Mac OS X 10.6 Brings Serious Performance Gains

[krazy]

It hasn't effected the PD code in any such way that it can harm it. The original code stays PD. If someone goes and makes their mods and decides not to reveal those mods that is their right. Once in the PD code cannot take a step backwards, unlike licensed code. The only thing it can do it progress forward.

At first glance, your argument has merit, but I think there are several benefits to copyleft licenses such as the GPL which PD does not offer.

A major issue with PD is Embrace, Extend, Extinguish. If you develop a useful piece of software, what is to stop a large corporation from taking your code, creating an extended clone of your software, and using their greater marketing power to render your code effectively obsolete?
This is similar to what Microsoft tried to do with their implementation of Java (they developed their own code so this isn't directly analogous with PD licensing, but there are similarities so please bear with me ).

Sun sued Microsoft for $35 million in 1997, saying Microsoft breached its contract by trying to extend Java so it would work differently, and presumably better, on Windows computers.
...
Sun executives said they were able to use Microsoft as a distribution tool to get Java into the hands of users when the software was in its infancy. Then Sun's lawsuit froze Microsoft's plans to modify Java for its own benefit, and Java became established in the meantime.

http://news.cnet.com/2100-1001-251401.html

Recall that Java was explicitly designed to be "Architecture Neutral and Portable". Had Java been released under a more naive licence, then there would have been nothing stopping MS from using their enormous market power to crush Java's cross-platform interoperability - effectively taking over control of the technology by ensuring their 'version' was most widely used. Surely if Java apps were only usable on Windows, this is a "step backwards". (it's also no coincidence that C# and .NET appeared at about this time)

Thus I think that while PD is obviously the least restrictive license, that is not enought in the "real world": GPL and similar licenses do more to protect the freedom of users and developers.

Also note that Microsoft's proprietary JVM implementation ceased development in 2001 (after they lost the lawsuit and thus the right to use "Java Compatible" trademarks). Hence, due to their licensing, any improvements/modifications they made to Java are now lost forever and so any "steps forward" that were made are also lost.

Of course, it is your choice to license your code however you wish. So as long as you don't mind having someone embrace, extend and then place restrictions your code for their own benefit there's no problem. However I'd argue that society as a whole loses in this case.

[nanonyme]

A major issue with PD is Embrace, Extend, Extinguish. If you develop a useful piece of software, what is to stop a large corporation from taking your code, creating an extended clone of your software, and using their greater marketing power to render your code effectively obsolete

If it's extended in an original way to become better than the original product, should be plenty justifiable... If it's just the same code wrapped up under a different name, I agree.

[nanonyme]

Pointless because macos do not shutdown the 3D-Desktop wile the ogl Benchmark run witout the modificantions.

So you imply OSX would fare even better compared to Linux than nowadays if you disabled the compositing while doing the benchmarks? :3

[kraftman]

Same in reverse: why should the resources of other companies that you don't serve, serve you. Thus we end up in a situation where they keep theirs, we keep ours. It's all about Intellectual Property, no matter whether you're dealing with megacorporations or FSF.

They shouldn't and if I don't want to serve them I choose the GPL or proprietary, but I want to have my code useful by others and only GPL projects (of course it's clear theory, because I don't make good code :P and some projects can be just not interesting and useless).

@Deanjo

Hey, if you want to keep something to yourself then by all means do so. It is you code after all. This can be achieved though GPL or Proprietary licenses (again same shit different pile). If you want to benefit all however you use PD.

Oh yes, that's a point. I should mention I want to other GPL projects eventually benefit from my code not proprietary or others using different licenses. I've got your point now If I want everyone have benefit I would use PD as you mentioned.

[nao.]

Well, most of the performance gains (and drops) of the benchmarks in the article are due to the difference of gcc. gcc in 10.6 creates 64-bit executables by default, but in 10.5 not. So I think the article is almost nonsense in terms of the comparison between 10.6 and 10.5. At least it should be revised to use the same (i386 or x86-64) executables.

