> So far, the consensus seems to be in the compromise solution: keep the 3.x
> branch targeting .NET 3.5 and 4.x targeting .NET 4.0.
I don't see that. IMHO the consensus is that focusing on .NET 4.0
will limit adaption for that version, unless .net 4 is used everywhere,
which I don't see happening soon (I agree with the description Bohlen gave)
> That does imply a bit of additional work (backporting patches), but it
keeps
> things simple enough. A guideline for submitted patches should be
> established, so the contributors do the work themselves when fixing bugs,
> instead of deferring everything to the committers.
> We could then establish a threshold of .NET 4.0 adoption that will result
in
> killing the 3.x branch, like 90% (I have no idea how to measure it, but
I'm
> wearing my free brainstorming hat at the moment. Maybe downloads/month)
Remember that killing a branch or 'leaving it up to the user to
patch things' won't help the acceptance rate among enterprises and large
teams: they don't want to spend time (and thus money) on poking into a
codebase they're not familiar with to get data in/out of a database.
FB
