proponents of software patents until now are the representatives of the various national governments, and the strongest opponents until now has been the European Parliament (save for the Legal Affairs committee, both in 2003 and now).
Cell processors are in-order and therefore quite unsuited for general workstations and notebooks (unless all you do all day is performing matrix operations).
There are several other binary translation projects which have nothing to do with Transitive. You might have heard of Valgrind. Others are for example Walkabout (also with a host of derivatives) and DIOTA.
I did say in my original reply that the numbers were abysmal and that I wonder about the cause. My remark was that the reason you mentioned ("it's Mach at the core, so of course it's slow") is wrong (I left out completely, hope you can sleep tight now). Peace.
I claim your deduction is completely wrong, yes, because it's based on the false premise that Mach is inherently slow. This premise is wrong because completely disregards the fact that xnu doesn't use a message-passing implementation of Mach (but plain function calls) and thus its performance is not inherently worse than that of a traditional monolithic kernel.
They're not "running a BSD kernel on top of Mach", the BSD and Mach both run together in the same address space as one process. They're also not using userland threading, as that article suggested. That said, it did show some abysmal performance, and I'm really curious about what the cause is. It's probably nothing those people came up with though, as they apparently have no clue at all about the underpinnings of Mac OS X.
OSX hacks a BSD kernel into a Mach microkernel, and thus performance is nearly as bad as Mach despite the existence of the mature, standardized interfaces of a BSD.
This is completely wrong. The traditional bad performance from Mach is due to the fact that pretty much every "kernel"service is a separate process which runs in user space and all those processes have to communicate via messaging. On Mac OS X, the whole kernel (including the BSD personality) runs in one address space and pretty much everything is just function calls like in monolithic kernels.
MacOSX is not about performance.
It seems do not read the darwin-dev and darwin-kernel Apple mailing lists.
That said, I have no idea what caused the abysmal MySQL and Apache performance. Maybe MySQL is using F_FULLSYNC? Or maybe there is indeed somewhere a big inefficiency in the kernel, I don't know.
You're right it's under a heatsink, but for some reason I've been convinced for years it was an Intel chip, I really thought I had read that somewhere. Thanks for setting the record straight.
LinuxInsider also has a pretty strange track record. I've only followed it on the topic of software patents, but at least there they are only publishing pro-swpat lawyer opinion pieces without giving any room for rebuttals. See the collected documentation on the FFII wiki
Where they afraid that Borland would end up with monopoly control of the software industry? That Delphi would overpower.NET? Was the cost of buying Delphi higher than the development costs of Free Pascal?
The original author (Florian Klaempfl) started working on it because he wanted a 32 bit Dos Turbo Pascal compiler, and it was clear Borland was not going to write it. I started hacking on the assembler optimizer because I liked diddling with assembler code and programming in Pascal. Later I did a large part of the port to PowerPC and finished the Mac OS X port, because I like to program on my Mac as well.
It's just a hobby project for all of us, not something we feel we have to do to change the world or so. Your statement that "the computing community would be better off" if we had put efforts into something else is strange. It assumes that we would have actually made another huge project like this. Additionally, many people do consider it quite useful (otherwise we wouldn't have 20+GB downloads per day I guess).
The BSD and Mach personalities run together in one and the same (kernel) address space. The BSD layer does not consists of merely some user space libraries. See e.g. this graphic.
Slashdot Mirror
No, not many small companies have done that. The BSA (yes, the Business Software Allience) recently ordered a study trying to prove your point, but it resulted in proving exactly the reverse.
'Technical = natural science" (which is equivalent with forces of nature) on its own is not enough, you need a lot more than that. Article 2b states that by definition everything which is new is technical. Article 3 then explicitly states that all new features (technical or not) can be used to fulfill the inventive step.
This even includes business techniques and things like that, which is even worse than the EPO (they at least require you to talk about some "technical effects" of the implementation of your business method, and though this is easily fulfilled, that principle in theory at least allows for backtracking to a more sane situation).
The strongest proponents of software patents until now are the representatives of the various national governments, and the strongest opponents until now has been the European Parliament (save for the Legal Affairs committee, both in 2003 and now).
Cell processors are in-order and therefore quite unsuited for general workstations and notebooks (unless all you do all day is performing matrix operations).
I doubt you'd be able to run it on Qemu, since even Darwin/w86 doesn't boot on Qemu.
There are several other binary translation projects which have nothing to do with Transitive. You might have heard of Valgrind. Others are for example Walkabout (also with a host of derivatives) and DIOTA.
I did say in my original reply that the numbers were abysmal and that I wonder about the cause. My remark was that the reason you mentioned ("it's Mach at the core, so of course it's slow") is wrong (I left out completely, hope you can sleep tight now). Peace.
I thought he meant OpenFirmware (the Macintels won't use OF)
I claim your deduction is completely wrong, yes, because it's based on the false premise that Mach is inherently slow. This premise is wrong because completely disregards the fact that xnu doesn't use a message-passing implementation of Mach (but plain function calls) and thus its performance is not inherently worse than that of a traditional monolithic kernel.
I doubt it had much to do with revenge.
They're not "running a BSD kernel on top of Mach", the BSD and Mach both run together in the same address space as one process. They're also not using userland threading, as that article suggested. That said, it did show some abysmal performance, and I'm really curious about what the cause is. It's probably nothing those people came up with though, as they apparently have no clue at all about the underpinnings of Mac OS X.
They only ever managed to achieve a speedup on HP processors. They later ported Dynamo to x86 and did not manage to get any speedup at all.
Yeah, they say that all the time. And when you then go to work for them based on that misinformation, you figure out the truth pretty quickly.
You're right it's under a heatsink, but for some reason I've been convinced for years it was an Intel chip, I really thought I had read that somewhere. Thanks for setting the record straight.
No, it's a chip soldered on the motherboard.
The gigabit ethernet chip in my old G4/400 in fact is an Intel chip.
No, you don't
LinuxInsider also has a pretty strange track record. I've only followed it on the topic of software patents, but at least there they are only publishing pro-swpat lawyer opinion pieces without giving any room for rebuttals. See the collected documentation on the FFII wiki
No, Adobe patented it quite a long time ago already. E.g. US 5,546,528 and EP689133.
The original author (Florian Klaempfl) started working on it because he wanted a 32 bit Dos Turbo Pascal compiler, and it was clear Borland was not going to write it. I started hacking on the assembler optimizer because I liked diddling with assembler code and programming in Pascal. Later I did a large part of the port to PowerPC and finished the Mac OS X port, because I like to program on my Mac as well.
It's just a hobby project for all of us, not something we feel we have to do to change the world or so. Your statement that "the computing community would be better off" if we had put efforts into something else is strange. It assumes that we would have actually made another huge project like this. Additionally, many people do consider it quite useful (otherwise we wouldn't have 20+GB downloads per day I guess).
Ask and ye shall receive.
The BSD and Mach personalities run together in one and the same (kernel) address space. The BSD layer does not consists of merely some user space libraries. See e.g. this graphic.