Return to the Press page

LWN.net Weekly Edition for December 4, 2008

Tux3: the other next-generation filesystem

LWN.net Weekly Edition for November 27, 2008

LWN.net Weekly Edition for November 20, 2008

Weekly edition	Kernel	Security	Distributions	Search
Archives	Calendar	Subscribe	Write for LWN	LWN.net FAQ

Myths Open Source Developers Tell Ourselves (O'ReillyNet)

[Posted December 12, 2003 by corbet]

chromatic has some suggestions for free software developers in this ONLamp.com article. "Many of the users you most want to test your code before an official release won't. The phrase 'stable release' has special magic that 'alpha,' 'beta,' and 'prelease' lack. The best way to get user feedback is to release your code in a stable form."

(Log in to post comments)

Posted Dec 12, 2003 22:04 UTC (Fri) by mmarq (guest, #2332) [Link]

"Myth: Installation and configuration aren't as important as making the source available.

Reality: If it takes too much work just to get the software working,many people will silently quit."

Inexact Reality above: But that is one big flaw of the "Marketing Oriented" M$ proprietary software industry, that had convinced common user that every thing is just one click away,..., whyle at the same time it filled the entreprises with MSCEs just to secure the intermittent nature of their systems. When Computers stops of been sold has a toaster or a radio in big generalized commercial stores, then installation and configuration would get real help!

"...Reality: Potential(Linux) users become actual users through several steps...Conversely, the more difficult the installation, the more people will give up, often without giving you any feedback."

Inexact Reality above: IMO since the "big buumm" of the "White Box" that common users could have had enough help even for the more complicated systems, but,..., who would "give more" specialized help to anyone, by the price of a common PC, and when ~35-40% of that price is for Windows and Office alone ?? ... not even DELL or HP !!
The M$ effect and the forced commodization that accompanied it as put common users between a rock and a hard place.

""As to Linux, users give up simply because there arent good enough drivers for all that hardware out there, period.""

IMO this has to have a "TECHNICAL" solution;..., a split driver model, a HAL of some kind... ingeniosity is at Linux side! But it has to take off soon, because if M$ succeeds in locking everything behind NGCBS/Paladium, then there is no hope for Linux.

Myths Open Source Developers Tell Ourselves (O'ReillyNet)

Posted Dec 12, 2003 22:46 UTC (Fri) by freethinker (guest, #4397) [Link]

Fascinating. How long have you been channeling Dr. Bronner?

Myths Open Source Developers Tell Ourselves (O'ReillyNet)

Posted Dec 14, 2003 1:41 UTC (Sun) by mmarq (guest, #2332) [Link]

Almost 12 years in "that" business...

Packaging

Posted Dec 13, 2003 15:53 UTC (Sat) by addw (subscriber, #1771) [Link]

Myth: Installation and configuration aren't as important as making the source available.

Reality: If it takes too much work just to get the software working, many people will silently quit.

How true that comment is. How frequently I have come across something that looks useful, downloaded it, tried to get it going but failed and decided to spend my time somewhere else. What seems obvious to the software writers can be obscure to me, often I will download something because I want to learn about it - but I can't learn until I have it running.

A quick and easy install is vital. The first time round I don't really care about configuration more complicated than ``I want it to go into /usr/local or /usr''.

Myths Open Source Developers Tell Ourselves (O'ReillyNet)

Posted Dec 18, 2003 10:06 UTC (Thu) by beejaybee (guest, #1581) [Link]

"The best way to get user feedback is to release your code in a stable form."

And the best way to get your software a Bad Name is to release a beta (or even alpha) version with a "stable" tag.

stable = beta

Posted Dec 19, 2003 17:37 UTC (Fri) by giraffedata (subscriber, #1954) [Link]

And the best way to get your software a Bad Name is to release a beta (or even alpha) version with a "stable" tag.

I believe the point of the article is that you're releasing a beta version with a "stable" tag anyway, because your beta testing is a myth. The serious beta testing does not begin until you use the "stable" tag.

And that means we're just wasting our efforts having a complicated release stream that involves alpha, beta, prerelease, and patch releases. Better just to have "releases" with the implicit understanding that the earlier ones in a given release branch are more likely to have problems than the later ones.

And to manage expectations and protect your reputation, you should never refer to a new release branch as "stable." Call it "stabilizing." It isn't "stable" until you decide to stop releasing fixes in it.