Book People Archive

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

jon said:
>    The question is if the general public using public domain texts 
>    will respond by submitting error reports in sufficient numbers 
>    to make the system worth operating?

to take the last part first, a general-public error-reporting system
costs us nothing, so there's no question that it's "worth operating".

moreover, the important thing is the psychological point that we
have brought the general-public into the operation as _partners_
in the process.   they are owners, with responsibility, of the books.

this sense of "stewardship" has given wikipedia a _vibrancy_ that
project gutenberg -- bless its heart -- has never really attained.

honestly, the "error-reporting system" at project gutenberg now
-- send an e-mail to an anonymous box -- seems designed to
_extinguish_ any error-reports rather than to encourage them;
especially compared to wikipedia, where when you see an error,
you fix it, yourself, right now, and see the fix applied, right now.
(and regulars then monitor the fixes, to guard against vandals.)

and remember that my notion is that an end-user should be able
to click a button and have the original scan displayed right next to
the text itself.   without such a capability, how is a person to know?

besides, i'm not advocating that we do a less-than-stellar job to
start with, knowing that we can "blame" the public for any flaws
that remain long-term.   that's _not_ my attitude, in the slightest.

i strongly believe we need to make the e-texts as perfect as we can
before we turn them over to the general public for "final proofing".

i'm all about writing computer programs to ferret out the flaws,
_before_ the e-text goes out to the end-user.   nicholas hodson
has developed a number of routines that can find mistakes well.
and as a result, the 400 books he has digitized -- by himself! --
are _amazingly_ clean.   (and a huge counterpoint to the idea that
a distributed process is always superior to that of an individual,
on either quality _or_ quantity.)   tools are _extremely_ important.

i went to the p.g. listserves as a programmer volunteering help.
for the first year or so, i posted only to the programmer listserve.
it was when i was treated so rudely there, mostly by d.p. people,
that i decided i'd move over to the _general_ volunteer listserve,
figuring that those antagonists would never act so badly on the
_main_ listserve, in front of an audience of 300+.   (i was wrong.)

the point here is that, with the proper tools, finding errors is easy.
when the tools become advanced enough, removing _all_ errors,
so that we've obtained a perfect text, will be relatively elementary.

for any elusive errors that remain, the public, reading for content
(inside a system that lets them click a button to summon the scan),
will be able to track 'em down and get 'em fixed.   i guarantee it...

(and if nobody ever reads a text, then yeah, it might have errors.
but if nobody ever reads it, those errors do not matter, do they?)

a side-effect of reaching out to _everyone_ in the _general_public_
is that we will get some _programmers_ interested in the challenge.
if we're smart enough to be nice to them (or at least minimally civil),
they can help project gutenberg move all of its texts to perfection...

-bowerbird