From the article: “In order to allow our users to follow the most important changes over the last few months, we provide a brief summary in these official status reports on a regular basis. These status reports, released with irregular regularity, are suitable for reproduction and publication in part or in whole as long as the source is clearly indicated. This report summarizes the changes within NetBSD during the first three months of 2006.“
The NetBSD Foundation Quarterly Report: January – March 2006
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
16 Comments
-
2006-04-20 7:01 amnick
May I suggest that you read what I wrote in respects to the actual issue at hand; I’m not talking about the programmer, but the end user.
Err, I did read what you wrote. How about you read what
I wrote. I am talking about the end user. License for
distributing the software has no effect on them at all
(when it comes to BSD vs GPL).
And people make their decisions, strictly on what benchmarks spew out?
Umm, yes, they do (especially when Linux is 50% faster
than Solaris 10).
Benchmarking is big business, take a look at tpc if you
want to see companies regularly spending millions of
dollars on benchmarks.
Yes I am ‘qualified’ to know what a ‘hack’ is, as well as knowing what a well thought out design to address a flaw in the system.
You are qualified to comment on kernel code are you?
Because I don’t think you give Linux kernel engineers
very much credit. What are your qualifications, then?
What kernels have you worked on?
http://lwn.net/1999/0121/a/vmreview.html
Here is someone, a couple of years back, who is
qualified to comment on kernel code. He actually said
this of Linux:
“What FreeBSD can learn
Well, the main thing is that the Linux VM
system is very, very clean compared to the
FreeBSD implementation.”
-
2006-04-20 11:50 pmhustomte
That was from freaking ’99. Some water has passed under the bridge since. The current state of the VM in both projects is, to say the least, quite different today.
So: it was a very bad example.
-
2006-04-21 5:08 amnick
That was from freaking ’99. Some water has passed under the bridge since. The current state of the VM in both projects is, to say the least, quite different today.
You’re right. Linux has got much cleaner and more
algorithmically sound since then. Not sure about
FreeBSD, but considering it is based on the same
underlying design (FreeBSD’s pmap variant), I don’t
think Linux will have gone from much cleaner to
“randomly throwing hacks at the problem” in comparison.
So: it was a very bad example.
No it wasn’t. It illustrates exactly what I want.
-
2006-04-21 6:01 ambubbayank
That was from freaking ’99. Some water has passed under the bridge since. The current state of the VM in both projects is, to say the least, quite different today.
So: it was a very bad example.
But an excellent troll!
A good example of a hack is “coding for benchmarks”.
-
2006-04-21 7:22 amnick
But an excellent troll!
How was it a troll? kawai’s FUD was a troll.
A good example of a hack is “coding for benchmarks”.
Umm, those benchmarks were run by an Apache developer.
No Linux kernel tuning was done, and no Linux
developers “coded for it”.
Also, a good example of good engineering is “improving
performance on benchmarks when it doesnt compromise
other areas of performance, or where those
compromises are known, and accepted”. This is what
it takes to get patches into Linux. Look on the
OpenSolaris lists, this is what Sun developers do too.
Often, the only things that kernel developers have is
benchmarks, representative of a particular workloaad
that some user is experiencing poor performance on.
So they code for the benchmark, then go back and ask
the user to verify their solution. Nothing wrong with
that.
Let’s have another look at Linux “getting the job done”
http://marc.theaimsgroup.com/?t=114416927400001&r=1&w=2
Sure this is only a benchmark, but with the two year
old Linux beating FreeBSD 6 HEAD by more than a factor
of 7, this suggests Linux might have a slight
real-world performance advantage with MySQL.
The NetBSD internals documentation looks especially decent.
I didn’t realise it wasn’t possible to boot NetBSD from GRUB till now..quite a big miss imho.
I didn’t realise it wasn’t possible to boot NetBSD from GRUB till now..quite a big miss imho.
