RE: [security-services] Minutes minutes SSTC/SAML concall Tue 21-Oct-2008

["Scott Cantor" <cantor.2@osu.edu>]

> Scott: What's the reason that you're told it's left open in XML-DSIG?

Because other encodings are legal and presumably if they exist, people have
a reason to use them. Exactly the same as with XML. Everybody pretty much
uses UTF-8 most of the time for protocol usage, but people do use other
encodings.

The question is whether implementations can handle others or not. I don't
think mine would because I had to brute force an awful lot of certificate
handling in C/C++. But I got the impression the WG felt that most people's
implementations would handle other encodings, so I assumed mine was an
outlier.
 
> In general, though, this is a potential interop problem that we should try
> to solve, IMO. I guess we don't want to do something like extending the
> ds:X509Certificate element to add an Encoding attribute?

We can't, at least not without a new schema. I believe an Encoding attribute
was discussed as a possible future addition to whatever the eventual output
of the W3C WG is.

-- Scott