Thor Lancelot Simon wrote to the NetBSD-Advocacy Mailing List: “There are many upgrades we’d like to make to the NetBSD project infrastructure, but which we cannot make because, to be blunt, our project is poor. Not poor in innovation nor poor in developer resources nor poor in features — poor in cold, hard cash, the kind we need in order to buy hardware that would let us better serve our users.“Although it is unusual for the project to solicit donations from its users, the project is in dire need of money for upgrading its build infrastucture. In particular, it is looking to update the autobuild system and the anonymous CVS system. To help donate, please read the linked message and go to the NetBSD Project website. There is a PayPal link where users can donated directly to the project. Furthermore, there is an official NetBSD store at CafePress where one can buy various merchandise to support the project.
gutted, I got modded down..
Tho my points were valid, NetBSD does sport one of the best CPU port collections and the BSD license is much more attractive to some.
The embedded arena use OS’s like BSD a lot. Wassbi systems for example, is a NetBSD house providing embedded solutions and tool kits.
Hit up the companies making money off NetBSD (support companies, solution companies, and consulting companies), they are the ones who feel the pain if the OS lags other current offerings.
What confuses me is that I cannot see why you were moderated down. Your usage of language was subpar, and you were for all intents and purposes responding to an obvious troll, but other than being only vaguely condescending to the person trolling, your post wasn’t objectionable at all?
Presumably they’re asking for donations from any party that would find value in their work, whether that happens to be businesses or individuals.
This is what we can Free as in beer. It is just free. You do not have to give back. It is free in all means.
This is what we _call_ free as in beer. It is just free. You do not have to give back. It is free in all means no restrictions no BS just give us credit if you want to use it and yes you can donate if you like
Do you honestly think corporations care about open source or open source organizations? All they care about is money. If they can make money off of it they will. I don’t think they will donate anything they don’t have to.
Now, they may consider it for a tax write-off though.
However, some companies’ core software may be a BSD OS and it needs to be extended then they may upgrade the hardware.
Sounds like a nice build server, but why not go with an Athlon64 to save some money? You’re poor, right? Use poor man’s hardware, it goes a bit further.
Also why do you need it to fit in a 1U rack space? Isn’t there a cheaper way to do it? I could get that kind of system resources for less than $3000 for both systems if I didn’t have to meet the 1U requirement.
I would also consider using a dual-core Athlon64 for the build server.. Its about the same price as the Opteron.
If you make smarter decisions your money will go further. That’s what all this “use a linux cluster” stuff is about. Cheaper hardware with a free OS cuts costs significantly, unless you go buy more expensive hardware than is necessary. Let the big corps buy the Opterons until you guys can afford it.
Oh, and make it easier to donate.. credit card order forms are nice, but even paypal or something might help. I’d consider contributing a few dollars if there was an easy way to debit my account. And I run Linux, never use NetBSD. Tho I love the fish and openssh and songs and stuff. You guys are awesome!
I love OpenBSD, NetBSD sucks.
Its all about the fish.
But there is no price on the kind of great advice that is dispensed on osnews.com.
Next time I will try to read the whole article before making another stupid comment.. looks like Opterons are comparible price/performance to Athlon64s for that class of system, too. So yeah, everything I said can be redirected to /dev/null.
Oh, and make it easier to donate.. credit card order forms are nice, but even paypal or something might help. I’d consider contributing a few dollars if there was an easy way to debit my account.
It is certainly possible to donate via Paypal, for more information, have a look at:
http://www.netbsd.org/donations/
This makes a point, so please mod it down.
It’s kindof funny really. Why is it when I see threads riddled with troll infestations, nothing happens. But, when someone raises a point that is maybe slightly abrasive due to its ignorance or perhaps is simply just against popular opinion, its trashed? Shouldn’t those who have been provoking ridiculous OS wars be mod’ed down? It seems we should be letting even those who “don’t know as much as me” TM contribute. This for me has been the biggest problems about this site… Oh, well, mindless heated battles raise hits, I guess.
Btw – I feel NetBSD is a fine OS for some things, but prefer FreeBSD for my needs.
Even if I don’t use NetBSD I feel always sad when an open source project is in need of money.
I’ve always been a BSD fan, NetBSD particulary.
