Home opinion-and-analysis Open Sauce Six-monthly releases: OpenBSD shows the way

Six-monthly releases: OpenBSD shows the way

Six-monthly releases have become something of a talking point in free and open source software circles after the problems Ubuntu has faced with users unhappy over major bugs.

While Ubuntu, which is now five years old, appears to struggle with this pace of development, the OpenBSD project has been doing six-monthly releases for the last 12 years - with no major bugs.

The head of the project, Theo de Raadt, says he came to the six-monthly schedule as a result of his experiences with NetBSD; he started the OpenBSD project in 1996 after being shut out by the NetBSD team.

"NetBSD was dragging out releases - 'not ready yet!' - which ended up full of stupid bugs which should have been caught but were not, and I felt this was because it was just a replay of the vendor model of software development: 'product driven', 'must meet milestones' 'ship the new candy we planned'," he said in an interview with iTWire.

"Doing it for fun, we had an opportunity to make it about ourselves. The minute I write some code or fix it, I want as many people to run it so that they can spot bugs while my mind is fresh. The commercial model does not match that desire."

De Raadt says that if software is not always ready for users to use today - every day - then it never will be. "Prod our users to test often, fix bugs while we can still remember the area of the code," is his motto.

"The Linux developers say the same things, but their commercial sales buddies doing official releases act just like Sun, IBM, Apple, Microsoft... except without the testing manpower or feedback and dollar-based backing to ensure things are fixed on time," he says with something of an air of superiority.

"I don't think both approaches are possible. I don't think there is a hybrid. The hybrid models are failing to produce the promised quality."

He does make some compromises but feels they are worthwhile. "The #1 complaint we do is that we don't support old software. Well, it isn't in my nature to care about old problematic software I don't remember because we are slowly replacing it because it has problems... a lot of our users have realised that they can't have it both ways."

De Raadt says that by the time one release takes place, the developers are already six weeks into the next development cycle. "The release is a branch off the main development
tree. In many other projects, it is a live branch, as in a separate team cuts it off and then keeps making minor tweaks to it to make sure it is a good release (why? that is because the mainline is crap).

"In OpenBSD, the release branch is a dead branch. The day it split from the trunk it was determined to be good enough for making a real release. Without any changes. That is because the trunk is good stuff."

CONTINUED


He says everyone knows the release schedule. "They don't know the particulars - there are random slowdown and lock dates thrown in to throw the optimistic new developers off-track- as this serves to make sure they pay attention. In general, everyone knows that if they get in the way, Theo will not be happy, and then the release process will be highly unpleasant.

"Everyone is involved in the release process, it is loud and obnoxious and no-one is exempt, and then it is fixed, and it ships. Everyone participates because otherwise stuff gets turned off and then their efforts are in vain."

Each developer does what he or she wants to do and rarely more than that. "We share ideas, and sometimes someone who is more in tune with a certain area takes it on," says De Raadt.

"Every release I have a few ideas that are 'must do'. That typically requires the focus of three people for specific sub-projects. Sometimes some of them get put off. But other times hackathons really resolve these issues, since people can operate full-time for a week. Hackathons make OpenBSD different."

He is somewhat cynical about the bling that many operating systems pursue. "The world doesn't live off jam and fancy perfumes - it lives off bread and meat and potatoes. Nothing changes. All the big fancy stuff is sloppy stuff that crashes. I don't need dancing baloney - I need stuff that works. That's not as pretty, and just as hard.

"Our developers are addicts on their own paths. I don't think they choose. They are chosen by their circumstances, entirely outside of OpenBSD. They come hack on OpenBSD because they found reliable value, and they want to increase it."

"My priorities are (2) to not screw ourselves by being too optimistic and (3)to move ahead. Ahead of that, my #1 priority is that OpenBSD is a small enough group that we are friends, and that should not be hurt.

"There are things I can ask for, things I want, things I think we need. People do listen to me, but I don't dictate what will or won't happen. When I do dictate, it has a chance to backfire on me.

"The only thing I really demand - in our development community - is that we not introduce risk against quality by putting preference on future features, and the most important part of that is follow-through (if you start hacking on something and start committing parts of it, you better finish it and get it bug-free, or back it out before it wastes some other developer's time).

"But I am just preaching to the converted when I tell my audience that. As a close-knit team, we all understand that. If something is broken, it gets backed out."

De Raadt sees no difference with avoiding problems in releases, no matter to which range of machines the release is catering. "We don't cater specifically (to desktops). Laptops, desktops, servers. I don't see what the difference is unless you are addicted to candy and dancing baloney."

He says the projects has hundreds of security angles in mind to explore, but is very limited in development talent. "We are ahead, so I think we can afford to move a bit slower these days."

LEARN NBN TRICKS AND TRAPS WITH FREE NBN SURVIVAL GUIDE

Did you know: Key business communication services may not work on the NBN?

Would your office survive without a phone, fax or email?

Avoid disruption and despair for your business.

Learn the NBN tricks and traps with your FREE 10-page NBN Business Survival Guide

The NBN Business Survival Guide answers your key questions:

· When can I get NBN?
· Will my business phones work?
· Will fax & EFTPOS be affected?
· How much will NBN cost?
· When should I start preparing?

DOWNLOAD NOW!

Sam Varghese

website statistics

A professional journalist with decades of experience, Sam for nine years used DOS and then Windows, which led him to start experimenting with GNU/Linux in 1998. Since then he has written widely about the use of both free and open source software, and the people behind the code. His personal blog is titled Irregular Expression.