Schneier's source code

Tim May said:
> * we also have fragments of C code accumulated and laboriously
> developed by Bruce Schneier. How many of us have bought the C code book
> and used the code?                                      ^^^^^^^^^^^^^^^

As far as I know, there's the book, Applied Crypto, and the source code on disk.
Are you referring to the C code within the same book, the disk, or a book of
source code with the stuff in the disk (which contains more than the main book)?

It matters to me, because though I have the fat book, I have yet to find the
energy to OCR the source code. As I'm in India, I can't get the disk. If the
contents of the disk are printed in a separate book, I could get *that*.

The disk contains code described, but not presented in the book, including 
DSA, Diffie-Hellman, etc.
-------------------------------------------------------------------------------
Rishab Aiyer Ghosh                            "What is civilisation
rishab@dxm.ernet.in                             but a ribonucleic
Voicemail +91 11 3760335; Vox/Fax/Data 6853410      hangover?"
H-34C Saket New Delhi 110017 INDIA
-------------------------------------------------------------------------------

Rishab Ghosh wrote:

> > * we also have fragments of C code accumulated and laboriously
> > developed by Bruce Schneier. How many of us have bought the C code book
> > and used the code?                                      ^^^^^^^^^^^^^^^
> 
> As far as I know, there's the book, Applied Crypto, and the source code on disk.
> Are you referring to the C code within the same book, the disk, or a book of

That was a typo on my part. There is no separate "book" of code, just
the diskettes.

Someone in Cypherpunks has a novel solution: print code in the most
easily OCRable font---I think the suggestion was that OCR-A and OCR-B,
or somesuch, are optimized for this (one would think so from the
names, but I had thought they had something to do with the magnetic
ink printing on checks...).

I'm sure Schneier would be entertain the idea of a special
"exportable" version of his code in which purchasers paid the $30 he
charges and received a loose-leaf book of very neatly and precisely
printed code, ready for easy OCRing.

And why not make it error-correctable? That is, provide a printed
version that can be scanned, OCRed, then error-corrected for any minor
character recognition errors? It may not be "human-readable," (*) but
it's _printed_ and that meets the letter of the law.

(*) And I can imagine human-readable versions that have the ECC stuff
at the end of each line, or in a block below, or whatever. But this
would not be the standard, of course.


--Tim May


-- 
..........................................................................
Timothy C. May         | Crypto Anarchy: encryption, digital money,  
tcmay@netcom.com       | anonymous networks, digital pseudonyms, zero
408-688-5409           | knowledge, reputations, information markets, 
W.A.S.T.E.: Aptos, CA  | black markets, collapse of governments.
Higher Power: 2^859433 | Public Key: PGP and MailSafe available.
"National borders are just speed bumps on the information superhighway."

"tcm" == Timothy C May <tcmay@netcom.com> writes:

tcm> Someone in Cypherpunks has a novel solution: print code in the
tcm> most easily OCRable font---I think the suggestion was that OCR-A
tcm> and OCR-B, or somesuch, are optimized for this (one would think
tcm> so from the names, but I had thought they had something to do
tcm> with the magnetic ink printing on checks...).

  One of the computer magazines ("Compute"?) in the '80s used to
supply source in a bar code format which was readily scanned into
one's machine using one of those "light wands." I don't know what
export restrictions might apply to this distribution method.

michael

michael shiplett <michael.shiplett@umich.edu> sez:
> tcm> Someone in Cypherpunks has a novel solution: print code in the
> tcm> most easily OCRable font---I think the suggestion was that OCR-A
> tcm> and OCR-B, or somesuch, are optimized for this (one would think
> tcm> so from the names, but I had thought they had something to do
> tcm> with the magnetic ink printing on checks...).
> 
>   One of the computer magazines ("Compute"?) in the '80s used to
> supply source in a bar code format which was readily scanned into
> one's machine using one of those "light wands." I don't know what
> export restrictions might apply to this distribution method.

Something that an unaided human can't read easily might run into
problems.

Why not use a font that's pleasant to read and include a checksum for
each line?  The reduced character set should make errors reading the
checksums themselves less frequent and easier to detect.  No OCR is
perfect so you may as well be prepared to deal with errors.

Stephen

Quoth michael.shipett@umich.edu:
>>   One of the computer magazines ("Compute"?) in the '80s used to
>> supply source in a bar code format which was readily scanned into

Actually, BYTE used to publish things in "BYTEcode", a simple barcode
system (narrow for 0, wide for 1, or something like that, no
modulation of the gap like you find in UPC) and they had articles
spread over several years on how to build simple readers, both
hardware side and software side. (One even involved wrapping the page
around a coffee can, placing it on a turntable, and then having a
latching device to move the wand "up" one "track" on signal from the
computer...  so it could automatically retry bad tracks...) If people
really care to resurrect it, I could go digging, email me if you'd
like me to try.

I don't think BYTE ever had any trouble with exporting it -- but then,
I don't recall ever seeing crypto software in that form. (Carl
Helmers, one of the founders of BYTE, is on the net these days, and
might have useful input...)

Quoth tcmay@netcom.com:
>> easily OCRable font---I think the suggestion was that OCR-A and OCR-B,
>> or somesuch, are optimized for this (one would think so from the
>> names, but I had thought they had something to do with the magnetic
>> ink printing on checks...).

Magnetic ink printing is done with MICR fonts (Magnetic Ink Character
Recognition, or something like that... Under version 10 of the X
Window System, there was a screen font based on MICR. Pretty ugly.)

The OCR fonts really are designed for OCR... I don't recall the
distinction between A and B, I think the latter actually has lower
case as well as upper case :-), but you can find an OCR font for
TeX/MetaFont in one of the standard places (archie CTAN if you don't
have a place to start from...) There are also print-wheels (remember
daisywheel printers?) for the font, and many of the Computer Output
Microfiche services from the 70's and 80's printed all microfiche in
one of the OCR fonts for easy future retrieval.

Anyone out there have experience with modern OCR systems (not the
highest tech Kurzweil units, but something your average hacker could
get cheap for his PC or Mac) and know if OCR fonts are even worth the
trouble these days? I'd guess that a good monospace Courier font would
be just as readable to modern scanners.

After all, Dr. Dobbs (April 1994) has listings for Blowfish encryption
code, in C, in about a 6pt Courier font; I note, however, that they
also have them up for ftp (ftp://ftp.mv.com/pub/ddj/1994.04/blowfish.asc)
so perhaps it doesn't matter how easy it is to scan.

				_Mark_ <eichin@paycheck.cygnus.com>
				... just me at home ...

Oh come on, all this talk about OCR makes it sound like nobody would ever
be willing to just type in the code by hand. It only need be done once,
and the task could easily be divided up for a group. The listings for any
particular cipher just aren't that long.

Phil

Phil Karn says:
> Oh come on, all this talk about OCR makes it sound like nobody would ever
> be willing to just type in the code by hand. It only need be done once,
> and the task could easily be divided up for a group. The listings for any
> particular cipher just aren't that long.

Furthermore, assuming your request to export Bruce's floppy is
accepted, this whole nonsense is over, and assuming its rejected, a
lawsuit can be started to assure that the nonsense will be over with.
I think it pays to spend more time hacking law and less hacking code
in this instance.

I must admit, however, that I may be the origin of some of this stuff.
I was talking about a year ago about printing the PGP sources in a
book in OCR B.

Perry