In other my posts I always said what I really miss from NetBSD:
Let me keep my system, even pkgsrc, up-to-date with binary upgrades with some simple actions.
Someone told me here at osnews that binary patches exist, please point me where they are.
I read in the mailing lists that login to anonCVS was suggested to apply patches…
The pkgsrc process to build packages is too slow, yes you have to wait 3 months.
More than one year ago I tried to compile gnome from sources and I got an error after many many hours of compilation.
It’s frustrating for an user with a slow machine as me.
If the NetBSD project is going to go to these directions I’d be happy to be your fan, even support you, otherwise I will stick with my debian machine.
I’m not using NetBSD, I’ve just tried it a few years ago. But as I read this request It was clear for me that I will spend 20 bucks to support them. An I did. What I become is an alternative and a nice warm feeling in my stomach 😉
As a said it, I’m not using NetBSD, but It’s a good feeling to know that there are alternatives, but only as long as they get some support (through donations or by submitting some code, etc.)!
I hope that the free software movement will exist for a long time despite of the all the threats that are existing at the moment (mainly patents)!
Cut that out Linus.
Why they ask money through paypal only and just can’t take mine because paypal’s stupid country limits?
Why they ask money through paypal only and just can’t take mine because paypal’s stupid country limits?
It is also possible to donate money in other ways. From http://www.netbsd.org/donations/ :
If you would prefer not to use PayPal, or would like to make other arrangements, please contact <[email protected]>.
I would like you to go look at the CPU support under NetBSD and then check the CPU ports under Linux.
I think you would be suprised.
Yeah I just did, and I am very surprised at all the noise people make about how many architectures NetBSD runs on – because Linux supports quite a lot more CPU architectures than NetBSD does.
So what was your point?
Linux may support some architectures that NetBSD doesn’t, but the meaning of “support” is quite different here. The kernel itself may be ported to a lot of CPUs, but this is often done in a very hackish, make-it-look-like-i386 way. And the kernel alone won’t give you a usable OS. On obscure architectures, you may have to collect a usable userland by yourself. Conclusion: Linux will look different on different archs.
NetBSD OTOH, can easily be cross-compiled for any arch, from the same source tree. So you have the same kernel and the same userland, with the same version, the same configuration, the same pkgsrc, etc, on any supported arch. This is invaluable if you have a heterogenous hardware park.
Yeah I just did, and I am very surprised at all the noise people make about how many architectures NetBSD runs on – because Linux supports quite a lot more CPU architectures than NetBSD does.
Just so nobody says I’m trolling:
Linux:
alpha
arm
arm26
cris
frv
h8300
i386
ia64
m32r
m68k (with, without MMU)
mips (32/64)
parisc (32/64)
ppc
ppc64
s390 (s390x)
sh
sh64
sparc
sparc64
v850
x86_64
xtensa
= 22
NetBSD:
alpha
arm
i386
m68k (std / sun mmu)
mips
ns32k
parisc
ppc
sh3
sh5
sparc
sparc64
vax
x86_64
= 14
Not to mention NetBSD has a few real old dinosaurs in that list, like vax and ns32k, while missing the 2 fastest general purpose CPU architectures available today (ia64, ppc64), mainframe (s390), and increasingly important embedded architectures (!mmu systems, cris, frv, h8300, v850)
Linux may support some architectures that NetBSD doesn’t, but the meaning of “support” is quite different here. The kernel itself may be ported to a lot of CPUs, but this is often done in a very hackish, make-it-look-like-i386 way.
Nonsense. Please stop trolling.
And the kernel alone won’t give you a usable OS. On obscure architectures, you may have to collect a usable userland by yourself. Conclusion: Linux will look different on different archs.
That’s true, Linux by itself won’t do much for you. Take Debian for example though (IIRC supports 11 or 12 architectures), and they are very well supported – probably better than many NetBSD ports that have “interesting” meanings of the word support too.
NetBSD OTOH, can easily be cross-compiled for any arch, from the same source tree.
Linux can be and is compiled for 22 architectures from the same source tree, no problem.
But there is no price on the kind of great advice that is dispensed on osnews.com
I’ll second that
Linux can be and is compiled for 22 architectures from the same source tree, no problem.
