How about just publishing the sha1 hashes on an SSL-protected page? That
would leave it up to the user to decide whether to trust the server's CA and
to get a correct sha1 implementation. One such implementation could be a
previous version of the openssl utility, using the dgst -sha1 command. It
has the chicken-and-egg problem you mentioned, but that's life.
_____________________________________
Greg Stark
Ethentica, Inc.
[EMAIL PROTECTED]
_____________________________________
----- Original Message -----
From: "John Pliam" <[EMAIL PROTECTED]>
To: "Rich Salz" <[EMAIL PROTECTED]>
Cc: <[EMAIL PROTECTED]>
Sent: Friday, March 09, 2001 11:41 AM
Subject: Re: Why no sig for openssl src?
>
> Rich Salz wrote:
>
> > An interesting question. Should it be PGP-signed? Well, since it's an
> > X.509-based system, that wouldn't look great.
>
> Hmmm, but I don't think it would look bad. PGP and X.509 are very
different
> trust models; web-of-trust vs X.500-directory/hierarchical. Both
> are useful in different circumstances.
>
> For example, I "trust" my mod_ssl because:
>
> * It was PGP-signed by Ralf Engelschall
> * Ralf's PGP public key (& fingerprint) is served from
> https://www.engelschall.com/ho/rse/
> * The X.509 site cert for www.engelschall.com is signed by
> Thawte Server CA.
> * The Thawte Server CA Cert:
> (i). Is in my browser (downloaded from https URL).
> (ii). Is in my hard copy of Ross Anderson, et al _The Global Internet
> Trust Register_ (ISBN 0-262-51105-3).
> (iii). Fingerprints in (i) & (ii) match. FYI:
> C5:70:C4:A2:ED:53:78:0C:C8:10:53:81:64:CB:D0:1D
>
> This mixture of PGP/X.509 together makes for reasonable ad hoc
verification,
> so I think there should be no shame for OpenSSL to provide a PGP sig.
>
> Yes, there is always a kind of chicken-vs-egg philosophical question which
> prevents 100% trust, but I try to go through some sort of verification on
a
> piece of open source software which has security implications. Don't you?
> (Ok, I don't *always* do this, but when I can do it, I feel better :-).
>
> > And if it's signed with an X.509 cert, you can only verify with an
outside
> > source, and how many folks have convenient access to software that can
do
> > that?
>
> Most people have a web browser capable of SSL. Under an SSL download
> paradigm, confidentiality and integrity go together, because I think
> that cheapest attack against integrity is to break the symmetric
> cryptography -- and *that* is considered difficult.
>
> Sincerely,
>
> John
>
> ______________________________________________________________________
> OpenSSL Project http://www.openssl.org
> User Support Mailing List [EMAIL PROTECTED]
> Automated List Manager [EMAIL PROTECTED]
______________________________________________________________________
OpenSSL Project http://www.openssl.org
User Support Mailing List [EMAIL PROTECTED]
Automated List Manager [EMAIL PROTECTED]
