User: Password:
Subscribe / Log in / New account

A proposal for an always-releasable Debian

A proposal for an always-releasable Debian

Posted May 13, 2013 11:30 UTC (Mon) by renox (subscriber, #23785)
In reply to: A proposal for an always-releasable Debian by ovitters
Parent article: A proposal for an always-releasable Debian

>That's just due to lack of manpower.

Totally false, it's that the "manpower" prefers to do new things instead of maintaining old versions.

(Log in to post comments)

A proposal for an always-releasable Debian

Posted May 13, 2013 11:39 UTC (Mon) by ovitters (subscriber, #27950) [Link]

Pretty cool you want to force people to maintain old versions, but we rather leave it to people who want to maintain old version. As you have no ability for force anyone to do anything, in practice it is what I already stated: lack of manpower. And actually some maintainers do maintain old versions, just not across entire GNOME.

Saying "totally false": Read up on the "No obligation" bit in to get an idea on how things work in practice.

A proposal for an always-releasable Debian

Posted May 14, 2013 13:43 UTC (Tue) by raven667 (subscriber, #5198) [Link]

What I don't get is how the Linux kernel can be so far ahead when it comes to development practices and how long it takes for userspace projects and distros to pick up their habits. The kernel has been maintained in an always-releasable state for a decade now with regular releases every quarter, including stable releases of longer-term maintained versions. The kernel has a fixed ABI to userspace that is ruthlessly maintained. The kernel also has plenty of contributions from amateurs and hobbists who provide labor on their own terms that co-exist with a large number of professional paid developers. As far as "No Obligation" goes, if you want to have your code be a part of the "Linux" kernel then you have a mighty obligation to not break the build, not break the ABI, no break peoples systems and adhere to coding standards, or your code is not going to be accepted and Linus is not going to put his name on it. How is this different for maintaining standards for Debian or GNOME?

A proposal for an always-releasable Debian

Posted May 14, 2013 21:14 UTC (Tue) by jmorris42 (guest, #2203) [Link]

Pretty simple explanation. But if LWN had moderation I'd be getting a -5 troll mod for what I'm fixin' to say....

The kernel devs are professionals. They develop software that is used in the real world in massive installations ranging from Google, supercomputing clusters and pretty much every friggin' smartphone without fruit on it. Most televisions, cable boxes, BluRay players, NAS boxes, etc. depend on that codebase working. If they screw up the people who pay them will not be happy. There would be consequences.

The GNOMEs are pretty much the opposite in every way that matters. The other desktop efforts aren't much better. Because they don't have to be. GNOME released a product that pretty much everyone and their dog hated and other than a lot of heat on Internet forums did it really make a difference to anyone's bottom line? Did it either bring about or prevent 'The Year of Linux on the Desktop!"? No.

Until there are consequences nothing will improve. And as long as it is kids playing with their pet projects there won't be consequences because there won't be enough users to matter. Do I have the answer to how to break that loop? No.

A proposal for an always-releasable Debian

Posted May 13, 2013 13:05 UTC (Mon) by mpr22 (subscriber, #60784) [Link]

People who don't care about maintenance of old versions generally do a terrible job of it. The way to get effective maintenance of old versions is for the people who want it to band together and either do it themselves, or offer some reward (not necessarily tangible) to get someone else to also care and do it for them. "Strangers will be less rude about you on the Internet" is not a particularly inspiring reward.

Copyright © 2017, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds