BSDForums Interviews Scott Long
Dan writes that BSDForums is featuring and interview with FreeBSD's Scott Long. Scott fills us in on some of the new things in FreeBSD 6.0 including Apple G4 PowerMac, AMD64, and wireless compatibility. In addition to specifics Scott also abstracts on the overall snapshot of BSD development with respect to OpenBSD, NetBSD and the ongoing debate between BSD vs. Linux.
There's a debate between Linux and BSD? Hmm, I must have missed it.
Finance tutorials and more! Understandfinance
Scott Long is dying!!!
Speaking of which, whatever happened to all the "FreeBSD is Dying!" drumbeats in the last year or two? I haven't kept up with FreeBSD/NetBSD news, but it would seem some series of events has really turned people around, even though it doesn't seem BSD use is necessarily skyrocketing.
FreeBSD seemed to have some issues around the 5.0 release because of the major features that release brought (and the ensuing nervousness about upgrading). Hopefully 6.0 won't be plagued by these kinds of issues and should be taken up rapidly. I've had nothing but good experiences of FreeBSD in server environments, and the fact that increasing out of the box hardware support is being included for desktop platforms is great.
Business Voyeur
Very few things regarding computers die quickly (with exceptions for the HST standard and some other things that were 'hot' for awhile but disappeared) ... another example, Netware has been passing away slowly for years... if FreeBSD is going to be dying as well, it just takes time.
... [depending on your meanings on those two words]... ;)
Hell, COBOL is still "alive" and "well"
MoM++ - A Classic Expanded - [Master of Magic 1.5]
http://mompp.sourceforge.net/
The main thing I've noticed is just better 'out of the box' support for hardware in 6.0. I don't have massive requirements, as I'm running FreeBSD 6.0 for my primary server (mail, web, chat, database, file) at home. I didn't need to rebuilt the kernel as I did with 5.2 - but that was to support an older NIC. Basically it 'just works' and I've stuck with GENERIC for the kernel with no issues. I use Ports like they're going out of style, and I haven't had anything break (that I couldn't fix ;))
Anyway, better 'out of the box' support, which would manifest mostly for folks installing 6.0 for a desktop, or someone who has some new(er) RAID or 1G NIC to support. I couldn't be happier, not using Linux for a server anymore, but it's still my Desktop of choice.
fak3r.com
From a TI-99/4A to years of wintel boxes, then Linux, I've now made the move to AMD 64 and freeBSD. I've moved onto freeBSD because somewhere I saw a quip that went...Linux is for people who hate windows; *BSD is for people who love Unix. I see Unix is an OS written by scientist for scientist and, as much as possible, I'd to discover that idea.
First 3 questions said nothing about Linux.
First 3 answers did.
muahahahahhahaha
Anyways my take on the debate: Multiple times I've written some simple software without even CONSIDERING portability, written on my (Linux/BSD/OS-X) box, and when the time came to make it run on a new platform (Linux/BSD/OS-X), it did, with barely any modifications. So who cares, target any UNIX, they all rock, and you probably wont end up permanently stuck to one of them.
Why stick up for big business?
From the interview:
4. Some critics claim that FreeBSD has not been as popular among corporations as Linux, not even close. Microsoft acknowledges Linux as a viable threat but FreeBSD is not even on their radar screen.
Wasn't FreeBSD the only other operating system Microsoft ported C# to? Didn't Hotmail run for a LONG time on FreeBSD? Doesn't Microsoft use BSD code in their operating system?
There's your debate.
People will wonder why this is great. Darwin is not FreeBSD... there are too many things going on with the Mach Kernel. Yes, Fink or DarwinPorts solves some problems but still. I'm a happy fella.
Please explain what the irony is in that. FreeBSD also works in the real-world quite well (Yahoo, Hotmail before MS, not to mention 2.5 million other sites.
While FreeBSD doesn't have the "sexiness" of Linux right now (although I'd argue that is starting to wane a good bit) it is rock solid and dead simple to admin.
The point is, and this is not specific to FreeBSD, that there are some really innovative experiments going in one bsd flavor or another. Dragonfly comes to mind - some new concepts in computer science comes from its development.