Could you cross-compile (without access to the actual hardware) a complete Linux kernel + userland + X + install media for, say, sparc, on your i386, by tomorrow?
It is sad to see the stupid comments that have been modded that go along the lines of “I don’t use NetBSD so who cares”. If that is the case, and you have used none of their work, which I would frankly find quite surprising, don’t make a comment.
NetBSD guys…thanks for all the good work you have done, particularly pkgsrc.
Sirs:
If you would like a box for snappier builds, try distcc. How about the devteam actually setting it up to work on that large scale. I, for one, would LOVE to get my Compaq 5500 working for them. ( Quad Pentium Pro, 4MB Cache ).
Or, how about buying that old CRAY that Apple Computer Inc has laying around? I also have a DEC Alpha, and a few spark boxes to add to the mix. How about a LIVE NetBSD CD that can become a distcc client?
All this would certainly be more productive than searching for little green men?
1U box? you only need a server to collect the results.
Best of Luck.
Actually these guys are both smart and reasonable: to wit:
http://mail-index.netbsd.org/netbsd-advocacy/2005/06/12/0002.html
“PARTS LIST #1: Standard NetBSD Foundation Build System
ITEM QTY PRICE TOTAL
Acme 8176T 1U Server 1 $685 $685
Chassis Rail Set 1 $ 45 $ 45
1U 64-bit PCI Riser Card 1 $ 19 $ 19
AMD Opteron 244 2 $252 $504
WD Raptor 74GB SATA disk drive 2 $196 $392
512MB PC3200 DDR ECC REG RAM 4 $ 78 $312
Labor, assemble & test 1 $ 25 $ 25
Shipping (approximate) 1 $ 40 $ 40
Total for system: $2022
The more of these we can buy, the more we can speed up autobuilds for you.
PARTS LIST #2: Replacement Anonymous CVS Server
Acme 8177T 1U Server 1 $799 $ 799
Chassis Rail Set 1 $ 45 $ 45
1U 64-bit PCI Riser Card 1 $ 19 $ 19
AMD Opteron 242 2 $201 $ 402
WD Raptor 74GB SATA disk drive 4 $196 $ 784
1GB PC2700 DDR ECC REG RAM 8 $138 $1104
Labor, assemble & test 1 $ 50 $ 50
Shipping (approximate) 1 $ 40 $ 40
Total for system: $3203”
Free as in free beer? I’d like to see it as “free as in free beer, with a donation box at the bar”. What I’d like to see is more stuff donated by Intel and AMD; a machine or two isn’t too much, I mean, out of a profit of $15billion, $20,000 for some servers is pretty piddly.
I’m a FreeBSD user and a Ubuntu Linux user, but I just donated $10 because I feel NetBSD plays an important role in the greater community as well. I wish them great success.
Uhm, 2% of the post has been about DONATIONS and the rest has been silly Linux/others wars and talk about things beeing free.
NetBSD makes a great system and I actually currently prefers it over FreeBSD and OpenBSD, and I would use all of them before whatever Linux dist if it wasn’t because of the evil hardware companies (and sun) out there not giving support
But I’m quite low on cash myself to so there’s not much I can do right now.
Every bit of cash helps. Imagine if *EVERY* user of *BSD (OpenBSD/FreeBSD/NetBSD) gave $20 each year; that would be a decent amount of money! imagine the cool things that could be accomplished!
Maybe now they understand the importance of a real business plan. Of course you will be totally broke if you just give your product away.
Regarding whatever open-source I feel my current problem is more an issue with the hardware companies. Like ATIs crappy gfx drivers and Canons unwillingness to give me decent Cups drivers for their new printers.
The amount of RAM they need/want isn’t an option in any Athlon 64 motherboard that I’ve ever seen.
If you would like a box for snappier builds, try distcc. How about the devteam actually setting it up to work on that large scale. I, for one, would LOVE to get my Compaq 5500 working for them. ( Quad Pentium Pro, 4MB Cache ).
Or, how about buying that old CRAY that Apple Computer Inc has laying around? I also have a DEC Alpha, and a few spark boxes to add to the mix. How about a LIVE NetBSD CD that can become a distcc client?
All this would certainly be more productive than searching for little green men?
