Whoa, don't blow my mind quite so hard. I'm not sure I can handle all this wisdom at the same time.

C'mon, you think I execute shell commands by writing a C program that calls fork(), exec(), and pipe()? You think I write web pages pixel by pixel? Obviously high-level languages and programming paradigms are appropriate in many cases.

I'm sticking it to the Java weenies who think that C and C++ are obsolete. The people who year after year say that *now* Java is "often as fast as C++ and sometimes faster." The people who still won't acknowledge that there is a real reason C and C++ are still the languages of OS kernels.

It's not premature optimization to write libavcodec in C. Likewise with OS kernels, virtual machines, rendering engines, DSP plugins, and many other applications where the code will almost certainly be on the critical path of a resource-intensive application. It's not premature optimization to use manual memory management in applications that need to move lots of data around with low latency.

Chrome got me where it matters most - raw speed.

This is why I never understand how people can say "sure, maybe Java/garbage-collection/50mb-binaries/etc. are a little slower, but computers are SO FAST these days and programmer productivity is SO much more important. Hardware is cheap, programmers are expensive." etc.

Speed still matters! And it always will.

Applications can store all of their data files and resources on your SD card. Many do already. It's just the executable code that has to be installed locally.

The stuff in the linked articles is nothing, read this: The Injustice Collector: Is James Joyce's grandson suppressing scholarship?

Stephen Joyce to a James Joyce scholar he disagreed with: "You should consider a new career as a garbage collector in New York City, because you'll never quote a Joyce text again."

> No, no, no. The GPL is an agreement between the author and another
> person. The GPL does not prohibit the author from making any other
> agreement she wants with any other person.

I'm not sure I'm reading GPL the same way you are. To me it seems like a contract between code producer(s) and consumer(s). AFAIK the law doesn't really make much difference if the two are the same, though that's a rather rare case (the notable case happening a few months ago when Wells Fargo sued itself).

But that's academic. As you noted, as long as you are the only one creating and using a library, you don't really need a license. But if you release a library as GPL, then accept somebody else's patch that patched code is GPL. And you can't use it in your own non-GPL product.

> You don't want to license your code under the GPL, because it would
> force the recipient to abide by the terms of the GPL? Okay, I'll grant
> that one.

Phew, thanks. :)

But seriously, yes, that's my main point here. I don't like when GPL is sprung on me, so I don't spring it on other people.

> That's not "embrace, extend, extinguish."
> ...
> In EEE, you (E1) take a popular protocol, one that allows several
> products to interoperate happily. You release your own product using
> that protocol. Next, when your market share is great enough, you add
> undocumented "features" (E2) that make your tools more useful, while
> causing competing products to go "WTF?" Finally, you hope, people
> start using your product exclusively (E3), in order to ensure that
> everything works.

For me that's no different than what FireFox did to IE with the, for example, Ctrl-K shortcut that takes you to the search bar. Only we are talking mindshare, not protocols, but IMO that binds even stronger.

(Side note: when I tried to use IE8 for a few days I got frustrated because Ctrl-K doesn't do the same thing as in FireFox. I found it amusing that Microsoft got their tactics turned on them and made the same mistake WordPerfect and Lotus made in not emulating the better sides and UI of their competitor but instead decided to create their own standards. Good luck with that.)

> Microsoft did it with Kerberos, they did it with ActiveX,
> and they're even now trying to do it with ODF.

ActiveX is the main reason why I think embrace & extend is overplayed here. The playbook is:

1. Use a protocol or something.
2. Become dominant.
3. Extend it.
4. Fuck over all the others that don't know how to reverse engineer it or are too proud to do it.

But #2 is overlooked here, or done with handwaving about monopolies. And while Microsoft's distribution channel used to dominate before the age of the Internet (and still does for operating systems), it can't make an inferior product dominant. ActiveX clearly demonstrates that.

Also, if you remember the late nineties, the Java guys and the Netscape guys were all saying that Microsoft is dead because the web is the new OS (whatever that means). So what do you think Microsoft should have done? Suck Sun's and Netscape's dick or fight back?

> Because the GPL would be a good defense against them storming in
> and wiping out your entire niche? MS has done it several times,
> and tried it a dozen more.

Dude. I run a small software shop. I'm not on Microsoft's radar. So you are saying that out of some fear of them I should screw all other programmers like me who are trying to make a living selling software?

> But you should at least understand what the GPL is before you decide
> whether to use it (or, more to the point, before you go on a public forum
> and spread misinformation about the GPL in the course of explaining why
> you won't use it)

I think I did understand it. Possibly I'm wrong. Your scolding above seems to indicate that you hold no such reservations. Funny, that.

Dejan

> You're saying that you wrote a bunch of libraries yourself,
> included those libraries in your own commercial products,
> then released them under BSD because releasing them under
> the GPL would have forced the GPL-ification of the products?

Yes. Ditto for anybody else who wants to use my libraries.

