Jaroslav Hajek wrote:
On Tue, May 19, 2009 at 8:49 PM, David Bateman <address@hidden> wrote:
Robert T. Short wrote:
Just a question. What are the criteria for a release? I am not arguing
about the release, just want to know what the process is.
I have a few OOP items that need to be done, but the only real thing that
should be finished before releasing 3.2 is some documentation and the clear
classes function. How does this fit in?
Also, Jaroslav seemed to think that nobody liked the stable repository
thing. I didn't get that impression at all. I am not sure we ever agreed
on the front-to-back process, but I thought the basic notion was pretty
agreed on. It seems to me he should just do it in a way that seems to work
for him using the stable repository. If there are problems, we will just
have to do something else.
Documenting the OOP inheritance stuff would probably be a good idea before a
release as well. I was trying to write a chebyshev class inheriting for the
example polynomial class for this with an overloaded filter method to create
chebyshev filters for omega=0:1, but didn't have the time to get very
far....
D.
David,
as I replied to Robert, I see the OOP stuff in 3.2.0 as experimental.
As such, I don't think incomplete docs are enough of a drawback to
further delay 3.2.0. In fact, docs can make it very easily into 3.2.1,
and I really hope I was not too optimistic about the 2 months
interval.
cheers
I struggle a little with releasing undocumented code. Of course, the
OOP stuff IS partially documented, just not the inheritance stuff that
John and I added. Still, it seems unfortunate to not take the extra
time to do SOMETHING. I could probably do something over the weekend
that would suit for an experimental release, and do a proper job by
3.2.1.
That is why I asked the question "what are the criteria for a
release?". If it is simply time between releases, perhaps the
OOP/inheritance stuff should simply wait until the next release. On
the other hand another perfectly good model is feature based. In that
case, the features should be at least reasonably complete, and no
feature is complete without documentation.
I am fine with whatever is customary and normal for this group. I
haven't done much of this open-source stuff before, so don't really
know the normal approach.
Bob
|