Align version-map with EAT sw-version?

[thomas-fossati]

EAT defines a sw-version claim that is semantically equivalent to CoRIM's version-map.

The difference is that EAT uses an array, while CoRIM uses a map.

For reasons of RATS types uniformity, avoiding useless transformations, reuse in implementations, etc., it'd be nice to align.

[nedmsmith]

That will break concise-evidence and depart from TCG Endorsement. I'm just wondering if there is that much value given EAT, CoSWID, and so forth have to translate to an internal representation anyway.

[thomas-fossati]

That will break concise-evidence and depart from TCG Endorsement

:-(

I'm just wondering if there is that much value given EAT, CoSWID, and so forth have to translate to an internal representation anyway.

In implementations, it allows code reuse and avoids remapping types that are trivially equivalent.
Similarly, in specifications, it allows for the creation of a shared vocabulary across same-domain protocols and CDDL reuse.
It also removes cognitive dissonance; see Dionna's comment

[nedmsmith]

It ends up being a question of who's ox gets gored. If you're a verifier then the justification makes sense. If you're an attester and you have committed implementations, then the justification seems really thin.

[thomas-fossati]

It ends up being a question of who's ox gets gored. If you're a verifier then the justification makes sense. If you're an attester and you have committed implementations, then the justification seems really thin.

hmm, the case is more that some implementations may have committed a tad too early, IMO.

Hence my sad-face emoji above.