Hi Jim,

Jim> I have concluded, with regret, that I have to abandon Crypto++ in favor
Jim> of more suitable libraries.  Before I go, however, I would like to offer
Jim> some unsolicited advice.  I offer these in good faith.

Thanks for your contribution and considered opinion on Crypto++.

I agree with everything you've said, and I appreciate the clarity with
which you said it.

For my applications, I just decided to go with the latest version
(5.x) so I didn't experience the conversion problems you mentioned.

Also, I only need a handful of lines of Crypto code, and I found the
filter processing was fine, once I got my head around it.  But having
to reverse engineer the code to figure out what things actually do and
how to actually use them, is a pain.

As for code size, I actually built my own library by copying the
headers and source files from Crypto++ to get just what I needed.  It
took me a day and was really painful.  Logically, it probably couldn't
have reduced the binary size - I mainly did it for stable source code
management purposes.  But, at least I know I only get what I actually
copied.

Anyway, I hope Wei takes on what you've said.  I don't know much about
Wei, but maybe he's hoping that some of his Crypto++ users will take
on some of the challenges you've raised (e.g. documentation).

Certainly, if/when I ever get some spare time, I have had a plan to
start a proper documentation project for Crypto++.

But, in many ways, the serious problems (really, only the code size is
a show stopper - assuming the future versions don't ever obsolete
previous code again) you've highlighted indicate that the library may
have no future unless they can be resolved.

(Just comments, for what they are worth.)

-- 
Russell Robinson (mailto:[EMAIL PROTECTED])
Author of PHP FormMail & Tectite
Download your free CRM from: http://www.tectite.com/


Reply via email to