On Fri, Sep 24, 1999 at 05:53:35PM +0200, Ivan Popov wrote:
> Hello,
> 
> am I the only outside USA (in Sweden) who is interested in running Coda
> with Kerberos for obvious reasons?

Well, the only thing it gains you is getting a coda-token without
retyping your password. I don't believe it gives any additional security
benefits.

> As far as I know it would be perfectly legal to move the code over the
> border - in explicitely human-only-readable form.

Not really, the precedent that is often cited, the PGP source code, was
an officially published book with an ISBN number and such. Also the font
that was used was _explicitly_ not suitable for OCR, and people had to
read through the code line-by-line to verify it was correct (Actually
they used checksums taken from either the book or some US version).

Simply printing out the code and putting it in an envelope is AFAIK not
a legal way of exporting anything that falls under these silly export
laws. And even if it might be, I'm not sure whether it is very smart to
test their flexibility.

> It takes certainly less effort to reinput the source manually than to
> reinvent it.

Actually, the whole krbsupport.c file is only about 500 line of source,
which is for both kerberos 4 and 5 including comments. So the actual
amount of code you'd need to `reinvent' is probably around 100 lines.
And a description of the routines and their intended purpose should be
exportable.

Jan