Thanks

[Apopas]

I've got your point now If I want everyone have benefit I would use PD as you mentioned.

That's the trick of the case. PD benefits everyone at the begining, from large companies to individuals and simple users. But since everyone can get the code, it's a matter of time to be extended in such a way that the original code to become obsolete since what matters will be the new technology that will be built on the original code. And since PD permits to a large corporation to close the later code, patent it and struggle education and science like this and sell it to the end user (who ofcourse can afford the money) with tons of restrictions then do really everyone benefits?
The idea of PD which leads to real benefits it's very utopian. I don't doubt that if we were elves it would work greatly, but we are humans and UNFORTUNATELY in the real world it produces much much more harm than good. Here if you don't protect your freedom you can easily lose it.

[Apopas]

Thus I think that while PD is obviously the least restrictive license, that is not enought in the "real world": GPL and similar licenses do more to protect the freedom of users and developers.

I absolutely, agree with you, but deanjo simply believes freedom and anarchy is one and the same.

[Pikavee]

Some suggestions

Phoronix: Mac OS X 10.6 Brings Serious Performance Gains

While our focus at Phoronix is on testing hardware under Linux, we remain friendly and interested in other BSD and UNIX operating systems too, including Mac OS X . With the launch of Mac OS X 10.6 "Snow Leopard" we have been particularly interested in it considering the technological advancements that have been made in this update thanks to their large focus on improving the performance of Mac OS X. With that said, we have spent all week working on a grand Mac OS X benchmarking showdown by comparing the performance of the retail build of Mac OS X 10.6.0 to the earlier Mac OS X 10.5.8 through a number of different quantitative tests. We firmly believe that as of right now these are the most detailed desktop performance numbers available concerning Snow Leopard, but we already have more figures on the way. We have performance numbers from not just one Mac computer, but two different setups. Here's to the first 60+ tests we ran!

http://www.phoronix.com/vr.php?view=14155

Just a suggestion for the future: put all the benchmarks on one graph, using a metric that's useful across benchmarks (such as performance increase in percent). Here's an example using some data from this article:



The benchmarks here are sorted by the NVidia performance numbers, which makes it really easy to scan the results and see that three tests had significant regressions, about a third had no little to no change in perfomance, and the rest saw moderate to dramatic performance increases. It's also obvious that except for a few benchmarks (bork, threaded write, threaded read), the NVidia machine doesn't gain significantly more than the Intel one and it actually sees performance regressions in a similar number of cases where there are none for the Intel machine (stream: copy, stream: scale, gzip).

In your follow-up article, you compare OSX with two versions of Ubuntu. Similar suggestions apply; I would suggest using Mac OS X 10.6 as the baseline for the performance deltas in that case, since the point is to compare OSX against Ubuntu.

Frankly, I don't think I could bear to actually read one of these articles if it wasn't for the infinite scroll greasemonkey script[1] that de-paginates them automatically for me. Increasing the amount of information per page (or having less page breaks) would really help this.

[Michael]

Increasing the amount of information per page (or having less page breaks) would really help this.

Then join Phoronix Premium where this is easily done,

[kraftman]

That's the trick of the case. PD benefits everyone at the begining, from large companies to individuals and simple users. But since everyone can get the code, it's a matter of time to be extended in such a way that the original code to become obsolete since what matters will be the new technology that will be built on the original code. And since PD permits to a large corporation to close the later code, patent it and struggle education and science like this and sell it to the end user (who ofcourse can afford the money) with tons of restrictions then do really everyone benefits?
The idea of PD which leads to real benefits it's very utopian. I don't doubt that if we were elves it would work greatly, but we are humans and UNFORTUNATELY in the real world it produces much much more harm than good. Here if you don't protect your freedom you can easily lose it.

Yes, my post about PD was very simplified and not accurate.