From: Bruce Schneier <schneier@winternet.com>
Date: Tue, 19 Sep 95 05:57:40 PDT
To: cypherpunks@toad.com
Subject: Errors in Applied Cryptography, 2nd Edition
Message-ID: <199509191257.HAA09502@icicle>
MIME-Version: 1.0
Content-Type: text



I have taken a three stage approach to reducing the number of
errors in the second edition:

     1.  Correct the errors from the first edition.
     2.  Ensure that I don't create any new errors while writing
          the new material for the second edition.
     3.  Make sure that additional errors don't creep in during
          the production of the second edition.

In order....  I have done a lot more than incorporate all the
corrections from the errata into the second edition.  Minor
things, too trivial to make the errata list, were also corrected. 
The first edition was translated into French, and the translator
sent me hundreds of queries: things that were unclear, ambiguous,
or erroneous.  All of those queries resulted in changes to the
second-edition manuscript.

I was a lot more careful adding new material, reading it over
several times to ensure accuracy.  And a lot of other people read
the chapters over, too.  For the first edition, I asked about a
dozen people to proofread the manuscript for me.  Most people
read the first chapters carefully, and then petered out somewhere
in the middle.  This time I asked about seventy people to each
read a chapter (or two)--a chapter within their area of
expertise.  Each chapter was edited by at least two people, often
more.  Additionally, two people were paid to read the entire book
for accuracy and comprehensibility.

I learned a lot about the publishing process during the
production of the first edition.  For the first edition, I had no
control of any of the production decisions.  The copyeditor added
errors to the manuscript.  Some computer idiot lost all of my
italics, boldface, superscripting, and subscripting while
translating my files from my word processor to their page layout
program.  The proofreader didn't catch errors that the production
people added.  I had almost no time to review the manuscript at
any of the stages.  This time around I was able to choose my own
copyeditor, proofreader, and indexer.  In fact, there were two
separate rounds of copyediting with two different copyeditors. 
The proofreader is not only the most meticulous proofreader I
know, but someone who knows enough math to catch errors in the
equations.  I verified all file translations myself.  And I
reviewed the copyedited manuscript twice, and two separate passes
of page proofs.

I'm not stupid enough to state that the book is 100% error-free,
but I am confident that there are far fewer errors in the second
edition than there were in the first.  I was embarrassed at some
of the errors in the first edition, and I put a lot of effort
into making sure it doesn't happen again.

Bruce

**************************************************************************
* Bruce Schneier
* Counterpane Systems         For a good prime, call 391581 * 2^216193 - 1
* schneier@counterpane.com
**************************************************************************