Release Early, Release Often [LWN.net]

From:		Jonathan Day <imipak-AT-yahoo.com>
To:		letters-AT-lwn.net
Subject:		Release Early, Release Often
Date:		Fri, 19 Mar 2004 06:01:00 -0800 (PST)

Dear editors,

The one thing most software developers forget is that
if the project is dead to the world, it will often
become simply dead.

At the moment, I'm looking at software routers - an
area notorious for slow to non-existant releases.
Zebra, for example, has a commercial offshoot, and the
Open Source version has since come to a halt. Who,
though, is going to buy the commercial product, if
they perceive the project as dead?

Click, a router from MIT, is better in that they have
just made a release. A very large, bulky, and no doubt
bug-ridden release. That's the reason for releasing
often - bugs breed in the dark, and die off in the
light.

Mind you, if you are after a software router, Click is
the only one out there with any decent releases at
all. It's also very fast and does support a lot more
than the others ever did. All the others are sleeping,
comatose or dead.

If you want to encourage Open Source - and I think we
all do - then release early and release often. LWN
does, with its headline news items, though you could
argue it's not really software. The point is, though,
the model works and the alternative doesn't.

Jonathan Day

(Log in to post comments)

Isn't Quagga alive?

Posted Mar 25, 2004 5:18 UTC (Thu) by nicku (subscriber, #777) [Link]

Have you tried quagga? It looks pretty alive to me.

Isn't Quagga alive?

Posted Mar 26, 2004 0:31 UTC (Fri) by nowster (subscriber, #67) [Link]

Quagga forked off Zebra, probably because Zebra was moribund.

Isn't Quagga alive?

Posted Apr 2, 2004 12:08 UTC (Fri) by ingvar (guest, #1530) [Link]

Quagga forked because there was a feeling that absolutely nothing was happening with Zebra (active community, but nothing readily happening in the Official Release). Thus, it was decided to take the existing Zebra source, integrating patches that had existed for 1-2 years in the field and simply go from there.

Release Early, Release Often

Posted Mar 26, 2004 23:22 UTC (Fri) by giraffedata (subscriber, #1954) [Link]

I'm confused. You start off by indicating that you should release early and often to avoid a project appearing to be dead. Then you give examples of projects that, for all I know, are actually dead.

Is there some reason to believe these projects are alive, but just saving up changes for release far in the future?