[Python-Dev] Enhanced tracker privileges for dangerjim to do triage.

[Steven D'Aprano]

On Tue, 27 Apr 2010 03:45:39 am Antoine Pitrou wrote:
> Steven D'Aprano <steve <at> pearwood.info> writes:
> > Who are we worried about offending? The crowds on the Internet who
> > never volunteer for anything, who never submit patches, let alone
> > offer to do the unglamourous work?
>
> Perhaps you should look more carefully. We do have contributors who
> submit patches and advice on the tracker. There isn't just the
> committers and the passive masses.

Yes, I know, I'm one of those contributors who have submitted patches. 
Only a couple, so far, but provided I'm not discouraged by the lack of 
anyone with the time or motivation to review them, there will probably 
be more in time.

(By the way, if anyone wants to review issues 4037 and 8128, I'd be 
grateful.)


> (oh, and following your logic, we should ignore your advice, unless
> you actually contribute to the "unglamourous work" - do you?)

That depends on what you call unglamourous work. No, I don't triage 
bugs. I don't have commit privileges, so I can't. Does hand-holding 
newbies who don't know the difference between a list and a dict count 
as unglamourous work? I'm not looking for a medal, I'm just trying to 
give back whatever little I'm able.



> > In a meritocracy it isn't enough to be
> > good at what you do, you also have to be known to be good.
>
> If this were the criterion then the answer would be simple: nobody
> seems to knows dangerjim in the Python community.

Sean is nobody?

We trust Sean to check in code. We trust him not to hand his username 
and password to dangerjim and let him loose. But do we trust his 
judgement that dangerjim is ready, willing and able to triage bugs?

I think we can all take it as a given that commit privileges shouldn't 
be just given out to anyone. I think we can all agree that one good way 
to gain that trust is to submit lots of high-quality patches. But what 
I don't understand is why there is so much resistance to the idea of 
accepting a personal recommendation from somebody who is trusted with 
commit privileges, even in principle. The reasons given don't strike me 
as convincing, especially the idea that it is nepotistic. It's not like 
commit privileges is a reward that is valuable in and of itself, or 
that they can't be revoked if dangerjim turns out not to have the chops 
that Sean said.

Dangerjim doesn't know Python, he can't contribute by writing patches, 
but he could make a valuable contribution by reviewing them. Sean has 
said he'll mentor him. Meritocracies are reputation-based, and the 
thing about reputation-based systems is that reputation propagates: I 
trust Alice because I've seen direct evidence of her merit, and I trust 
Bob on the basis that Alice vouches for him and I trust her to be a 
good judge of merit.

Such propagation is lossy, of course. I trust Alice more than Bob. The 
further away from direct evidence of merit, the less confidence I have.


-- 
Steven D'Aprano