Another example (now from FreeBSD) is their support for multiple threading libraries, like libthr (1:1 - like in linux) and libkse (kernel scheduling entities), which implements M:N threading, that on paper is supposed to be a superior impementation, but it hasn't been implemented previously (afaik, SUN wanted to do it in Solaris, but later went with 1:1). See this and this posts by one of their kernel developers. Quote: "Hopefully this makes a good platform for people to do thread research.." Now that is what "academic spirit" is all about.
Need I mention OpenBSD's innovations in the field of security, or NetBSD's in portability (and their driver infrastucture)?
When challenged with the notion that Linux is more popular that FreeBSD, Scott answered:
``There is no denying that Linux is gaining popularity, but that doesn't necessarily mean that the BSD family is not. A great example is Mac OSX. Under the hood it is very much a BSD operating system. That means that BSD is now the second most popular desktop OS, far more popular and widespread than Linux.''
That's utter crap. First of all, the question was about FreeBSD vs. Linux, not *BSD vs. Linux. Secondly, Darwin isn't very much like the other BSDs. Based on a Mach kernel, with lots of GNU and Apple stuff thrown in, and with some dynamic linking weirdness, it differs significantly from other open source Unix-like systems at every level. Darwin is at best the bastard child of BSD, and using its popularity to support the notion that FreeBSD is popular makes it look like FreeBSD isn't popular in its own right, which is far from the case. So why resort to deception?
Please correct me if I got my facts wrong.
I have a love for operating systems (and programming languages, but that's for another time), and I've tried most of the open source operating systems that I'm aware of. These days, I have three favorites: OpenBSD, for being the only OS _really_ serious about security; NetBSD, because I love the portability, minimalist feel, and pkgsrc; and GNU/Linux for generally being up to date (in features and hardware support) and easy to maintain (apt-get is the best). FreeBSD seems to have fallen by the wayside; it's a great system, but to me it seems it doesn't really stand out. Are there any great features unique to FreeBSD that I'm overlooking?
Please correct me if I got my facts wrong.
*wonders out loud* I guess Linus didn't invent Linux in a academic environment at all then.
"1. Partitioning. Why in the fucking hell do you have to partition TWICE? Why do you need a standard partition that can be seen by other OSes and then those weird ass partitions within that partition that only BSD can see? How is that useful? Here's the answer: IT ISN'T. Get into the 21st century and realize that your stupid partitioning scheme should have disappeared with the 'ed' line editor."
/, swap, /tmp, /usr, /usr/home, /var....
Because I may want more than 4 partitions/slices. You know, for
"2. Lack of LVM support"
I guess that's why you have a issue #1.
"3. Serious instability."
Right. Did you send in a bug report? No? "Fucking tragic."
"4. The use of a non-standard make."
I did not know there is an International Standard Make Standards Body.
"5. No Bash by default?"
No, it's not linux.
"This is where Linux is cleaning your cocks."
So, linux DOES suck?
-- I have a private email server in my basement.
It has generally been my experience that the people who argue "BSD versus linux" fall into one of two categories.
1) BSD people who have no in-depth knowledge of linux, and therefore speak from a position of ignorance
2) Linux people who know very little about any particular BSD, and therefore speak from a position of ignorance
The people with truly deep knowledge of both systems always say "use the right tool for the right task" and typically have no time for OS religious wars.
Why is there no "-1 Retarded" ?
I had used FreeBSD excusively in the 4.x's on both my desktop and servers up until about a year ago. I really liked ports and found that everything just seemed like it fit together like one seamless product instead of the hodge-podge that my previous experiences with RedHat and Slackware had been. Subjectively I also found that it felt faster than the Linux at the time of my switch. I stuck with FreeBSD but also with the 4.x tree because I was a bit put off by the whole stable vs development nature they kept putting forward for 5.
That changed when Novell bought SUSE and started offering their certifications. I was asked to evauluate it by some people at work who had fond memories of Novell and wanted to see what they did with Linux and I was given the opportunity to sit for the Novell CLP (Certified Linux Professional) practicum exam if I wanted as a carrot for doing it. I decided that the only way to get comfortable enough with it for the test was to dive in and install it on my primary desktop OS and force myself to use it.
