On Tue, 28 Jun 2016, Paul Jakma wrote:
- Doing it in a batched way, so there's X time (e.g. 1 or 2 weeks) of
the boring, tedious, annoying patch-herding work; then Y time (maybe
also 1 to 2 weeks) of 'quiet' time for the integrator where /others/
do work (reviews) and the integrator can do something more
interesting, before doing the final collating based on review comments
to create the tentative 'accepted' - another little 'quiet' period
for testing, before final pushing.
- Rotating it around. If there's 3..5 others, then each potentially only
needs to do the intensive 'on roundskeeper duty' thing - that will
suck 2 to 4 weeks of time over the course of 4 to 6 weeks - once or
twice a year.
Oh, and the 'queue, review, decide, test, accept' pipe-line of each
'round' can itself be over-lapped with the next round - when there's
more people working on this. The early queueing up of r-(N+1) can start
while the round-N is in the final stages.
So even if each round is 4 to 6 weeks, we might get still get through
rounds on a 2 to 3 week basis perhaps.
That's what has driven us to here. I simply can not fail to mention
the facts of this, because this is the _prime point of disagreement_
between us.
And I tried just to gently pushback on the majority voting stuff
initially, however to no avail. The more that wasn't heard, the more
explicit I had to get on why this bothers me particularly in this
context.
Sorry. But hey.
regards,
--
Paul Jakma | [email protected] | @pjakma | Key ID: 0xD86BF79464A2FF6A
Fortune:
I have ways of making money that you know nothing of.
-- John D. Rockefeller
_______________________________________________
Quagga-dev mailing list
[email protected]
https://lists.quagga.net/mailman/listinfo/quagga-dev
