the sitka book, which was produced over at fadedpage.com,
was posted over the weekend, so i will finish my version and
do a comparison on it.

the fadedpage/p.g. version is here:
>   http://www.gutenberg.org/files/31862/31862.txt

i can, however, already report that the posted version _does_
contain a few errors.  from page 19 in the original p-book:
>   http://z-m-l.com/go/sitka/sitkap019.png
"yukali" is incorrectly entered as "yuhali", and "prazdnik" is
incorrectly given as "prasdnik".  in addition, both words are
italicized in the original, but unitalicized in the p.g. version,
for a grand total of 4 errors on this single page.  (troubling
is the fact that "yukali" had another occurrence in the book,
and the similarity of these two non-dictionary words in the
book's word-list _should've_ triggered closer examination.)

ironically, there was a note in the forum about this page,
saying that it was marked as "done" incorrectly after _no_
proofing, due to an inadvertent click on a wrong button,
a note which also then pointed out these changes which
needed to be done, but evidently rfrank missed that note.

and with nobody else doing that page, the errors persisted.

plus there are a few other errors i have already noticed...

the word "sheetkah", on page 25, is missing its italics:
>   http://z-m-l.com/go/sitka/sitkap025.png

and the period after "december 14th" should be a comma,
in the footnote (#25) which was on page 80 in the p-book:
>   http://z-m-l.com/go/sitka/sitkap080.png

so there are some things worth pointing out about all of this...

the first is that, by _my_ standard of one-error-per-10-pages,
these 6 errors are _not_ a damning indictment, not nearly so...
in and of themselves, these errors are trivial.  don't forget that.

of course, you don't like to concentrate four errors on one page.
but stuff happens.  so in terms of this particular case, for _me_,
i don't think it's a big deal.  however, in terms of the _workflow_,
the fact that a page with 4 errors can float through the system
without _anyone_ having had a "second look" at it -- or, indeed,
even a _first_ look -- is not a good sign, not a good sign at all.

moreover, according to clear results of a poll i ran over at d.p.,
the majority of people over there believe that 5 errors in a whole
_book_  is right at the maximum that they are willing to tolerate
as "acceptable".  even worse, some of them -- in a defiant act of
wishful thinking -- actually have convinced themselves that they
really do _attain_ that level of accuracy with the books they do!
now, clearly that's ridiculous, and they're just fooling themselves;
they don't actually _know_ how many errors are in their books, so
they just let themselves believe that there aren't any errors there.

but in spite of this break from reality, their _expressed_desire_
is that they release books that have 5 or fewer errors in them...

viewed from that perspective, this performance -- with 4 errors
on a single page and (at least) 6 in a posted book -- is _bad_...

-bowerbird

p.s.  i'd point to the note that was left about that page, except that
rfrank seems to take down the thread for a book once it is posted...
once again, this data is important for a proper analysis of the test.