What I found was surprising. There, obviously, were some growing pains when it came to various BSD vs SYSV things and directory layout and ports vs RPM etc. What I was surprised by was that everything worked out of the box. I am used to, and almost looked forward to, having to roll up my sleeves and figure out the config files and recompile the kernel and go through newsgroups and mailing lists for fixes. This has been especially true since my primary machine is a laptop (Dell Inspiron 8600). What also surprised me was that Yast configures, with either a console or X-Win GUI, just about eveything that I wanted to configure and every setting that I wanted to change. I kept waiting to run into a gotcha so I could swear it off and convince myself I had to do it all by hand but it hasn't come yet. The whole magic-black-box aspect of it scares me a little but I am amazed how little I have had to get my hands dirty. It almost feels like Windows Server 2003 -- in a good way. Also, while I was put off by the 6 CD thing at first (I have always had a pretty streamlined and small FreeBSD install for my desktop) I find that having pretty much any piece of software that you might want in RPMs you can trust (and don't suffer from the dependency hell I remember) right on the CDs is actually pretty nice.
After I take my practicum in the next few weeks I am going to try Novell's desktop offering. If it is as slick as SLES then Novell, especially when you figure in NDS and ZenWorks, is going to make huge inroads versus the other distros and FreeBSD. And, strange as it sounds to me who missed the Novell hayday, there are alot of people in the industry who seem to remember their interaction with Novell fondly for whom their name and support seems to be a big plus.
BSD not complying to standards? At least it has some, which is more than can be said for Linux
1. Partitioning. Why in the fucking hell do you have to partition TWICE? Why do you need a standard partition that can be seen by other OSes and then those weird ass partitions within that partition that only BSD can see? How is that useful?
How useful is the partitioning scheme of Linux and Windows? First you have primary partitions, then extended partitions, and then logical partitions. Huh?
Here's the scoop, numbnut: way back in the beginning of harddisks on the PC, the idea was that every OS got its own primary partition. Then the OS could slice up its own partition however it wanted to. Microsoft decided to use extended/logical partitions, while 386BSD decided to use the traditional BSD slices. Then Linux came along and said "let's do it the Microsoft way!" The truth is that neither method is better or worse than the other. In the absence of any standard, each OS is free to subdivide their primary partition however they want.
A Government Is a Body of People, Usually Notably Ungoverned
Please be accurate - Wind River (note it's two words) bought the BSD/OS assets from BSDi in spring 2001. Not the late 1990's.
Sorry, wrong. I challenge you to find any posting that said that 6.0 would be released in June. It was widely publicized that the release process would START in June and hopefully conclude by August. It was also widely publicized that it would be released when it was ready, and not before.
Cat, the other, tastier white meat.
"Scott, I thought you used Linux? Anyway, I never knew you were so important!"
DAMMIT, here we go again!
I have no mod points. You are a troll. See sig.
If it weren't for the rocks in its bed, the stream would have no songs.
Boy we have a lot in common... I too was in the Navy, got started with BSD in the mid 90's after I got out, and live in/near Denver as well.. and now we have the TI-99/4a in common too... You didn't by any chance own a Apple II clone called a Franklin Ace did you?
There will be a 5.5 release once 6.0 is done. It will likely be the last 5.x release, and it will also be the seventh 5.x release. Not too shabby. 6.0 is the start of the 6-STABLE branch. There will be a 6.1 release a few months after 5.5. After that, there will only be 6.x releases until it's time for 7.0. That will be in approximately 2 years.
Cat, the other, tastier white meat.
> Then Linux came along and said "let's do it the Microsoft way!"
More like the IBM/Microsoft/Novell/SCO/Sun way -- the partitioning was always seen as a feature of the hardware architecture and not some nefarious MS invention.
BSD slices exist due to VAX legacy or some reason, it's a totally uncommon design on PC hardware, and IMO a valid if not mostly unimportant complaint. If you go to Troll Point #2, a LVM would make the issue irrelevant.
Whenever I hear the word 'Innovation', I reach for my pistol.
Surely is drivel - but useful drivel. Good of you to take the time.
How many beans make five, anyhow ?
Someone should let Andy Tanenbaum know - he has a chapter about Sun's N:M threading model...
NetBSD also has an interesting framework for Scheduler Activations which allows N:M threading.
I am TheRaven on Soylent News