[rpp] Why the limit to RMM level 2?

[Pierre Thierry <[e-mail]>]

Hi,

I started reading the RPP RFCs and as there already is a good level of 
discoverability in the requirements, I was wondering why the level 2 of 
the Richardson Maturity Model was set as a requirement instead of fully 
adhering to REST and using State Transfers in Representations?

As a concrete example, in RPP Core, message deletion goes through the 
hard-coded but relative URI of "/messages/{id}", but the response 
message for message polling could as well contain a "deletion" link.

Section 3.2 from RFC 9205 lists a few advantages of that approach.

I'll note that fully using REST would also accomodate some exisiting 
requirements better than stopping at RMM L2:

  * R9.11/R10.1: representational state transfers make it possible for
    the RPP server to present links to attenuated facets of the
    provisioning objects
      o a read-only view (that needs authenticated access or not,
        depending on the use case)
      o a link to a provisioning process that has single-use
        confirm/abort links
  * R10.1/R10.2: when collections and object operations are provided as
    representational state transfers, a generic RPP UI can present new
    data elements, new object types and even some new operations to some
    extent, even without prior knowledge

Curiously,
Pierre Thierry
-- 

[e-mail]
0xD9D50D8A