It was, in two ways even: you could either chainload a NetBSD partition (i.e. invoking NetBSD’s own bootloader), or boot a NetBSD kernel directly (using kernel –type=netbsd). But with the latter option, you couldn’t pass any arguments to the kernel (like -s for single-user mode). Now that there is a third method (using the multi-boot specification), you can pass any argument to the NetBSD kernel from GRUB.
I have 5 years of linux experience, and don´t pay much attention to netbsd…until now (I’m very frustrated about linux development).NetBSD is a clean and simple system, and we can concentrate in the real job (not in the configure&&make&&install).This reports is a demonstration of the atitude for the users of the system. GREAT JOB!!
In no way is this intended to be trolling or whatnot, but being someone who has used both Linux and BSD, I feel most Linux users, in general, will find the BSDs to be “clean and simple” and should definitely give them a spin. Both OSs have their place, pros and cons, but after using the BSDs, I try to put them into use wherever it makes sense.
Got BSD?
I feel most Linux users, in general, will find the BSDs to be “clean and simple” and should definitely give them a spin.
more likely different and somtimes difficult to use because netbsd like the rest of te BSDs still follow the UNIX way of doing things. unlike linux like ubuntu where the user has the option of using GUI configueration wizards.
http://www.openbsd.com/faq/faq9.html
it does not apply to netbsd but it applies to BSDs, in general more likely to openbsd, but it’s a good start for a linux user that wants to add a BSD os to their list of OSs
I feel most Linux users, in general, will find the BSDs to be “clean and simple” and should definitely give them a spin.
more likely different and somtimes difficult to use because netbsd like the rest of te BSDs still follow the UNIX way of doing things. unlike linux like ubuntu where the user has the option of using GUI configueration wizards.
There are actually alot of Linux users, who used Linux as training wheels to UNIX, and move onto BSD – now sure, there are those who hug onto the GPL like some sort of safety blanket, but there are those who go with what ever does the job.
For me, I started off using Slackware years and years ago, and eventually made the move to FreeBSD, and simply switched between FreeBSD and Solaris, constantly moving back to FreeBSD for its structured and well tested ports, and concerntration by the developers on getting things right, rather than randomly throwing hacks at the problem in a hope of correcting problems.
There are actually alot of Linux users, who used Linux as training wheels to UNIX, and move onto BSD – now sure, there are those who hug onto the GPL like some sort of safety blanket,
Why would this be the case? Choice of license for
distribution has no impact on users.
but there are those who go with what ever does the job.
There are times that Linux is the best tool for the job
too http://www.stdlib.net/~colmmacc/
For me, I started off using Slackware years and years ago, and eventually made the move to FreeBSD, and simply switched between FreeBSD and Solaris, constantly moving back to FreeBSD for its structured and well tested ports, and concerntration by the developers on getting things right, rather than randomly throwing hacks at the problem in a hope of correcting problems.
I don’t think you’re qualified to know what is a
“hack” and what is “getting things right” when it comes
to kernel programming.
if netbsd had a system of binary updates and updating similar to apt-get update/upgrade that would make the system very attractive to many people.
i’ve tried both pkgsrc and freebsd-ports and it is not a stable way to maintain a system. this is why sometimes the technically inferior linux distributions win out becuase that are easier to manage.
i realise that netbsd is often used as a *basis* for others to build systems and appliances out of. in that case perhaps its time for an desktop and server distribution of netbsd.
First and foremost, sorry for being off-topic in regards to NetBSD.
In the FreeBSD case, please refer to the manpages of portupgrade, portsnap and portversion, in particular the switches that control whether you prefer installing from ports or from packages, and pkg_info, in particular the -r and -R options. It’s not GUI-based, but it’s so darned easy to manage that I simply find it more convenient than trying to track what dependencies Synaptic (for example) is adding behind my back. After adding some aliases to your preferred shell’s local .rc file, port/package management is a no-brainer.
About binary upgrades to the system core, please refer to freebsd-update. It IS actively mantained.
Why would this be the case? Choice of license for distribution has no impact on users.
May I suggest that you read what I wrote in respects to the actual issue at hand; I’m not talking about the programmer, but the end user.
There are times that Linux is the best tool for the job too