On Fri, Jul 15, 2011 at 18:14, Jeff Graham <Jeff.Graham@brukeroptics.com>wrote:
OK. I missed that thread. It is enlightening. It sort of looks like maybe a shadow of a plan: wait for Phoenix V3 to stabilize then refactor because the interface will likely change in significant ways. That is reasonable, though as others have articulated I too would question the wisdom of holding up a very useful library for TWO YEARS because you know that you will need to refactor some of it in the future. This is the larger question.
I agree, a wiser way (AFAIK from some experienced devs I don't count me in yet ) would have been to first fix the problems that were important but didn't rely on external dependencies, release a first version, then target the next version at using the dependency IF it was out at the time. (I think an additional per-library version would help clarify for a lot of people this kind of process, but that's another discussion - maybe Ryppl would help there?) Now am I wrong in thinking that the decision was made to wait because the transition to Phoenix3 would anyway have cost too much for the library developer(s) if not prepared "early"? Joël Lamotte