Live data from Hacker News

FreeBSD the most distributed Linux in the World

blog.milkingthegnu.org

11–15 of 15 posts

Re: FreeBSD the most distributed Linux in the World

#11
post #8

I'm not so sure about this guy's policework. For instance, NetApp ONTAP is decidedly not FreeBSD. Juniper uses FreeBSD as a control platform, but neither the core functions nor the user CLI are FreeBSD. I'm not sure why I care what IronPort uses. From several years commercial reversing experience, my guess is that even if you lump NetBSD in with FreeBSD (NetBSD is the preferred BSD OS for non-X86 BSPs), you still see…

> FreeBSD excels at (due to a better tuned filesystem and VM), for things where raw I/O and compute are the bottleneck, Linux tends to win.

I've noticed that FreeBSD works more responsively in low memory environments. Perhaps that's the VM tuning.

Re: FreeBSD the most distributed Linux in the World

#12

Did you RTFA? BSD isn't Linux. I think the reason it's used in a lot of hardware and OS X is because of the licensing. They don't need to release the source code after they've used other people's work for their own product. I suppose that's OK since the developer agreed to the license but the rapid explosion of Linux is in my opinion indicative that developers prefer the GPL even with its flaws.

BSD gets an edge over proprietary software because if you distribute changes then you don't have to contribute source. GPL gets an edge over proprietary software because if you distribute changes then you must contribute source. Both models have merit and motive. BSD code is distributed widely within proprietary devices and operating systems. GPL code is distributed in a more homogenous form. However, which licence has the largest sum of installed modules?

Re: FreeBSD the most distributed Linux in the World

#13
post #9
post #5

Earlier quoted context omitted.

developers prefer the GPL Speaking as a FreeBSD developer: I prefer the BSD license. :-) In all seriousness, while Apple hasn't set a great example for giving code back to FreeBSD, they have been very good at sending bug reports back -- and really, a report of "it looks like you're doing bazzle wrong on line 234 of foo.c" is often more useful than a patch which fixes the bug but doesn't explain why. Looking at FreeBS…

Take Apple out of the list. And, obviously, the Google SoC. Now strike DARPA, SPAWAR, Secure Computing Corporation and McAfee Research as well --- that's TrustedBSD, a government-funded project that was required to give its code back (the "McAfee Research" there refers to TIS Labs). Now strike Chelsio, Dell, Myricom, probably Pair Networks, and Timing Solutions, which is driver code. Now strike Ethon Technologies, En…

Of course, more companies contribute back to Linux than contribute back to FreeBSD; they're obligated to.

I don't think it's as simple as that. The license has a major influence on the ecosystem around the project, and it is that ecosystem that determines the contributions that flow back to the open source code. The GPL encourages a certain type of ecosystem, and BSD another; that the GPL will automatically yield a healthier ecosystem than BSD is not at all obvious to me.

One of the major differences is that with the BSD license, companies are free to build proprietary products on top of the open source code, which opens up a lot of business models that aren't feasible with the GPL. Take Postgres, for example -- there are many companies that have used Postgres to build a proprietary database of some description. Some of these companies choose to contribute changes back, or to employ Postgres developers full-time, or to hire them as consultants -- mandating that all changes be contributed back would probably mean the companies wouldn't exist in the first place, or wouldn't be using Postgres, which would probably be a net loss for the project. And if a company chooses not to contribute code back? Well, personally that's fine with me, but I can understand why others might disagree.

Re: FreeBSD the most distributed Linux in the World

#14
post #13
post #9

Earlier quoted context omitted.

Take Apple out of the list. And, obviously, the Google SoC. Now strike DARPA, SPAWAR, Secure Computing Corporation and McAfee Research as well --- that's TrustedBSD, a government-funded project that was required to give its code back (the "McAfee Research" there refers to TIS Labs). Now strike Chelsio, Dell, Myricom, probably Pair Networks, and Timing Solutions, which is driver code. Now strike Ethon Technologies, En…

Of course, more companies contribute back to Linux than contribute back to FreeBSD; they're obligated to. I don't think it's as simple as that. The license has a major influence on the ecosystem around the project, and it is that ecosystem that determines the contributions that flow back to the open source code. The GPL encourages a certain type of ecosystem, and BSD another; that the GPL will automatically yield a h…

I'll be glad if you don't make me take the time to do it, but I could argue that many, many more companies have built businesses on Linux than on BSD. A huge number of products are built on Linux now. Linux is a more popular embedded BSP OS than BSD is.

Re: FreeBSD the most distributed Linux in the World

#15
post #14
post #13

Earlier quoted context omitted.

Of course, more companies contribute back to Linux than contribute back to FreeBSD; they're obligated to. I don't think it's as simple as that. The license has a major influence on the ecosystem around the project, and it is that ecosystem that determines the contributions that flow back to the open source code. The GPL encourages a certain type of ecosystem, and BSD another; that the GPL will automatically yield a h…

I'll be glad if you don't make me take the time to do it, but I could argue that many, many more companies have built businesses on Linux than on BSD. A huge number of products are built on Linux now. Linux is a more popular embedded BSP OS than BSD is.

Sure, in the case of Linux vs. FreeBSD, you're probably right (although there are certainly more factors involved than just licensing).

My point is that requiring code to be given back doesn't necessarily yield a healthier ecosystem, or even more code contributions, than a more laissez-faire attitude.

Post reply on HN