default IdP metadata for Shib IdP V3

[Tom Scavo]

On Wed, Jan 20, 2016 at 9:38 AM, Cantor, Scott <cantor.2 at osu.edu> wrote:
> On 1/20/16, 8:10 AM, "users on behalf of Tom Scavo" <users-bounces at shibboleth.net on behalf of trscavo at gmail.com> wrote:
>
>>I understand why the back-channel TLS and signing keys are different
>>(because they represent completely different security models such that
>>the compromise of one key does not necessarily imply compromise of the
>>other) but I still don't understand why the signing certificate and
>>the encryption certificate aren't the same. Can you explain?
>
> It's bad key hygiene to use one key for both, just as with TLS, and the implications of changing them (or of one of them being compromised) are very different.

I still don't see it, so I will press for an answer, please bear with
me. I'm not asking about key rollover. I'm specifically asking about
key compromise. I understand how a TLS key can be compromised in
multiple ways (think: HeartBleed) but I don't understand the
distinction you're making between XML Signature and XML Encryption.
Why do I need separate keys for each?

Thanks,

Tom