[bksvol-discuss] Re: Need Help to Confirm Bug In Bookshare Tools

  • From: "EVAN REESE" <mentat3@xxxxxxxxxxx>
  • To: <bksvol-discuss@xxxxxxxxxxxxx>
  • Date: Mon, 06 Jul 2009 02:47:31 -0400

Hi Jill,

Sure, you can see the list of misspelled words. As soon as you upload the book, 
a page comes up with a quality report. Just above where you put the ISBN, you 
will get info on how many words the software found that it thinks might be 
misspelled. There is a link there that you can click which will bring up the 
list of misspelled words. Here is an example:

Quality Analysis
The book you submitted has
63 unique misspelled words
, occurring a total of 419 times out of 17,810 total words (proper nouns, 
acronyms
and jargon are counted as misspellings by the spell checker.)

Evan 

  ----- Original Message ----- 
  From: Jill O'Connell 
  To: bksvol-discuss@xxxxxxxxxxxxx 
  Sent: Monday, July 06, 2009 1:11 AM
  Subject: [bksvol-discuss] Re: Need Help to Confirm Bug In Bookshare Tools


  Evan, Do I understand you to mean that when submitting a book, we can see the 
actual mis-spelled words? If this is correct, how do we do it? 
    ----- Original Message ----- 
    From: EVAN REESE 
    To: bksvol-discuss@xxxxxxxxxxxxx 
    Sent: Sunday, July 05, 2009 6:25 PM
    Subject: [bksvol-discuss] Need Help to Confirm Bug In Bookshare Tools


    Hello All,

    Lissi and I have discovered that in one recent book we did, errors were 
introduced by the Bookshare processing that were not in the original rtf 
submission. These errors take the form of words being split, often in strange 
places--that is, not between syllables. I was checking the quality of a recent 
book we did and discovered that the word "girl" appeared in one place as "gi 
rl", and the word "will" appeared as "w ill". That was in just the first few 
pages. It seems likely that more of these will appear if I keep reading. We 
checked Lissi's proofread rtf file, and neither of these errors were there. I 
had seen this odd word splitting in some other books, some of which I was 
thinking of rescanning, and just assumed they were scannoes that were not 
noticed by the proofreader. (I'm sure that many of you know that lots of books 
get into the collection without a REAL proofreading, even now.) It never 
occurred to me until today that these errors would come from Bookshare's tools. 
It was a bit of a shock, in fact.

    Lissi did mention that when she has clicked on the link for misspellings 
when checking in a book, she has sometimes seen letter groupings tagged as 
errors that she was not able to find in the rtf file, so she was unable to 
correct them. When she was not able to find the misspellings in the file, she 
ignored them, figuring that Bookshare was just wrong. That is not as strange a 
position to take is it may sound, since Bookshare cannot even give a consistent 
word count on the submission form!

    I use brf files, but if these errors are being introduced before braille 
translation, then they would probably appear in the Daisy version of the book 
as well.

    We plan to check some of our other books to see how prevalent this problem 
is, but we can at least say for certain that the Bookshare tools introduced 
errors in at least this book that were not in the file that was checked back in 
after proofreading. It seems very unlikely that this particular book would be 
unique. If this happened to this book, then it is likely happening to others.

    That's what we want your help with. We would be very grateful if some of 
you  out there would be willing to check a proofreading or two of your own and 
see if this problem exists in those books. I think it would be best to check 
proofreading's, not submissions, to eliminate the possibility of errors that 
might have crept in between submission and proofreading. I know this may take a 
bit of time, since not only do you need to find any split words, but then you 
need to compare the Bookshare version with the original rtf file so that it can 
be shown that those errors are not in the file that was checked back in to 
Bookshare. It is probably best to check the most recent proofreading's to 
ensure that any problems you discover were not the result of tools that have 
been superseded. The book we found these split words in was added to the 
collection on June 18. On the other hand, I don't think that huge numbers of 
examples need to be found in each book. If just a few examples are identified 
in several recently added books, that should be enough evidence to prove that 
the Bookshare tools are adding these errors. Maybe, if a few more people can 
confirm this, we can get it looked into.

    Thanks a lot for any help in this investigation. It's a little 
disappointing to think that those of us who do our very best to get the highest 
quality books into the collection might be having our work degraded by a 
software glitch of some kind. We're hoping to get some confirmation of this 
particular glitch so we can quash it as soon as possible to ensure the best 
reading experience for readers. I know most of you here agree on that, 
otherwise you probably wouldn't be subscribed.

    Thanks again for helping us check into this if you can.

    Evan



----------------------------------------------------------------------------



    No virus found in this incoming message.
    Checked by AVG - www.avg.com 
    Version: 8.5.387 / Virus Database: 270.13.5/2220 - Release Date: 07/05/09 
17:54:00

Other related posts: