Hi Cait!
I posted this some time during the summer. I think it deals with
everything that you asked about. It's long, so get ready!
Before I paste my document below, I should say that Bookshare only
has three requirements for books. At least 90% of page breaks be present
and properly located in the book. And title page and copyright information
present and legible in the book. All the rest is extra, but makes a book
as
perfect as I know how to make it.
For all of you who saw my list of prevalidation instructions before,
just hit the delete button now.
See below.
Mayrie
COMPLETE PREVALIDATION PROCESS
Okay, I did everything that I do to a book to prepare it for
submission to Bookshare today short of reading it and documented the time
it
took. The total minus actually reading the book was four and a half
hours.
One hour of that was spent in recognition that I did while eating and
doing
laundry. Probably not necessary to count that, laugh. But I included it
anyway. So, here you go.
The book had 292 pages including back cover, jacket flaps, preliminary
pages, and, of course, the text of the book. I'll tell you in general what
I
did, and how long it took, then elaborate on the particular process. As
has
been said before, not everyone's process is the same, and there are
probably
at least three ways to achieve any given result. This is what I did with
this particular book, and my process might vary slightly from book to
book,
but here is what I did.
I was using Kurzweil 1000 and one of the find and replace parts can easily
be done in Microsoft word. In Kurzweil the paragraph mark is represented
by
\n. In word, that character is represented in the find and replace
dialogue
by ^p. That might help folks validating using Word instead of Kurzweil.
1.
Scan took 90 minutes
I am using an opticBook 3600 scanner in single-page mode. Scanner
settings
are as follows:
Scan to images, automatic page orientation, gray-scale data, resolution at
300 DPI.
Recognition settings were:
Collumn identification disabled, one page recognized per scan, speckle
removal disabled, Text quality is normal, partial collumns kept,
suspicious
regions kept, blank pages kept, recognition engine is FineReader 8.0,
English will be recognized.
Reading settings:
Line endings will be ignored by the editor and tables will not be
identified.
I do not identify tables in straight fiction because junk sometimes scans
as
a table and is more of a pain to remove that way, more time consuming. I
have to know when I'll need table recognition so I can enable it.
While scanning to images, I am always reading another book that I have run
through this process to catch errors that ranked spelling didn't.
2.
Recognize images took 1 hour.
I do this when off eating, or doing laundry, or sleeping, something that
doesn't require my computer to be doing anything else.
This time may vary a lot depending upon how hardy your computer is, or how
lame mine is.
3.
Save the file under the name of the book. No time taken.
4.
Clean up preliminary pages and confirm accurate page count: 15 minutes
Label: [From The Back Cover] [From The Front Flap] [From The Back
Flap][This
Page is blank.] if any blank pages exist. Read through all preliminary
pages
and correct all scannos.
Determine where the publisher thought page one should go and set an
opperator defined page number there as page 1.
Check that the last page in the book is numbered properly, telling you
that
you do not have any missing or duplicated pages. If the numbers don't
match,
either rescan and insert pages that you missed, or delete duplicated
pages.
5.
Remove headers, protect chapter headings, number and label any blank
pages,
get rid of end-of-line hyphens, and ensure that blank lines at the tops of
pages will be preserved: 30 minutes.
Protect all chapter headings by placing the page number followed by a
blank
line above the chapter heading.
Remove all headers. Do this only after protecting chapter headings, as
very
often the absence of a running header is the only indication of where a
poorly scanned chapter heading should go.
Page down through the document numbering and labeling all blank pages, and
looking at the first word on each page to be sure that it is a complete
word, and reconnect hyphenated words on one page.
On each page beginning with a lower case letter, insert a space before
that
initial lower case letter. This will help later.
6.
Insert page numbers at the tops of all pages: 30 minutes.
Delete all page numbers at the bottom of pages. These don't always scan
at
all, so can't be counted upon to be there in the page numbering for daisy
navigation, and especially in the html of the Bookshare final copy in the
collection.
Insert page numbers at the tops of all pages not already numbered above
chapter headings followed by two carriage returns.
Remove all extra blank lines by using the find and replace dialogue as
follows:
In the "find box" insert \n\n\n\n\n\n (\n is the character string that
will
search for a carriage return.)
In the replace box type\n\n Do this with the replace box remaining the
same,
but with five, then four, then three carriage return symbols each
successive
time in the "find" box. This will get rid of all instances of more than
one
blank line between any blocks of text, or between page numbers and chapter
headings or text on a page.
7.
Remove any extra carriage returns inadvertently inserted by the OCR: 5
minutes.
This involves using the find and replace command 27 times.
In the find box type " " (That is quotation mark followed by space
followed
by quotation mark."
In the replace box type "\n"
This will separate any paragraphs between speakers that might not have
been
separated by the OCR program. This does happen regularly.
Now you are going to look for paragraph marks that shouldn't be there.
You will do this with each letter of the alphabet in lower case.
In the find box type\na (That is backslash followed immediately by the
lower
case letters n and a)
In the replace box type space a that is hit the space bar followed
immediately by the lower case letter a
Replace all.
Inserting a space at the tops of pages before each occurring lower case
letter allows your carefully inserted blank lines between page numbers and
text on the page to be preserved now.
8.
Run ranked spelling: This took 20 minutes with this book.
I started out with a 99.28% accuracy rating.
Correct all scannos as ranked spelling or the spell checker finds them.
9.
At this point I read the book and correct any errors that the spell
checker
or ranked spelling didn't find. Hopefully I catch them all.
10.
Convert to rtf and close the file. No time taken.
11.
In Microsoft Word, Protect page numbers and page breaks, standardize fonts
and margins, and convert em dashes to double hyphens: 5 minutes. (This is
a
generous estimate of how much time taken).
Open the file in microsoft word.
Standardize font and justify margins
Make sure if validating someone else's submission that there are no smart
quotes in the document, making sure that all quotation marks are standard
quotes. Open book tends to produce inaccurate quotation marks in my
experience.
Protect page numbers and page breaks by using the find and replace
dialogue
as follows:
In the find box type: ^m
In the replace type: ^p^m^p
Replace all.
Convert em dashes to double hyphens by using the find and replace dialogue
as follows:
In the find box type: ^+
In the Replace box type: -- (That is two hyphens or two dashes, depending
upon what you call that key to the right of the zero on the number row.)
Save the file.
NOW YOU'RE DONE!
-----Original Message-----
From: bksvol-discuss-bounce@xxxxxxxxxxxxx
[mailto:bksvol-discuss-bounce@xxxxxxxxxxxxx] On Behalf Of My Nickels Worth
Sent: Monday, November 10, 2008 8:46 AM
To: bksvol-discuss@xxxxxxxxxxxxx
Subject: [bksvol-discuss] How to be a black belt validater?
Ok, I am wondering if we can put together a list of things a validater
can/should do to make a book a really good one for the collection. This
is
a sort of companion thread for the older one about being a black belt
submitter.. What are some things which a validater should always do, what
are things which are extras, according to bookshare's requirements, but
which could really improve the book for the collection?
Maybe an experienced validater could give an example of what they do with
a
book from downloading it off step one to uploading it for publication..
Thanks,
Caitlyn
To unsubscribe from this list send a blank Email to
bksvol-discuss-request@xxxxxxxxxxxxx
put the word 'unsubscribe' by itself in the subject line. To get a list
of available commands, put the word 'help' by itself in the subject line.
