Slashdot Mirror
FreeBSD 5.0 Delayed One Year
Satai writes: "FreeBSD 5.0-RELEASE has been delayed a full year, until November of 2002. The reasons included a lack of support for SMPng - including a developer fall-off ratio of 15 to 1 - a desire to finish the PowerPC/Sparc64/IA64 architectures, and a general desire to robustly test the additions. The economic downturn even makes an appearance in the announcement."
I thought it said FreeBSD 1.0 delayed for five years! What a relief!
"Where's my peanuts?" said the parakeet.
Can somebody, who knows what happened, explain to the rest of us why so many developers left the boat ?
15:1 is way above what can be regarded as "bad luck".
echo '[q]sa[ln0=aln80~Psnlbx]16isb572CCB9AE9DB03273snlbxq' |dc
*cough*TIM*cough*
anyway, at least the bsd-hackers forum can be quite hostile, and i've seen it keep more than a couple people away..
I don't want to start a FreeBSD vs Linux battle. I get enough of that from some of the people I know. But I have to admit that after using several Linux distros and using FreeBSD, the choice (for me) was quite clear. That's not to say I didn't like some of the Linux distros I tried. Not at all. I really liked Storm and I fully intend to install either Debian or Slackware on an IBM I have sitting in the corner. But when it came time to choose a system of the many I tried to run my web-server off of, I had to settle on FreeBSD.
At first I was a little wery about going with something slightly less mainstream than Linux, but good Linux binary compatibility (not to mention the Ports Collection) was a plus that won me over to FreeBSD.
With FreeBSD the first few days were really rough because there were several major annoyances I had, and none of my Linux friends had any useful insight. But I quickly solved most of my problems on my own. I feel I have learned much more this way. Plus, when I needed quick answers, web-searches almost always provided immediate and exact answers because there is only one FreeBSD and many other users have experienced the exact same problems.
It's something of a shame that Storm went the way of the wind, but after I made my choice to run FreeBSD it hasn't mattered too much. As for my soon-to-be Linux system, that just shows that I'm not knocking Linux at all (how could I?) it's just that I made the choice based on my needs and what I like. I personally don't feel I was moving forward fast enough with any of the Linux distros, but I felt comfortable with FreeBSD very quickly.
Conformity is the jailer of freedom and enemy of growth. -JFK
Well, it's pretty bad news if you are a manager in a company and fought for using BSD.
Boss - "So, are we on schedule to start rolling out the 4-way file servers in July next year?"
Me - "Um, no, that will have to wait until, maybe, Jan 2003".
Boss - "Errr, why's that? You said to me last quarter that the new SMP stuff would be ready by the end of this year? Surely 6 months is plenty of safety margin?"
Me - "Actually, the release date slipped by 12 months. I just found out now. I think it was due to most of the developers leaving the project."
Boss - "What?!?!?!! They fired 14 kernel developers?! I thought you said this organisation wouldn't be affected by the economy, on account of not being an evil capitalist outfit that only cares about their quarterly results!"
Me - "Yeah, well, no-one got fired, it's more like, they, uh, just kind of stopped doing any work. I guess maybe they got bored."
Boss - "OK, that does it. We're going with Solaris x86, I don't care what you say."
-----
As others have pointed out, it's good that the FreeBSD developers have decided to push the deadline by several (14) months.
But I can't help but wonder if the FreeBSD "core" isn't trying to do too much with too little.
SMPng is great. Porting FreeBSD to dozens of architecture may not be -- I thought NetBSD was the one group that was supposed to focus on portability? Stick with Intel CPUs, guys! =)
Nevertheless, a magnificent OS, and one that I use very often...
The right to offend is far more important than the right not to be offended. (Rowan Atkinson)
The reasons included a lack of support for SMPng - including a developer fall-off ratio of 15 to 1 - a desire to finish the PowerPC/Sparc64/IA64 architectures, and a general desire to robustly test the additions. The economic downturn even makes an appearance in the announcement."
