nic2109 wrote:...There seems to be general consent that Slitaz (among others) has better package management but wild disagreement about whether it's worth emulating.
Is this really the end of the discussion? I really cannot see that robust and flexible package management threatens the freedom many yearn for, but it can give the stability that others seek. Is it really that difficult?
The problem, as I see it, is not really the package management software, with the exception of some obvious bugs which can and should be dealt with. The problem lies in that repository of officially-sanctioned packages known to work without problems.
What I wanted to emphasize is that Puppy is not limited in the way many other distributions are. BarryK has created the roots and trunk of a tree with many limbs and branches. He has declined to trim this into a pole with a single path of development.
You are entirely free to create a repository you like, with your own choices and standards. You can post about this in the forum and wiki. You can set out your own manifesto of purpose. You can recruit others to your cause, even using this forum. You can borrow code from all over. You can modify a puplet to work exclusively with this repository. You can create your own package management system.
What you cannot do is call this the one and only official Puppy repository. Fights over control have crippled such promising software development projects as Gentoo Linux. Barry has gone to some lengths to avoid this.
"MU" created a substantial fork from the Puppy 3.x series called Muppy. "Big Bass" has a fork from 4.12 called Slaxer. "Ttuuxx" is supporting projects from the 2.x series to the latest development. As far as I know, Barry is on good terms with all of these people.
The hidden agenda of some people who call for standards is an attempt to force others to do things their way. They may not even be aware of this motivation when they start out, but it shows up with depressing regularity.
Go back to fundamental documents and you will see the statement "Puppy is an evolutionary system". The development process we follow is as messy as organic evolution, the results are uncertain, but we are doing better at innovating than a great many sizable organizations.
If you have ever been part of such an organization, you will immediately recognize the truth of this observation: you can say many things about software development in such a context, but you would not call it fun. This is the missing ingredient which brings developers to Puppy, and turns users into developers. Whatever changes you would make to the way we are operating, don't take this advantage away.
This is not to say I am complacent about the current state of the art here. I, too, would like to spend less time ferreting out places people have hidden files in the file system. I, too, would like to be able to tell people "just stick with these packages, and you will have no trouble." I, too, would like to spend less memory on remembering tricks which surmount obscure problems.
What I'm trying to convey is not that your goal is impossible, or not worthwhile, but that how you get there is every bit as important as where you are going. If you want to clarify the structure of the Puppy filesystem, you will have to make your case, even provide examples. If you convince enough people, the changes will become part of Puppy. If you want particular libraries, likewise, you will have to make a particular case for each.
Avoiding the hard work and tough decisions doesn't result in a lean minidistribution, it results in bloat that swallows resources. Fitting the minimum together to give many people most of what they want is a serious intellectual challenge, and a great deal of work.
This disorganized bunch has gone a lot farther than I would have predicted years ago. Whatever they are doing right, I don't want to destroy the basis for this accomplishment. Thus, my challenge to find a distributed method of harnessing all the wasted effort and unsatisfied requests seen on this forum. Finding this could be more important than any particular piece of software.
I'm afraid I've gone off again, my original intent was to make some pertinent observations on the construction and growth of program libraries. That will have to wait for another opportunity. However, rereading what I've written, I think it was worth saying. Let it stand.