opinions on porting software in alpha state?

[Mark Linimon]

On Mon, Mar 06, 2006 at 04:44:46AM +0100, Jean-Yves Lefort wrote:
> On Mon, 6 Mar 2006 01:22:05 +0300
> Dmitry Marakasov <amdmi3 at mail.ru> wrote:
> 
> > I've successfully ported some software products, but now I doubt
> > if I should actually submit these ports, because these programs are
> > in alpha state. They are usable, yes, but as it may be expected
> > from alphas they are quite buggy, and some functions don't work.
> 
> What's the difference with the other software we have in ports?

There are two theories of what should be in ports: 1) only things that
are known well-working and useful; 2) anything that someone might find
useful somehow, somewhere.

There is never going to be consensus on what the model is.  (My conclusion
is based on the last N times this topic was discussed).

I think it is fair that if something is really rough, that the user ought
to be warned somehow, so that an informed decision about whether to install
it can be made.  Users "expect" the ports to work -- that part I'm not
inclined to argue about, I'll just assert it.

So if it's something rough, either it should be in pkg-message, or if it
is known not to work yet, set an IGNORE message and let the user override
that if they choose.

(From past conversations you will be able to deduce that I lean towards
theory 1, but that I no longer believe it is possible to either reach a
consensus or force a policy on this.)

mcl