BSDForums interviews FreeBSD Release Engineering Team’s Scott Long relating to various aspects of FreeBSD. Topics discussed include FreeBSD general issues, its academic roots, how FreeBSD compares to other BSDs – OpenBSD, NetBSD, and the ongoing debate on FreeBSD vs. Linux. Scott gives us his perspective on the corporate adoption and popularity of FreeBSD. He brings us up to speed on FreeBSD 6.0, its new features and enhancements, including Apple G4 PowerMac, AMD64 and wireless compatibility. Scott also discusses FreeBSD 6.0’s upgrade path and release timetable.
FreeBSD: Interview with Release Engineer Scott Long
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
FreeBSD 6.0-beta indeed supports my nforce3 350 mobo suberbly.And it’s fast.
Does it support Intel 64 bit chips?
I have a Pentium-D dual core thingy?
The CPU isn’t an obstacle.Depends from my experience more on the motherboards chipset.Like Scott said;a lot more chipsets are supported.You might want to try FreeBSD 6.0 beta4.
FreeBSD 6.0 Rocks ! The only problem is the lack of graphic driver for the 64bit platform ! Nvidia gives one for the 32bits ! Anyway its really Fast and really Up-to-date !
The whole thing of Linux VS BSD, I really do think BSD is far superior, technically speaking. There are things I don’t like about the Linux kernel, but not really for technical reasons, aside from the file structure, that could be more unified and simplyfied. For those of you that have had a look at and read the source for Linux and any BSD, wouldn’t you agree BSD is better designed and function given what it does and how it seems to be designed to work? When I say about reading the source for Linux, I’m talking just a straight download from kernal.org and look at that, and not the source for any distribution with X window and all the other add ons on top of the Linux kernel.
But s far as FreeBSD, specifically 6.0, I tried beta 2 or 3 awhile ago, it’s now at beta 5, and it defineately did a better job setting my system automatically then 5.4 did, like setting up DNS, hostname, username. I just hope PC-BSD does not put out their official 1.0 release until they switch to the FreeBSD 6.0 kernel.
I’m just glad that at least Scott mentioned that ALL open source OS’s really lack end-user support. I am a very big nay-sayer that to constatly refer people back to message boards, forums, FAQ’s, and mailing lists just does not cut it. That’s why I provide full technical help for anyone’s computer that I install any Linux or BSD on. How are open source operating systems ever gonna get into mainstream desktop and laptops if people have keep searching on the net for answers instead of giving their computer to someone to fix something that broke in their Linux or BSD system. I mean really, how many people are willing to offer full technical support or assistance to say 15 complete strangers to help fix their open source operating system for them, and not try to tell that what to do. Until there is a significnt number of computer techies(as in at least 1 or 2 in any computer repair business) the can repair or fix any propblem with any operating system out there that can be used on any modern computer(the last 4 years), only then will alternative OS’s become more widely used in homes that does not have anyone who’s into computer’s living there. Like if someone is fulltime employed, plus married, plus raising kids, how can they always be searching message boards or forums about problems for answers, that’s what I meant about just letting a techie fix their Linux or BSD system instead of them trying to bother with anything.
Sorry about this long post, but I did so to make a few points, even if you do think it’s only trolling, I’m really not if you give proper careful consideration to what I’m saying and just leave your opinion behind just for a few minutes.
The Internet is a fortress of knowledge and bullshit. Today, many people routinely use the Internet to shop, pay their bills, find their heritage, look for their soul mate, indulge in forbidden fantasies, research, among other unimaginable and imaginable activities. Why then is seeking information about a computer problem regarding Linux or FreeBSD on the InterWeb evil?
Do me a favor. The first thing you should do when you introduce novice users to computers, Linux or FreeBSD is to launch a web browser pointing to Google. Then proceed to tell them the answer to life’s questions is before them. Forgive my hyperbole.
“Why then is seeking information about a computer problem regarding Linux or FreeBSD on the InterWeb evil?”
I’m not the GP, but I agree with his argument. I think we can all agree that the best place to seek information on Linux and/or FreeBSD is the Internet. Even for the biggest of businesses, the Internet is the best place to go to see what’s out there. In terms of growing the free software movement, the Internet is how we raise awareness and acquire users.
However, for a large portion of the market, the Internet is not a viable solution for KEEPING users. When businesses encounter a problem, they want immediate support. When computer-challenged home users have problems, they want to know who they can call to get help. The existing resources don’t address these demands.
Free software can acquire business customers and home users with the Internet, but it can’t keep them happy that way. They need a number that they can call to get help. They need be able buy a support contract that guarantees support on demand or be able to pay a reasonable fee for each support incident.
Most people prefer crappy software with decent support to decent software with crappy support.
“Most people prefer crappy software with decent support to decent software with crappy support.”
