Something to think about before your next post.

[melon]

Playing Devil's Advocate here...

Without tribalism, our attention is divided to others' ideas, and we cannot focus on our own work.

[Ex-Cyber]

Playing Devil's Advocate here...

Without tribalism, our attention is divided to others' ideas, and we cannot focus on our own work.

Rejecting tribalism doesn't require that we pay attention to everything or everyone equally. It does mean that we cannot ignore someone because he is "not one of us". In this case "one of us" means a category that has no meaningful connection to the topic at hand, rather than a category that meaningfully indicates someone's relevant interest or expertise.

Obviously, if you're a kernel hacker grappling with a kernel problem, you're going to value the input of other kernel hackers more highly than that of non-kernel hackers. That's not tribalism. An example of tribalism might be (for example) if you valued the input of kernel hackers over non-kernel hackers in something that had nothing to do with kernels, such as a debate about monetary policy.

Likewise, another example might be if you dismissed someone's opinion on the kernel because of their nationality, their ethnicity, their religion, their favorite football team, their alma mater, etc..

A more subtle example might be if you're trying to deal with a somewhat general problem (e.g. analyzing the algorithmic properties of a scheduler) and dismiss someone's advice because they work on a different kernel than you do (even though that kernel presumably contains a solution to the same general problem, possibly using an algorithm related to the one you're considering).

All of these sound somewhat ridiculous, but they're more realistic than you'd think, because people often make these kinds of judgments subconsciously, and it's a matter of degree rather than all-or-nothing. The absurdity of it is fairly obvious when it's spelled out, but it's usually something that happens without the person doing it even noticing.

[yotambien]

It was a general comment, not an indictment of Ubuntu, though I can't see how anyone can deny that they contribute far less in terms of upstream contributions than any major distribution.

Packaging Linux an making it pretty is completely fine and legitimate, I was just pointing out that it has never worked for anybody for a long time.

The reason for this, IMHO, is that people will happily switch from such distros as soon as the next shinier thing comes along. People stay with SuSE or RedHat or Debian for more reason, and the contributions to the kernel, KDE, GNOME, GCC, drivers and other technologies is one of the important reasons.

I guess it remains to be seen whether they're just packaging a bunch of software and making it pretty or there's actually some value added.

But those Gnome census metrics are pretty useless. On the one hand, they're just that, lines of code for that one project. No one can deny Red Hat's involvement and importance of its contributions, and these statistics are a poor picture of what they do for the Linux desktop: what's more valuable, their 16% Gnome contributions or their paid Xorg developers without whose work it'd be harder for some to, well, have a desktop? On the other, code contributions aren't everything. Canonical somehow managed to attract many users and change the perception of Linux as a viable OS alternative. I don't think it's just hype, although hype there is.