>>>>> "Dan" == Dan Price <dp at eng.sun.com> writes:
Dan> (a) not reestablishing the practice of having large piles of
Dan> asynchronously changing free software in OS/Net (i.e. I believe
Dan> that 'sendmail' is grandfathered, not a precedent).
This is the first time I've heard this being a concern, and I have to
wonder why exactly we care. Is it a concern about quality, that we
won't be able to do the necessary level of code review and testing
before updating to a new upstream release?
Dan> To reiterate, I'm interested mostly in the architectural
Dan> cleanliness of the project, and that things are done with
Dan> appropriate rationale-- number of commits, number of customers
Dan> requesting, and even backportability are in my mind at best
Dan> secondary concerns.
They need to be balanced. If you focus on architectural cleanliness to
the exclusion of other factors, you can easily end up with great
software, but a non-viable business. Sometimes compromises are
necessary.
I'm not sure this discussion is going to make much progress as long as
we talk about "cleanliness", which is abstract and subjective. What are
the real-life problems that Roland's current approach introduces? Are
they important enough to require fixing prior to putback, or are we just
subjecting the project to creeping perfectionism?
mike