I’ve been thinking about this post for a while, and while I was innitially inclined to agree with it, I’ve since changed my mind. I think it’s more accurate to say that most people expect software to be crappy, and therefore expect that the vendor will support the mess that they’ve created.
Furthermore, I am certain that the vendors themselves prefer the “ship product early, support it later” rather than getting the product done properly so that support would be far less of an issue. I find that the quality of damned near anything more complicated than a hammer to be lacking these days, regardless of who makes them, or where the production takes place.
In general, I’ve found that OSS is less crappy than most proprietary software I’ve used, with the notable exceptions of DEs (both KDE and GNOME seem brittle and flaky to me), and most Linux distributions I’ve tried, and as such, it’s less of an issue to have to support something myself.
In all, it’s just a miserable situation that we’re all in that quality is rarely the first priority, and for this, we all suffer.
In closing, FreeBSD is nice, NetBSD is interesting, DragonFly has a lot of potential, and although not the highest performer, I’ve never had a stock OpenBSD install crash on me on either x86 or PPC hardware. But there’s certainly much room for improvement.
Free software is created by developers for developers.
We release the source code so others can improve the software so we have the use of better software.
It’s quite a selfish thing.
Users serve no use but to file bug reports.
If a user doesn’t understand enough about the software they are using to file a bug then they are useless to the community.
If a user must go to the effort of learning about their software to take advantage of free software instead of running to tech support everytime something goes wrong then they will also learn enough to file bugs.
A lot of users refuse to learn. They are of no use to the free software community and would therefore be leeching.
Learning is good for everyone.
– Jesse McNelis
I just want to add to this (agree completley) that I didn’t migrate to Linux (the same statement applies) because I thogut that it would be or will be a great PnP desktop appliance aimed at my mum and dad.
I went linux because I liked the parts that where aimed at _me_ a geek and a hoppy developer.
I would thus be happier if future “usabilliy studies” and HIG’s where based on the assumption that the majority of users are tech savvy geeks and power users.
Ok, this was OT, sorry. Carry on discussing FreeBSD. Any more thoughts on GNU vs. BSD (as oposed to the ever boring Linux)?
…hmm, I should use the preview feature in the future. I’ll blame this one on no sleep.
Fair enough, though this doesn’t apply to all projects, some projects like Firefox are capable of being user-focused.
Users serve no use but to file bug reports.
Well, with such an attitude you’ll never be able to reach wide acceptance and wide acceptance is what you need if you want BSD to be truelly useable (even for power users, think of things like drivers for BSD, plugins etc).
There really isn’t anything related to userspace components in this interview, which was interesting to me. In my opinion, the answer to the “what makes FreeBSD different from Linux?” question must be (first and foremost) that FreeBSD provides a unified kernelspace/userspace stack at the core release engineering level. While Linux distributions also provide a complete OS stack, they respond to the releases of the kernel, glibc, etc., rather than driving these releases. When Linux 2.6.0 was released, there was no distribution releasing along with it, most of them migrated over time. In this sense, the new kernel series was a disruptive technology for the distributions.
Instead of highlighting the ways in which FreeBSD 6.0 will leverage the improvements under the hood, Scott mentions the under-the-hood improvements themselves as a reason to upgrade. It reads like a Linux kernel release notes. It seems to me that FreeBSD doesn’t get the media attention it deserves because there’s nothing glamorous about the kernel guts. The journalists writing about Linux are talking about Ubuntu and how awesome GNOME 2.12 is and how easy it is to manage packages. If FreeBSD wants to create excitement and develop the evangelism that drives Linux growth, then they need to highlight the ways in which users (whether they be desktop users or server admins) will see improvements to the usability, performance, or reliability of their machines. Otherwise it is just vague technical jargon that separates FreeBSD from Linux or any other platform.
My point is that FreeBSD is a platform, and the project, especially the head of release engineering, needs to start talking about FreeBSD as a platform. Not as a kernel, not as a fine example of quality software development, but as a platform on which one can be productive or run a business. How does FreeBSD 6.0 enable me to be more productive? Well, Scott says I can get double the throughput when writing to a RAID array with dd. This is a good start, I guess, but some information on how awesome FreeBSD 6.0 is as an AMP (apache, mysql, php) platform would probably get more attention from the media.
yay, rc1 at the end of the week
is nf4 supported as well? which components of it?
The support economy is practically built into the free software platform, and still there are no real takers. I’m talking independent support providers competing to provide services to free software users. We’re beginning to see a little of this from Red Hat and the like, but just to make their own platform viable in the marketplace. Projects like FreeBSD, Debian, etc. can’t and won’t provide premium support service on their own.
Someone should create a startup to sell support services for some of these platforms. Solicit some venture capital, make some splash in the media, and run with it. Whoever first manages to do this will get very rich.
i keep core dumping with mencoder and ffmpeg on linux. has anyone tried encoding vobs on freebsd?
