Agreed on all points...!
On 11/19/15 1:27 PM, Yingyi Bu wrote:
IMO, the investment for stabilization would greatly speed-up future
innovations and experiments (if we manage to fix all the major/critical
issues by the end of that phase).
Everyone could face way less blocking issues and need way less hacks to get
things work or grab experimental results:-)
Best,
Yingyi
On Thu, Nov 19, 2015 at 1:14 PM, Till Westmann <[email protected]> wrote:
My main goal it to get the number of painful issues down and I agree that
that doesn’t require that everything else stops. However, I think that it
is likely take some toll on innovation as some time would be spent on
fixing issues and merging fixes into the feature branches ..
On 19 Nov 2015, at 12:18, Mike Carey wrote:
+1 --- but --- to be clear, I don't think this is proposing an
every-hand-stops-innovating phase. I think the proposal is for a mode
where MASTER is locked down and all folks do try and fix their higher
priority bugs on a weekly basis - but that folks who're working on separate
things (e.g., spatial index performance or BAD approaches to data handling)
would still do that, just not in master (which isn't where it's happening
anyway). Master would be closed for business until all "Major" and above
bugs are indeed fixed. (Because that's where we'll cut release branches
from, and that needs to be stabilized.) @Till, is that a correct
understanding? Thoughts?
On 11/19/15 10:49 AM, Ian Maxon wrote:
+1. Our tests need a lot of work, so putting new features on the back
burner will definitely help with getting that work done.
On Thu, Nov 19, 2015 at 10:43 AM, abdullah alamoudi <[email protected]>
wrote:
++1
Amoudi, Abdullah.
On Thu, Nov 19, 2015 at 10:38 AM, Yingyi Bu <[email protected]> wrote:
+1!
Best,
Yingyi
On Thu, Nov 19, 2015 at 10:32 AM, Till Westmann <[email protected]>
wrote:
Hi,
there are a number of discussions (on and off list) and other
indicators
(people on the users list not getting ahead) that we have a relatively
big
number of issues that affect both the usability (and adoption) of the
system and the productivity of development.
I think that it would be good to move into a stabilization phase for a
while. In such a phase we would focus on addressing known issues and
only
add new features to master is there is broad agreement (on this list)
that
the feature is an exception. The goal of the phase would be to address
all
issues of priority “Major” and above (with the option of agreeing on
de-prioritizing issues …) and to increase test coverage.
Thoughts? Concerns? Questions?
Thanks,
Till
P.S. I think that the 0.8.8 release should not be affected by entering
such a phase.