You didn't take Butterfly Effect into account.
Failing to take into acount a butterfly flapping its wings in the country called Elbonia could cause the delay of its release up to several month, even a year.
Why are they setting dates in the first place? Can't they just say that they will release it when feature A, B and C are implemented and stable?
This isn't as bad as it might sound to Linux users.
FreeBSD has multiple branches:
* 5.0, aka -CURRENT, currently the target of
most new development.
* 4.4, the next release in the 4.x series,
due to be released today
* 4.3-RELEASE, which is updated with security
fixes as necessary
* 3.x, which is still being used, so it
occasionally gets a fix or two.
What this delay means is that the general public won't see most of the nifty 5.0 features until the end of next year.
That doesn't mean, however, that we won't get *any* new features; the list of 4.4 improvements will be evidence of that...
You can sit and whine about how its being delayed or you can get off your duff and help.
Even taking some time to run what parts of FreeBSD 5 do exist to give some valuable feedback as to how it behaves on your system could be useful.
I am just as dissapointed as anyone else about the news but I can't help but feel motivated to lend a hand in such bad times.
I will probably try FBSD 5 this weekend and see what's what. Too bad I don't have SMP...
IIRC, XFS does that too. In theory the same is possible with NTFS too, but MS chose not to implement it.
A deep unwavering belief is a sure sign you're missing something...
Guess what? OS X has 1 million users.
That doesn't include pirated copies, so it could be 1.25 million users of mac on a BSD base.
As apache is bundled with OS X, there could be 125,000 servers running on BSD from OS X alone.
OS X is doing great from all the unix apps for it.
- Kaos games and encryption systems developer
About a week ago, I saw the latest FreeBSD Development Report, compiled by Robert Watson. It's a simple report, including a paragraph or two on the state of all the major projects.
After reading the report, I decided to be a little bit scared of 5.0, because there were a lot of ambitious projects slated for inclusion therein.
This move strikes me as a recognition of a reality: it's going to take a lot of time to integrate all those projects and turn the result into something worthy of being called FreeBSD-RELEASE.
Ben "You have your mind on computers, it seems."
Well, it's pretty bad news if you are a manager in a company and fought for using BSD.
My answer to these people is: Maybe you should have taken the money you saved by not buying Solaris, and spent it by allocating half of one of your developer's time to the FreeBSD project?
Supporting Open Source has to mean actually doing some coding somewhere, or there isn't any source to support. Who better to code the features than those who genuinely need to use them?
Well, there's hope, it looks like Xandros is stepping up to fill the niche: debian-based desktop Linux distro. I wouldn't be surprised at all to see them bring some of the Storm team on board.
Life's a bitch but somebody's gotta do it.
"A cathedral will always be more beautiful and well-engineered than a bazaar. Even linux peasants should be able to figure that out."
Nice analogy. Should we take it a bit further?
1. A bazaar is always a better place to find open and valid information than a cathedral. Sure, you find a bit misinformation, but at least you have more than one source of it.
2. The sheer amount of activity at a bazaar is much much more than that of a cathedral. What point is there to beautiful engineering if noone uses it?
3. A cathedral takes ages to create. As a concrete example, the cathedral of Trondheim, "Nidarosdomen", was just finished this year, after starting in the 1400s. It took so long to build, that patchwork to make keep the current standards have been done for several hundred years before the building was actually finished.
4. The bazaar generates a huge amount of money and trade, compared to the cathedral.
A better approach:
1) Find some piece of hardware that isn't supported, but you think could be useful to more than a couple people worldwide
2) Post the question "Is anyone else working on a driver for [cool piece of hardware]?"
3) Commence work with any answering "I am" to #2 - or on your own, referring to driver-writing documentation included in the source tree.
4) Once working in a stable manner - post "Hey it works! Anyone want to help hammer it and make sure it's up to par?"
5) Gain respect and appreciation for your contribution to FreeBSD.