1U box? you only need a server to collect the results.
I’m not a part of the team that produces release and stuff, but I would personally be very uncomfortable if we started releasing stuff out that was built on untrusted computers.
Quentin Garnier.
The amount of RAM they need/want isn’t an option in any Athlon 64 motherboard that I’ve ever seen.
It is. We have a rack full of dual-amd64 servers with 8 Gb RAM each.
And already i am seeing how well engineered it is. Debian is the only distro that comes close to a feeling of being properly built. But NetBSD (all the BSD’s) are on another level.
it feels wierd getting used to the /usr/pkg concept, but its a natural fit once you get going.
linux may be prettty good, but netbsd doesnt suffer from those flaws.
for all you comparing it to linux, when linux is the only system you have used, please sit down.
you are flat out, out of your league, you cant compare two systems when you have only really used one.
go back to slashdot and hype gentoo or whatever the distro of the week is.
What motherboard brand and model? Highest I had seen until now is 6GB!
Sun Fire V20z servers. They support up to 16 Gb.
http://www.sun.com/servers/entry/v20z/specs.jsp
We have a cluster of 256 of these guys. One rack (32) of them have 8 Gb RAM, the others 4 Gb.
I think one thing about NetBSD that is often overlooked by non-programmers is the code itself. NetBSD’s general philosophy engenders the type of clean code that is hard to find in other similar projects. In comparison, Linux’s focus on performance at all costs leads to code that, while it is stable and functional, is sometimes rather ungainly from an academic point of view. Consider the MMU abstraction in *BSD versus the MMU abstraction in Linux. *BSD uses a properly abstract pmap layer, that hides all MMU management behind kernel routines. In comparison, Linux uses a virtual 4-level page table that clients manipulate directly through tricky macros that map the 4-level view to the actual page table format. Or, consider SMP locking. NetBSD’s giant-lock model might not perform nearly as well as a fine-grained threading model, but it makes the code a lot simpler and easier to understand.
Now, why rant about the code when the users never see it? Open Source, fundementally, is about code. Code that is clean, elegant, and reusable is thus intrinsicly valuable. NetBSD’s code has been reused very often. The NetBSD TCP/IP stack is in the PSP, while the NetBSD VM (UVM), is in use in several microkernel research projects. I don’t think anyone has ever tried to reuse the Linux MM!
They need to also put a price list up there for adding some wanted features to the system, and use the profit margin on coder time to buy the damn hardware. You know, “Full support for (Pick one: [ReiserFS|XFS|JFS|Ext3]): $2,000. Full VAIO hardware support: $900. DRI equivalent: $5,000. Revolutionary GUI to replace X: $8,000. Forth in bootloader: $1,200.” I would certainly pay to see some of that stuff.
Rayiner I’m not a tech, but I know that some of summer projects for NetBSD is to port it to MMUless cpu.
A fact, though and as I said I’m not a tech but I’d like to hear your opinion on porting NetBSD to MMUless cpu very often used in embedded world, is that actually Linux can be ported to MMUless cpus, NetBSD no.
Has anyone considered peer to peer distribution?
such as http://NetBSD.org/mirrors/torrents/info.html which works for more than a year now?
So where are the ports packages? Can I simply grab a package with pkg_add -r?
Quentin Garnie wrote:
“I’m not a part of the team that produces release and stuff,
but I would personally be very uncomfortable
if we started releasing stuff out that was built on untrusted computers.”
Well, for one thing, you can never fully trust your compiler!
See: “Reflections on Trusting Trust” By Ken Thompson
and second, you can collect the object files, and .asm files,and compare it to a totally random source, (double the work, but verifible ) and if you find a pattern of errors, you can
1. Check to see if there is a problem with the cross compliation,
2. Remove the offending machine that is generating bad compliations.
The final build would take place on the server, but you could only trust this
to the extent that:
1. Your compiler is OK in the first place,
2. You verify the results, ( triple check for critical parts,
like the TCP/IP stack handler )
3. Your server is not comprimised.
Great concern, but it can be managed and mitagated to produce faster and faster verifiable results.
you are gonna compare a really well hidden compiler trick like that to building software on a RANDOM computer on the internet.
there is trust, then there is just wackyness if you think those two even are in the same ballpark.