Wrong question … mencoder and ffmpeg (why do you need ffmpeg separately btw? ffmpeg is part of mplayer/mencoder, and it is always up to date) coredumping might have a variety of reasons: bad packaging of your particular distro, buggy (or bad packaging) of supporting libraries, etc… In other words, it is very unlikely that mencoder is buggy on linux in general and works on FreeBSD (in general). It would be rather silly to switch to FreeBSD if mencoder is buggy on your particular linux distribution.
On the other hand, yeah, I use it fairly regurarly on FreeBSD, encoding into a variety of formats: mpeg-4 – either ffmpeg or xvid -, h.264 (with x264) and occasionally mpeg2 – and it works fine, with my encoder settings. Which reminds me: there are so many options to mencoder, that it is easy to find one that won’t work, in fact if I remember correctly, there are certain settings that would trigger a coredump mentioned in the man page as well. Also, vob is just a container of mpeg2 streams, so you might have problem with mpeg2 decoding.
One good thing right now about FreeBSD and encoding is that they always have the latest x264 in ports (just checked, I currently have x264-0.0.20051004 installed). Which is a good thing for such a new and fast improving codec
– but if you can run ./configure && make && make install on linux, then this is not a good reason to switch either.
and that’s exactly why i think freebsd would be a better choice for me because the entire platform is very well coordinated and tested. that’s not to say that linux doesn’t have very well coordinated and tested distributions, but i use slackware 10.1 with a 2.6 kernel and probably the latest gcc and libc, libc++ etc libraries. mencoder’s ffmpeg option always fails. so i tried to use just ffmpeg libraries and that core dumped as well.
on my mac machine, ffmpegx, which wraps ffmpeg and mencoder, seems to work just fine on these same vob files.
you’re probably right, i do have a mismatch in libraries probably.
what kind of encoding times are you getting when going to say mpeg4 using mencoder on say a 2 hour movie and 2 two encoding?
“what kind of encoding times are you getting when going to say mpeg4 using mencoder on say a 2 hour movie and 2 two encoding?”
oh, there are so many options that can slow down encoding… you can’t calculate time based solely on the codec used and the lenght of the movie. I don’t have any idea how fast it would go without any options. But: apply a light denoise filter (hqdn3d=2:1:2) and encoding speed decreses by ~10%. use v4mv (4 motion vectors) which decreases speed by another 10. trellis quantization, mbd=2, etc – all will decrease speed. Now on my sempron 3100+, with almost ALL the cpu intensive options turned on, doing bicubic sofware scaling and all, I get 8-10 frames/sec. If I turn those options off, I’d probably get 25-30 or even more. I know this because I can rip tv programmes directly into mpeg-4 (which needs at least 25 fps – real time encoding) using mencoder, with some options (v4mv, deblocking filters) turned on. Check out this script: ftp://hatvani.unideb.hu/pub/FreeBSD/ADDONS/docs/TV_ENCODE/rip.sh
If you have further questions email me at molnarcsATgmail.com – I don’t want to drive this further off topic… (I might claim that the above script is a good template for encoding from TV on FreeBSD though
))
Kudos to Scott and the FreeBSD team for their hard work and dedication on the 6.0 release. I look forward to it.
– j
Great interview and many thanks to OSNews/BSDForums for making it available! Also, good luck to Scott and the rest of the FreeBSD community and thanks for your dedication and hard work!
It’s great to hear about new features and improvements but I think more discussion is needed in the area of marketing.
“Do you believe it is time to create some core team or other positions to increase FreeBSD awareness?”
“A “marketing team” was created several months ago specifically to address the need to get more involved with companies that use FreeBSD and interact more with the news media. They welcome comments from users, developers, and the public at [email protected].”
I think this coupled with the new BSD Certification effort is an important and needed step that will help raise awareness and gain Corporate mindshare.
Personally speaking, I’d wage that a large percentage of businesses currently using FreeBSD or any other *BSD flavor are using them without the knowledge of management. I think most managers are oblivious to what *BSD is and are only vaguely familiar with Linux. I’d also guess that most are unfamiliar with the whole concept of Open Source.
There needs to be an effort to reach/educate these people before FreeBSD or any other *BSD will see a higher level of adoption in the corporate arena.
BSD is such a great set of OS’es i prefer them over Linux . I have heard rumblings here where i work about using linux more for corporate tasks. something i am against with out a clear plan and idea of where the cost benifit will be. It was even suggested that we run Linux on our Sparc based servers we have <rolls eyes> (okay so i work with some people that are out of touch) Anyway i would push BSD but its hard to translate superb qualities of BSD and advantages over Linux when they can’t even understand how retarded it would be to run Linux on Sparc based SUN hardware especially when companies like RedHat package Linux so nicely.
The FreeBSD new logo contest was closed some three months ago. When shall we discover the chosen ? Will it coincide with the release of 6.0 ?