Simply asking "What needs to be done?" is like asking a star for a specific particle of light!
I AM, therefore I THINK!
He's talking about Trusted Solaris, not regular Solaris.
Quite frankly, that sounds like a very poor design. If you are changing your internal APIs so often and don't have a good abstraction layer in place for basic driver work, then you're shooting yourself in the foot. The internal APIs being in a "constant state of flux" shows that you need to get your heads out of the implementation detail, step back, and do some actual design work first.
A good OS should have clean interfaces for writing drivers that anyone can support. You should be ashamed that it requires an "expert" in the system to write one of the most commmonly needed contributions to the system.
I've written Linux device drivers as part of a class, including block and character devices with full /dev support and /proc filesystem support for configuring the devices. The Linux device driver model is so simple and easy to use that you can teach a class with a new, more complex driver as a biweekly assignment. It was basically a trivial task once you read through a little documentation. What is so wrong with FreeBSD that you can't simply do that?
Funny, if it wasn't an elite "boys club" on the list and if the kernel APIs were well designed, then you should be able to politely point someone to a HOWTO, FAQ, or book. Instead, snobbery and insults fly when it's really your own shoddy workmanship that is at fault.
Drivers are import to system adoption. Driver writing should be the low-level entry to kernel hacking for your OS. It's a good way to see who can code well and to enhance your system. Of course, if neither of these are your goal, then don't be whine when the supposedly "inferior" Linux kernel leads them to a higher mindshare in the developer community.
If it's for-profit but free, you're not the customer -- you're the product (e.g., the Slashdot Beta's "audience").
Fact: Jordan Hubbard did not leave the project - he simply changed employers. He is still the FreeBSD Release Engineer, and still active member of the CORE team.
Fact: FreeBSD-Current (5.x branch) has so many changes that pushing back the switch of Current to Stable does not mean that features from Current won't be MFC'd back to Stable during the course of the year. It just means the whole of it won't.
Assuming this is some sort of "writing on the wall" of FreeBSD's demise is incredibly short-sighted. If you truly have been involved in FreeBSD for 6 years, I would expect you to know better. The 4.x branch was delayed many times due to the amount of changes to various subsystems - some of which were then MFC'd to the then 3.x-Stable branch.
Passing FUD about the GPL beating BSD is just further evidence of your troll.
I AM, therefore I THINK!
It seems that this newbus bit may add a level of nice abstraction like that of the Linux kernel. It's good to see steps being taken in this direction. Hopefully, there will be more open-minded people willing to assist interested parties learn this interface.
What bothered me was the usual snobbery about it not being their "place" to help newbies learn how to write drivers. The whole API being in flux issue is mostly a red herring since they could simply request that the person submit their drivers against a frozen architecture, such as the FreeBSD 4.X branches -- unless these too are in a "constant state of flux." I mean, big deal -- the Linux driver models were quite different between 1.2 & 2.0 and between 2.0 and 2.2. That doesn't mean that people were turned away from submitting new drivers under the older stable tree during the 1.3 & 2.1 development cycles.
If they have a good interface, then they should really have documentation to help people add system support for the stable branches. I mean, really, the main "expert" developers shouldn't be bothering themselves with device driver writing. For one thing, it requires them to take time away from their usual projects to learn the interface for a new piece of hardware -- which should be the hardest part of writing a driver under a good architecture. Device driver writing is exactly what newbies should be doing. The kernel interfaces should be a trivial matter -- let the newbies worry about learning the actual hardware while the main developers work on more important core issues.
What bothers me is that this guy is. You can see his name all over FreeBSD mailing lists and code fixes. A quick search turns up that he's been a committer to the FreeBSD source tree since June 1999. This guy is in on things, and he's displaying this level of snobbery towards new developers. What a great way to gain mindshare! They're squandering a great resource.
If it's for-profit but free, you're not the customer -- you're the product (e.g., the Slashdot Beta's "audience").