Error Correction of Project Gutenberg eBooks

Your assistance is hereby requested:


As many of you know, I like to do something around this time

every year to take a new step forward in Project Gutenberg.

As luck would have it, I recently received an email reminder
from one of our volunteers who reads our eBooks out loud for
those who need or want audio eBook versions of our library.

This volunteer was kind enough to keep a log of errors found
while recording one of our classics eBooks out loud and then
sent us that list of errors, and now was following up.

Due to the fact that we receive more errors messages than we
have volunteers to  handle, these errors were not corrected,
which stimulated me to write a request for help on this in a
recent Project Gutenberg Newsletter.

The results were immediate, effective, and continuing.

The new edition, complete with ~23 corrections is online and
has been for a couple days already, and we are still getting
more volunteers for error correction.

This is a great and wonderful thing because the one thing in
the history of eBooks that separates Project Gutenberg is an
everlasting continuing process of improvement.

Hundreds of our eBooks are reissued each year with a variety
of improvements, some technical, some in format and/or style
of presentation, many with various error corrections.



How Good Can An eBook Get?


If we keep this process going for as many years more as this
has been going on already, there is no reason average eBooks
should not be as accurate, or even more accurate, than books
being published on paper.

Some people like to pretend Project Gutenberg eBooks that we
run through certain processes are "perfect," but I think our
own sensibilities tell us this is not the case.

The recent new edition mentioned above is a perfect example,
as it had been through just about all the processes we have,
and yet reading it out loud revealed ~23 more errors.

I would certainly hesitate to bet that our average 250 pages
long book would not have ~23 errors still in it.

After all, 25 errors in 250 pages at only 1,000 characters a
page, would mean the book had 1 error per 10,000 characters,
or that it was 99.99% perfect.

I won't bore you all with numerical details, other than just
a quick mention that the earliest eBook standards were 99.9%
and then The Library of Congress upped that to 99.95%, and a
few years later Project Gutenberg raised it to 99.975% and I
would certainly bet our average eBook that has completed all
our standard processes is at least that good.

However, there is always room for improvement, and that's an
awfully touchy subject for some, but not for CEO Greg Newby,
or for myself, or for a few others who are willing to create
a new Project Gutenberg Error Correction Team.

Believe it or not, we have receives perhaps 10,000 messages,
over 37 years, encouraging us to check certain parts of book
files for errors.

10,000 error messages!!!

We should expect to receive many more in the coming years as
we will have many more readers.



What Makes A Project Gutenberg eBook?


As I said earlier, the greatest difference between Gutenberg
eBooks and all others is in the proofreading.

No one spends as much time and effort on accuracy as we do.

In the end, after virtually all the easy to find eBooks have
been created, there will only be error correction to do, and
translations into other languages, the rest grinding slowly,
but assuredly to a halt, unless copyright trends reverse.

There is a reason that Project Gutenberg is used so greatly,
particularly when compared to the millions of other eBooks--
and that is because we work harder to make them better.

It takes an hour to work over the average book to correct an
already existing list of errors. . .you have to get the book
and then you have to open up in a program that won't leave a
trace behind, the various "artifacts" you often see when the
eBooks have been pumped through ill-mannered programs, and a
final pass to make sure all the margins still fit, etc.

Even then, one of our "Whitewashers" has to go over the book
with a final fine tooth comb that pops out every character--
every single character, even a comma, that changed from what
was in the previous edition, and make sure each one of those
changes was intentional.

It's really not terribly easy to be the last persons to work
on an eBook, and to know that any errors you leave behind or
accidentally create will be there for millions of readers in
the world until, hopefully, the next error checker finds and
corrects them.

It is a great responsibility, but it also carries a greatest
sense of achievement, as you realize all the future readers,
which could be billions, will benefit from your work.


So, I thank each and every one of our Error Checking Team in
great sincerity for their efforts, and at the same time I am
asking for new members for this team to step forward to make
yet one more level of contribution towards creating the best
library humanity has ever seen.


Please be encouraged to forward this message to everyone and
anyone you know who might be interested.


Again my HUGE thanks to you all!!!!!!!


Michael S. Hart
Founder
Project Gutenberg


_______________________________________________
gweekly mailing list
[email protected]
http://lists.pglaf.org/listinfo.cgi/gweekly