Of course, Hi Eugenia!
– First all BSD’s are NOT free Software and are not Open Source.
There is no doubt anymore that BSD dont ship at all anymore , its a small minority that is really shrinking and being replace by GNU/Linux on the few webserver that still use it. It as no presence at all on supercomputers , and as no presence at all on desktop beside on some fan desktop.
He make the higly laughable comparaison of comparing FreeBSD support with support from Red Hat and SuSe two of the major company , but not the only company unlike BSD’s , who offer GNU/Linux and does so from companies that are worth around 3 billion and compare them , when they do not at all.
He make the false analogy that because GNU/Linux is gaining market share and visibility and considered a threat by Microsoft that BSD’s does the same , its not the case at ALL. He offer MAC OS X wich is not BSD , not Open Source and not free software at all , the Biggest traitor to the BSD community , he also hang to the claim that its the #2 desktop , with its 5% market , problem is that GNU/Linux today is at 12% with all the numerous Asian , African , Australian , South American distribution and languages that GNU/Linux is translated in and used by the people in those more populous region that BSD dont reach , because they know what Thief , liar and coward are in those area.
FreeBSD already as smalls company ( all failing or on there way there ) and the biggest lying marketing , its failing because the product dont deliver at all , everyone know after trying it that its a piece of crap that is not up to par at all.
The Only thing that whas supporting corporate sponsoship BSD’s whas Darpa and Darpa pulled its funding recently.
He diverted money from GNU/Linux too he whas paid to do driver for GNU/Linux and Illegally diverted the code and research progress to BSD’s projects. ( Read #8 )
He consider FreeBSD 6.0 a success , before its release , before anyone beside the fan decide to go and download it or buy it to support the project. He is delusionnal.
He dont see that after 30 Years BSD’S have absolutely no real company supporting it , He havent known or seen all the BSD’s developper who all switched to GNU/Linux or Apple ( Apple use BSD’s but DONT contribute back ). GNU/Linux whas created by the Academics and students around the world because its Open Source and Free software. It also as company who do contribute.
The biggest secret of 6.0 is that it can be installed on older hardware , wow !!
Why is nobody using FreeBSD or running to get it when it just work ? Because it dont …
The best example for why BSD’s aint Open Source is APPLE vs all the other BSD’s and Red Hat vs All the other GNU/Linux distribution. There are many project based of Red hat directly , there are none for Apple.
BSD’s are NOT free Software and are not Open Source. Its time the BSD’s fixed that problem to begin with , had they done it in the past there would have been no need for GNU/Linux at all.
I truly love FreeBSD. The simplicity and elegance of the configuration files is unmatched. The stability and speed is breathtaking. FreeBSD is a beautiful system. My RedHat box takes over a minute to boot, FreeBSD does it in less than half the time. While it might not be the most advanced OS on the planet, it definately can hold it’s own and is definately the most innovative OS. FreeBSD leads, others follow. In fact, I have a suspicion as Linux’s SMP reaches it’s limit because of it’s simple design, a lot of lifting and hard work will be taken away from the BSD implementation.
That said, I used to advertise my BSD skills on my resume. I no longer do that anymore. There are a few jobs out there dealing with BSD but I don’t want them because I don’t know where FreeBSD will be in 5 years. I love FreeBSD, but my bookshelf is full of books covering routers, switches, Linux, Windows, Exchange, SQL Server, perl, dns, postfix, etc. The few books I had about FreeBSD are now in the attic. After years working in the computer industry, I have come to the realization that the superior solution hasn’t always won because of it’s superiority. Even though Linux is a lot dirtier, and Windows is a complete mess, I’ve never had problems getting either one to get the job done. And if companies want skill sets in those areas, I have to comply. I am not in a position to go against the grain.
I really do wish the FreeBSD project luck. I do hope that one day they own a major part of the market. When that day comes, I will be exstatic to get back into FreeBSD. Linux, Windows, Cisco are my source of income, FreeBSD is a source of happiness.
“…I don’t know where FreeBSD will be in 5 years.”
If I may….
Multics was supposed to kill BSD
VMS was supposed to kill BSD
Netware was supposed to kill BSD
SysV was supposed to kill BSD
Mach was supposed to kill BSD
Windows was supposed to kill BSD
I’ll see your Linux and raise you 30 years of BSD.
please, take this as a lesson in English:
it’s = it is
if you want to show possession then the word is:
its
definately is not a word. definitely is a word.
You sound weird, you loose!
I’m running FBSD 6.0BETA5 and it rocks! It seems much more stable than 5.x and a bit faster as well (of course, I have debugging turned off for obvious reasons). It is the best desktop OS (other than Mac’s OSX) that I’ve found and I’ve tried almost everything. Speed, reliability and ease of use makes it my desktop OS of choice.