In my humble, or not so humble view as it may be interpreted, I believe the only way to fix it, is to remove the technology from the developers hands.... and I'm not alone in this feeling.
I was invited to attend one of the meetings the PC-BSD developers hold some time ago. They do care, but despite whatever their best efforts are likely to be, I don't expect a
meaningful change in regards to PBI technology within the near future (e.g. several releases from now).
When it comes to maintainers, you basically have a chicken and egg problem. Documentation to help them know what they need to do (properly!) and people who actually have a clue what they need to do, without reading the docs. The quality of a PBI, depends on: the programs design being able to be "crammed" into the PBI model, the effort the maintainer / pbi developer put into making it happen, and the PBI Creator (PBC). Generally speaking, PBI/PBC is crap but it does get the job done, you just have to ignore (bad!!! but commonly done) or side step a few issues (annoying, makes maintaining harder), and the devs are little help beyond changing/bending rules to suit themselves or constituents.
The build server is a good idea, when the concept was originally purposed... those of us who advocated the idea were almost laughed at and told it was impossible. One person I know, has called the build server a joke in recent months; I haven't used it so I can't defend it. It is however a good thing in my opinion whatever it's +/-'s may be, because it is producing about the
only maintained set of PBI available, despite placing yet more unnecessary work on the PBI developers shoulders.
When it comes to supervision over PBIs, it has always been a joke. Having survived several engagements, all I've seen is incompetence, or conversely mismanagement -- and I am trying to be polite in my choice of words. If a PBI will install and run, it will likely make it's way to
www.pbidir.com. Many times I have been tempted to see "just who is watching". If I administered a group of PC-BSD machines for business usage,
www.pbidir.com would be
blocked from the network.
If they hired people to start making PBI and keying them into the build server, we'd probably have an end to the maintained problem; but we would still have crappy PBI to deal with.
I've seen much to much crap happen, to have any faith left in PBI. I've been stung/seen others stung/found stingers waiting to strike people; to often to use any PBI that I have not personally made, or was made from a source I can trust (there are none of those left now). And I eventually became fed up with wasting my time creating PBI, in a system so surrounded with rubbish. That is what I mean by the last part of '---> if still not effective, give up on PBI all together... (I am here)'
What it would take to change things, the developers seem to be either unaware of, or unwilling to do. As I have said else wheres about maintaince of PC-BSD releases, they have generally been maintained for a few months or so, roughly as long as it takes for work to start on the 'next release'.
PBI compatibility package
Linear Mode
