Re: [boost] Core libraries should separated from experimental libraries
Thomas Klimpel said:
I don't know whether I understand your concerns. I'm just not sure that it will be easy to agree on a set of core libraries. If it were just the official boost libraries themselves, it would at least be easy to figure out which library must necessarily be considered as "core". But you will never know exactly which boost libraries are used by closely boost related projects, like the libraries in the sandbox. For them, the official release boost libraries present the "core" libraries, and the other sandbox libraries are the experimental ones (and as Glas shows, it is quite possible that there is even a dependence to another experimental library).
My major concern is about how to encourage continued innovation. Having potential libraries languish in the review queue for many many months, in some cases, years, can't be a good thing for getting people to want participate in boost. Think about all the people behind the scenes that see the review queue and say, why bother, its too cumbersome. There are alot of good ideas out there, but they need a place to get some exposure, and to flush them out. Debian has demonstrated that a stable and a non-stable branch can be quite successful. Boost should adopt a similar approach.
Tom Brinkman wrote:
My major concern is about how to encourage continued innovation.
I have to admit that you are probably more competent than me with respect to this question. For me, libraries like Boost.Proto or Boost.Spirit.Phoenix represent continued innovation, but these libraries come from well established boost authors. On the other hand, some of the accepted boost libraries seem to receive only minimal maintenance by their original authors, and making it easier to contribute libraries to boost would probably only increase the number of libraries in minimal maintenance mode. So I guess the recent efforts to encourage the boost community members to spend some maintenance time on libraries they haven't authored is exactly the right thing to do for the current situation.
Having potential libraries languish in the review queue for many many months, in some cases, years, can't be a good thing for getting people to want participate in boost.
Think about all the people behind the scenes that see the review queue and say, why bother, its too cumbersome.
I'm not sure whether you want to highlight the excessive quality criteria for acceptance into boost here, or the lack of review managers.
There are alot of good ideas out there, but they need a place to get some exposure, and to flush them out.
Do you mean "flush" them out or "flesh" them out? But this is probably something were I can see some truth in your statement. But the sandbox is actually quite a nice place to flesh out a potential library. And there is quite a number of libraries in the sandbox that are currently in active flesh out mode, and will stay in that mode for quite some time.
Debian has demonstrated that a stable and a non-stable branch can be quite successful. Boost should adopt a similar approach.
Are you sure that Debian is in any way comparable to boost? It seems to my that their stable and a non-stable branch is solving a completely different problem. Regards, Thomas
participants (2)
-
Thomas Klimpel -
Tom Brinkman