Book People Archive

Re: Paul Duguid article: "Limits of self-organization: Peer production and quality"



david said:
>    From what I understand of the current errata system

just as an aside, "the current errata system" is simply
not transparent enough for us to assess its quality...


>    the issues are not in getting enough input.

lots of people say they have found errors in p.g. e-texts,
but almost uniformly say they haven't reported them, so
i believe there's a big problem in "getting enough input".

for a number of reasons, which should be _thoroughly_
explored so they can be overcome, people don't report.
they don't feel a sense of _ownership_ over the e-texts...


>    The problems are in the fact that
>    the input is wrong about 50% of the time

well, what do you expect when you haven't given people
the scans that they would use to verify suspected errors?

_that_ is the #1 problem with the current system.

heck, i think the fact that half the reports are _correct_ is
the striking statistic, since people are working in the dark.

half the mistakes they find are so obvious that they can be
located even _without_ a scan.   so why weren't they just as
obvious to the original digitizers who worked _with_ scans?

people who send in "suspected errors" are doing us a favor,
even if half of those suspicions prove to be unfounded, and
it does a _huge_ disservice to those well-intentioned people
-- the rare people who are actual sending in error-reports --
to do _anything_ that would dissuade them from continuing...

and i can't see that reiteration of this fact (if indeed it is true)
would have any other effect, except to dissuade those people.
really, we need to _encourage_ them, not _discourage_ them.

it's not a question of if the glass is half-full or half-empty,
it's a question of whether or not you want any water at all.

or, for another cliche, don't look a gift-horse in the mouth.


>    and the fact that we have many messages waiting 
>    in a queue for someone to check which 50% they sit in.

again, the system as it exists -- which makes the "checking"
of error-reports to be _so_ difficult (never mind the actual
_correction_ and then _re-posting_ of the corrected e-texts)
-- is a big part of the reason these problems are so thorny...

when you've built a system that requires excessive manpower,
and is unpleasant to operate, is it a surprise no one volunteers?

the whole error-reporting system needs to be rethunk entirely.

-bowerbird