Linux 2.6 Kernel Pool Results

jafo writes "Steve Ratcliffe, Master Software Project Estimator, has most correctly estimated the release of the Linux version 2.6 kernel. On January 6, 2001 (within 4 hours of when the 2.6 kernel pool was opened), he entered a guess which was accurate within 15 hours. Check the results for some interesting statistics and submit your guess for the 2.8 pool." See the original story if you like.

Comments

Missing Pool Option

[Anonymous Coward]

Duke Nukem: Forever release.

At least Linus didn't win ...

[FelixCat]

Since then we would have had proof the competition was fixed!

Well, made my guess...

[heironymouscoward]

But I'm not saying what it was. Otherwise everyone will want to vote for 29 September 2005!

How are we supposed to do it?

[Storm]

...With no LOTR releases slated for 2005...

Quick, somebody call Peter Jackson. Get him working on a Silmarillion release for December 2005...

Guess in the past

[Piggymon]

First guess submitted on 2001-01-06
Most optimistic guess: 2001-01-05 by Bill Segall
Bill, you're a GENIUS!

Call George

[ignipotentis]

And find out when exactly episode 3 will be released.

Link for 2.8 guess is wrong

[sparkes]

This link adds a guess for the 2.6 release. Anyone getting this wrong after the event deserves to lose ;-)

Re:Link for 2.8 guess is wrong

[PieEye]

It says that you're voting for 2.8, however the submission response page incorrectly states 2.6. You're voting for 2.8, and they'll correct the page eventually.

Interesting statistics...

[cperciva]

If you exclude the guesses which are obviously jokes (eg, 2038), then 96% of the guesses were optimistic. Obviously some elements of proprietary software development still hold true in OSS.

Also amusing: The median guess was April 1, 2002.

Re:Interesting statistics...

[StenD]

The same thing happens in proprietary software. Perhaps there's a lower rate of attrition in proptietary software once it's past the proposal stage, but I think that has a lot to do with the argument "We've already spent $bigbucks on the BlackHole project, do you want that to go to waste?", while with open source, if a project doesn't actually meet a need, the project is most likely going to wither away.

Finally!

[Guano_Jim]

I, for, one, have been waiting for a swimming pool that can run Linux.

Re:Finally!

[foobsr]

Here you have it ... http://truetex.com/poolcontrol.htm

CC.

SCO will tell us

[macdaddy]

Just ask McBride when he'll release the 2.8 kernel.

Diebold Machines?

[PincheGab]

Lemme guess... The kernel pool thing was run on Diebold voting machines, right? Give himself a 15 hour discrepancy to not make it look suspicious yet still win... No paper trail... hmmmm...

optimistic or stupid?

[gosand]

Most optimistic guess: 2001-01-05 19:36:18-07 by Bill Segall

Considering that the pool started on 2001-01-06,
"(First guess submitted on 2001-01-06 08:43:00 UTC by Ian Davis)"
could this really be considered optimistic, or just plain stupid?

Now this is an evil idea!

[SharpFang]

Imagine this, some hacker who takes part in kernel development submits a date, then sabotages the kernel, submits broken modules, introduces bugs, breaks stuff and keeps all the patches on his own harddisk. And finally, when kernel would be ready if not his own purposedly made bugs, he submits his patches one day before his "guessed date" and the kernel is released within next 24 hours, within 2h from his guess and four months later than it could've been released.

Re:Now this is an evil idea!

[mhesseltine]

While this seems funny, I doubt it would work for one simple reason; if someone is submitting bad patches, breaking modules, etc., I would hope that Linus, Alan, Marcello, (insert maintainer here) would quit accepting this person's patches.

After all, that's what the reputation game is all about. Those with talent are revered, those without are shunned until they can prove themselves worthy (probably through a stand-alone project).

What about the GNU/HURD Poll?

[miracle69]

Anyone run a release poll on that?

I wonder if Vegas has odds. If it does, I'd take the over.

Unfair advantage

[kbielefe]

Proprietary software managers have an unfair advantage in this pool. After all, for a living they make release schedules before anyone knows what the new software is actually going to do.

The schedule says we need to finish this in 6 months, but your estimates add up to a year so you'll have to do some stuff in parallel.

Uh, sir, it will still take just as long unless you hire some more staff. Even then, it will take some extra time for training.

But you said these modules were completely independent. You should be able to do two independent tasks in the same amount of time as one. Look how easy it is to schedule it that way in Microsoft Project.

Well, then why don't we all just do 12 at once so it will be done in a month instead of a year?

Don't be silly, it doesn't work that way. The schedule says six months. Boy, for how good you engineers are at algebra and calculus, you sure don't understand the basic mathematics of business.