OK, so this is a bit of a quick, un-thought through answer,
but here goes:
On Mon, Sep 10, 2007 at 09:28:24PM +0200, demerphq wrote:
> What can we change about how we work that will make
this better?
Not rely totally on volunteer developers.
> IMO part of our problem is that the people who
self-select to be
> p5p'er generally dont have good project managment
skills. Its well
> known that few developers do actually, and even less
good developers
> do. Its quite simple, all that time doing stuff to be a
star developer
> is time NOT spent doing stuff that makes one a good
manager/project
> manager.
But the project manager job will also be a volunteer job.
And even if we find a good one, they don't have paid
developers to manage.
People working on core stuff can't realistically commit to
deliver anything,
because any part of their life might suddenly take
priority.
It's best effort all the way as is.
[I'm not sure what I can accurately remember, and in turn
what I can
accurately repeat, about conversations I had some time ago
with the
professional donating his project management abilities to a
similar
volunteer developer project, and being very frustrated by
the
inability/unwillingness/ doesn't-matter-why of the volunteer
developers to
give reliable time estimates and progress estimates. But I'm
not convinced
that getting a project manager alone will solve anything]
Now, I don't know now we might go about getting
non-volunteer developers.
It doesn't seem that many companies can actually commit to
the idea of
offering someone part time to Perl, but then again, how many
have we asked.
And in theory there should be a support market for Perl,
given that firms are
using it for business critical activities. Yet despite
Activestate being in
prime position clean up on that, it being their core
competency, I don't get
the impression that it's a saleable service. So if with
their skills, they
can't get more trade, why will anyone else manage to do
better?
> Brushing this fact under the carpet and pretending that
we dont have
> serious project management weaknesses will just mean
that 5.12 comes
> out in 2016.
>
> Now maybe this is what we want as a community, but i
dont think so.
>
> Its time to start talking seriously about our plans for
the future. As
> far as im concerned (and ive said this before) we
should already have
> a 5.11 branch set up and be working on it.
I'm not *convinced* that we should.
As is, we force the focus of activity to be on getting a
stable release
shipped.
If (as is) we already had a maint-5.10 branch, then Rafael
would be very
much on his own, and everyone else would be working merrily
on trunk (or
elsewhere) quite correctly, so as not to mess his stability
up.
IIRC Linus forced Linux to stability for 2.4 and 2.6 by not
opening up 2.5
and 2.7 until the new "stable" kernel was.
Nicholas Clark
|