Patch Pilot report

Vincent Ladeuil v.ladeuil+lp at
Fri Dec 18 22:05:09 GMT 2009

>>>>> "jam" == John Arbash Meinel <john at> writes:

    jam> ...

    >> Some raw numbers to put that in perspective: I spent ~14 hours
    >> the first week and ~24 the second one. These numbers came from
    >> the hamster time-tracking application in Ubuntu and since I've
    >> started using it only a couple of weeks ago they may be a bit
    >> wrong as I can't guarantee I switched tasks as rigorously as I
    >> should have, so take them with a grain of salt. This time also
    >> includes reviews for core contributors which are not strictly
    >> part of the pilot's job IMHO but I was in the mood :) Finally, I
    >> think I did a pretty poor job about asking other to do reviews
    >> too :D

    jam> To be honest, something doesn't seem right with our
    jam> process if it takes 50% of a core developers time just
    jam> to keep up with the review queue. And even that, as you
    jam> say, doesn't include reviews done by 3rd parties. (I'm
    jam> sure I did a fair amount of reviewing during that time,
    jam> but I have no time tracking of it.)

I said raw numbers and grain of salt :)

If that didn't include reviews done by 3d parties (and you did a
fair amount as usual), it includes the time I spent coding and
testing on several submissions and that's certainly the most
important part (gut felling says in the 60/80%), I can try to
refine these numbers to at least separate pure review from active
participation to proposals to clarify.


    jam> Now it could mean that things were going very-very
    jam> right, and we just had a lot of stuff being
    jam> integrated. And I *do* think that we landed more stuff
    jam> this month than previously.

That is highly relevant to my eyes but then, it's hard to talk
about merge proposals as if they were all equal... Getting more
contributors has a multiplier effect, I don't fully realize how
this month differs from the others but there are clear
indications that things went different. Whether or not this is
due to more reviewer/helper time is yet to be better understood
and refined.

I mentioned the raw numbers because Robert asked for them on IRC
last week so I used the same source. We may have to compare notes
or you may try to get more tracking yourself to achieve that, but
until we use the same scales, I'd be cautious when comparing and
even there I think we all proceed differently and provides
various kinds of help, more differences that are hard to measure.


    jam> Certainly increasing community contributions is a good
    jam> thing. Especially if the time spent is beneficial. But a
    jam> contributor spending 10 min causes you to spend an hour,

Yes, that will be extremely bad and a dead end obviously. But is
that a good representation of the reality ?

    jam> then our balance is out of whack. (It becomes too easy
    jam> to DOS the patch pilot. :)

Hehe, I'm so DOSable :)

    jam> Anyway, I think you did a great job, 


    jam> I'm just wondering what we can do to get reviews flowing
    jam> smoothly, and still not consuming too much time.

I think less participation is the most obvious way to restore a
better balance and I know *I* can improve many things there and
I'm confident that I will. Others will proceed differently
too.. But as part of the experiment there were many areas I
wanted to explore...


More information about the bazaar mailing list