Not 100% sure [LWN.net]

Not 100% sure

Posted Dec 20, 2007 15:21 UTC (Thu) by mingo (subscriber, #31122)
In reply to: Not 100% sure by paulj
Parent article: Insufficiently free?

So RMS is factually wrong because you think something is plain and obvious? If it's obvious, could you provide some empirical data to back up your claim? I note that RMS did not claim to be factually correct, but rather stating his long-standing position on these things.

The proof is right in this very article. Less than 1% of Linux users use a distribution that Richard approves of, still FOSS is prosperous. 1 billion lines of free code has been written in the past 15 years. If that is not enough for someone to accept that "the mix of 99% free plus 1% unfree" is fine then nothing else will. This community has literally freed itself out of the shackles of closed code via an epic effort of writing 1 billion lines of code (which is the largest ever single scientific project that mankind has undertaken) - if that is not proof, what else is?

Contrast that with the following recent experience i had with FOSS developers who left Linux. I recently triggered a rather nasty bash problem (hung scripts) that i've seen for many years. I never actually realized that this was incorrect behavior, until i asked around. It turns out it was a long-standing bug in bash, but the current maintainer of GNU bash uses OSX so he never triggered it himself. Developers of FOSS working on other platforms are actively harmful to Linux and other FOSS projects - all the network effects are missing. A small proportion of unfree software on a free platform is a lot less harmful on the other hand. (but i could cite many other examples)

By Richard's argument, it's better to develop and use GNU Bash on OSX than it is to use a single piece of unfree software on Linux, amongst thousands of other free packages.

That is in plain contradiction with hard facts.

(Log in to post comments)

Not 100% sure

Posted Dec 20, 2007 16:00 UTC (Thu) by paulj (subscriber, #341) [Link]

Less than 1% of Linux users use a distribution that Richard approves of, still FOSS is prosperous. ... By Richard's argument, it's better to develop and use GNU Bash on OSX than it is to use a single piece of unfree software on Linux,

But that's nothing to do with RMS' position and his justification for it, which you yourself quoted:

People are unlikely to switch to a non-free operating system merely because a free program runs on it.

His position is whether non-free software should be recommended, not whether FOSS platforms can be successful without RMS' approval (note that RMS in that thread posted that he does indeed approve of OpenBSD), nor whether it is better to develop on unfree platforms. On these new, latter two points, you appear to be attacking a straw-man, rather than RMS' position which is the subject of this article. (Feel free to point me to posts that show you're not attacking a strawman).

Developers of FOSS working on other platforms are actively harmful to Linux and other FOSS projects - all the network effects are missing.

I'd be one such a developer, so I have to disagree with your assertion that I'm harmful to FOSS generally. I'll have to agree to disagree with you on a Linux monoculture being good for FOSS.

Not 100% sure

Posted Jan 3, 2008 15:28 UTC (Thu) by anton (subscriber, #25547) [Link]

Contrast that with the following recent experience i had with FOSS developers who left Linux. I recently triggered a rather nasty bash problem (hung scripts) that i've seen for many years. I never actually realized that this was incorrect behavior, until i asked around. It turns out it was a long-standing bug in bash, but the current maintainer of GNU bash uses OSX so he never triggered it himself.

Why do you think that this maintainer left Linux? Maybe he has always been a MacOS guy who came to free software because it was available there. In any case, if the GNU software was not available on MacOS X, would we have one Linux user more, or one bash maintainer less?

Also, note that the issue that RMS discussed was his recommendations; one is usually much more circumspect about recommending something than about using something.

The GNU recommendations about platforms explicitly state that GNU and GNU/Linux are most important, that one should test personally on these platforms, and that supporting other platforms is optional. Is RMS a hypocrite because the bash maintainer has not followed this recommendation?