On 04/01/16 09:47, Tristan Van Berkom wrote:
That aside, I think we also have to take a step back and recognize
GNU/Linux desktop distributions are largely community maintained. Which
means that what might be perceived as a problem may in fact be an
advantage in this context.
Yes. So not having a QA department means that's mostly done via users
testing and reporting bugs. The community is part of the verification.
As a GNU/Linux desktop user I would say that the ability to roll back
updates is not very important if I can have a high degree of confidence
that the updates will work. At least for the distros I use, the number
of times I've wanted to roll back an update has been maybe once or twice
out of many thousands.
I don't know about on Apple systems but software installation on
GNU/Linux has always been vastly superior to that of the Windows
experience. In Windows land the user has to hunt for executables of
On the question of unverifiability due to combinatorics I think it's
reasonable to expect that there will be a large amount of variety in the
way that users want to configure their systems. The world is complex and
software needs to match that. In this scenario the best verifiability
you can expect is to do a random sample on the most common system
configurations. If system configurations and their stability were
broadcast then that distribution could be known - similar to the Debian