jim said:
>   I don't think DP has changed anything else
>   so what I think they have done is simply
>   reduced the rate of books accepted into the system
>   in the first place, increasing the (implied) queue
>   of books digitized but not begun proofreading.
>   What this does, imho, is simply chase away
>   volunteers who want to proof the lower rounds. 
>   But that is probably better than simply
>   wasting those volunteer efforts by having
>   too much stuck on queue in later rounds.

i'm constantly amazed by the people who think they can
"analyze" what's happening at distributed proofreaders
from a distance, by looking at a chart or some statistics.

it's a social system, and thus a lot more complex than that.

but if you wanna reduce it all down to something simplistic,
_this_ is what has been happening at d.p. the past few years:
they have systematically driven away the p1 proofing crew...

the mantra at d.p. used to be "do whatever job you choose".

alas, no more.  you can no longer opt to stay in p1 forever.
much new material is _officially_ made "unavailable" to you,
either permanently or "temporarily" -- for a length of time
that usually means it's finished before it becomes available.
(material that doesn't get finished is typically undesirable.)

perhaps just as bad (or maybe even much worse), if you read
between the lines in the forums, you can tell that volunteers
who choose to stay in p1 are being harassed for that choice
by some other volunteers.  (at times, it even becomes overt;
the harassers operate with the indignation of the righteous.)

so frontchannel and back, officially and via "peer pressure",
volunteers are being chased away from doing p1 proofing...

this is a travesty, pure and simple.  probably even a tragedy.

many of the old-school p1 proofers, who had been doing it
for _years_, with literally thousands of pages of experience
-- even _tens_of_thousands_ -- decided to leave the site...

most of them were proofers who could be counted on to
catch every single error on a page and leave it _perfect_...

it is not inaccurate, in the slightest, to say that some of these
p1 proofers were the very best proofers on the whole d.p. site.

i've done lots of d.p. research, and everything i have looked at
-- everything! -- indicates that p1 proofers _in_general_ are
_just_as_good_ as p3 proofers, and _better_ than p2 proofers.

now, we know that p1 includes some _beginner_proofers_ and
others who are "unclear about the guidelines" (to put it kindly),
so -- to wash out those negatives -- it's patently obvious that
_some_ of those p1 proofers are doing an _outstanding_ job...

and it is those excellent proofers who are being driven away.
simply because they don't want to do p2 (which _bores_ them),
or p3 (even worse!), or formatting in any round (it's _stupid_),
or post-processing (because all they really wanna do is proof).

d.p. should _treasure_ these experienced, skilled p1 proofers.
instead, they are doing shit to make their lives uncomfortable.

and all in the name of "balance" and to "empty some queues".
because p1 proofers churn out the work, and the (admittedly
undermanned) p2 and p3 proofers cannot keep up with them,
and the post-processors can't keep up with any of the others.

but the reason the p1 proofers are being actively harassed is
because they churn it out.  thus "something needs to be done
about the excessive output from p1".  it's just plain ridiculous.

you know what d.p. should do?

send every page through p1 a half-dozen times.  or until it
goes unchanged by 3 successive p1 proofers, which means
it's as clean as it's gonna get, no matter who else looks at it.

indeed, roger's work shows that, for some of the p1 proofers,
the best of them, you don't even need a "confirming" proof...
if they say a page is _good_, odds are that the page is _good_.
but if you want to _really_ be sure, have 3 of 'em say it's good.

that's what d.p. _should_ be doing -- repeated p1.

d.p., stop dissing your p1 proofers!  put 'em to work, dammit!

but hey, guess how many times i have made this suggestion?
_a_lot..._  guess how many times d.p. has listened?  _never..._

d.p. sent itself down this wrong path when juliet decided to
set up a proofing hierarchy.  any "workflow" which demands
that _dozens_ of people (with the most difficult jobs) match
the output of _hundreds_ of people (with less difficult jobs)
who must in turn match the output of _thousands_ of people
(with ostensibly the least difficult jobs of all) is bound to fail.
it's _guaranteed_ to fail.  there's nothing else that it _can_ do.
indeed, it's sheer stupidity, on its face.  pure sheer stupidity.

so now p1 is being brought down to the level of p2 and p3...
and everything looks "balanced", and people think "it worked".
in reality, they have only succeeded in _deluding_ themselves.

but you know what's comical?  d.p.'s most congested segment
-- post-processing -- really could be completely unnecessary!
perhaps you didn't notice the "post-processing" part of when
we just went through those lessons on how to digitize a book?
that's because _my_ workflow -- like any _decent_ workflow --
has reduced that "compilation" part to just a few button-clicks.
but the workflow at distributed proofreaders is outright stupid,
so over there, post-processing is a messy complicated endeavor.

and that has now levied a human cost, on the p1 proofing crew.

the saddest part is that the p1 proofers who are chased away
will never come back to help project gutenberg again.  never.
so the world at large loses extremely valuable contributions.

that's the saddest part.

***

i haven't done many posts on d.p. lately, mostly because i've
completely given up on them.  they have created the seeds of
their own destruction, so their demise is just a matter of time.

the saddest part is they burned out valuable human resources.

-bowerbird