> That's not how it works. You own the copyright, you can use
> the code as you like, including packing it into proprietary
> products. The hoops you're complaining about having to jump
> through are imaginary.

Not really if I'm reading GPL correctly. What I _think_ you are saying is that since I'm the one writing the code and setting the rules for the license I can put out a license saying "this is under GPL for anybody but me" or something to that account, but that's not really GPL but a modification or a dual license.

But more important, people that use my library would effectively be forced to use GPL. As somebody that creates and sells software, that's exactly what I don't like done to me. I never was one of the kids to take their ball and go home because they don't like how the game is played. (And yes, I do send patches back to any libraries I use even when they are not GPL.)

> It's one thing to be okay with others making money off your code.
> It's another thing to be okay with another company pulling an "embrace,
> extend, extinguish" on you

People here use that phrase a lot, but I think it's an oversimplification. While I have seen MS do crazy shit (the original MFC license comes to mind), in the case of the browser, people forget how shitty Netscape was. IE won because it was a better product at that time.

Also, embrace & extend is a pretty good page for any software shop's playbook. For example, adding load & save for competing (closed source) product file formats to Word and Excel was a stroke of brilliance. FireFox, for example, did a similar thing when they supported IE's shortcut keys from day zero, and I applauded that move when switching to it. I wish more open-source GUI products did the same thing when trying to compete with the market leader.

But back to my main point: I don't spend one moment thinking about MS or Apple when releasing a library. Why would I screw all the people trying to make a living selling software just because of two companies?

Dejan

> That's what the LGPL is for.

That's what almost any open source (plus numerous closed-sourced) license except GPL is for. But the topic of this thread is GPL, so I was commenting on that.

Dejan

GPL is good for anybody not making money directly off software products. I don't buy all the ideology around it, but as Linus says it's a cool license because it enforces tit-for-tat.

However, GPL is the kiss of death for anybody trying to make money selling software products. If you have a software product and publish any of its libraries as GPL, then your product must effectively become GPL'ed. And you put hard work into it and want to charge money for that, but anybody can take that product and sell it cheaper or give it away for free.

You can then play games to work around it (spawn the GPL product from a commercial one and talk to it through a pipe or something) but whatever you do is just a kludge in order to dance around the license.

Personally, I gave away the few small, well-rounded libraries I made under the BSD license. I don't really mind if somebody takes them and uses them to build a product they'll be making money off. The knee-jerk reaction here is that when somebody says "commercial software" people imagine big dominant companies like Apple or Microsoft, but the number of programmers working there is dwarfed by the number of small 1-5 programmer shops trying to make a living.

In fact, I don't even mind if a programmer at Microsoft takes my source code and uses it in a product. I met a few of them and they are mostly nice folks trying to make the best software they can. If Microsoft shareholders profit to an infinitesimal amount from something I gave away for free, I don't really give a fuck.

Dejan

NX is great for programmers. When I'm coding, I have 30 emacs windows open, 10 terminals, 50 tabs in Firefox, etc., spread across several virtual desktops. It takes a while to recreate that state, especially when it's not just the windows, but a lot of useful stuff in each one. With NX, I can grab my laptop for a while and within 5 seconds I'm accessing the same desktop, with no perceptible latency. I just need to scroll a bit since I'm viewing a large desktop on a small screen - but it's much better to scroll a bit and continue coding where I left off, than to recreate my state on my laptop. Then later I close my laptop, go back to my desktop and now I have a nice big monitor again. It's a seamless transition.

Firefox is already multithreaded (if it weren't the UI would freeze during downloading, rendering, etc).

It does freeze during rendering and downloading. Simple way to reproduce: navigate to a 100MB file and see what happens to the UI.

FLAC is years older than Apple Lossless. FLAC was sitting out there for the taking when Apple decided to implement their own.

Your "abstract solutions" all take orders of magnitude more memory than C, and still suffer garbage collection pauses.

Why isn't your web browser written using these "abstract solutions"? The answer is the same reason that having real machine code on the client is a win, for those who want to go to the trouble.

Windows thread creation costs more than Unix thread creation because it does more. Whether that work is useful to most people is somewhat dubious.

Windows kernel is roughly based on VMS, which at the minimum has a different security model than Unix. The one in Windows is finely-grained, while the Unix one is fairly coarse.*

In addition, a bunch of things in Windows have thread affinity and that has to be set up too. The concept of thread affinity for things like windows is pretty good for a desktop OS, fairly lousy for a server one.

Dejan

* Windows security model is more powerful than Unix's user/group/world one (ask any large corporation admin), but comes at a significant performance and complexity price. I can teach any programmer the Unix security model in less than a minute, but I know very few Windows programmers that know anything about Windows ACL/SID/Token APIs. (Yes, ma, that's that last parameter in all those calls that you always set to NULL to inherit from the thread settings.)

A couple addenda to my last post:

• I said BigTable is "the" data storage system we use. I should have said "a" data storage system we use -- of course it is not the only one.
• I work for Google, but of course to not speak for them.