[Suit] draft-ietf-suit-information-model-00.txt

[Carsten Bormann <[e-mail]>]

I also read the information model draft.

The lead-in up to 3.4 (included) is actually already quite good.
In 3.5, the air gets a bit thin — these are less about requirements than about justifications for some design decisions.
Unfortunately, the justifications are not all holding water.

4 is a very weird information model.  Why is a manifest a list of “fields”?
Some of the fields reference each other — wouldn’t using a tree structure be a better representation of the underlying abstractions (and also be easier to process, as it maps directly into a data model)?
I have no idea from the text what a “Precursor Image Digest Condition” could be.
“Digests” is a map — what are the map keys?
Can’t even parse 4.14.

The whole thing needs to be based more on *types*.  
Types can avoid the repetition between 4.3 and 4.4.
They also are needed to be sure what exactly is *in* a field.
(If you think types is more of a data model term, you can think *terms* instead.)

Grüße, Carsten