RFC Words - Levels

Hi All,

After our discussions about maintaining or not the RFC words and creating
or not a mature model in conjunction with a set of BP levels, I grouped the
BPs by RFC words:

MUST
    Best Practice  1: Provide metadata
    Best Practice  2: Provide descriptive metadata
    Best Practice  4: Provide structural metadata
    Best Practice 10: Use persistent URIs as identifiers
    Best Practice 12: Use machine-readable standardized data formats
    Best Practice 21: Preserve people's right to privacy
    Best Practice 26: Provide data up to date
    Best Practice 29: Use a trusted serialization format for preserved data
dumps

SHOULD
    Best Practice  3: Provide locale parameters metadata
    Best Practice  5: Provide data license information
    Best Practice  6: Provide data provenance information
    Best Practice  7: Provide data quality information
    Best Practice  8: Provide versioning information
    Best Practice  9: Provide version history
    Best Practice 11: Assign URIs to dataset versions and series
    Best Practice 13: Use non-proprietary data formats
    Best Practice 14: Provide data in multiple formats
    Best Practice 15: Use standardized terms
    Best Practice 16: Document vocabularies
    Best Practice 17: Share vocabularies in an open way
    Best Practice 18: Vocabulary versioning
    Best Practice 19: Re-use vocabularies
    Best Practice 20: Choose the right formalization level
    Best Practice 22: Provide data unavailability reference
    Best Practice 23: Provide bulk download
    Best Practice 24: Follow REST principles when designing APIs
    Best Practice 25: Provide real-time access
    Best Practice 27: Maintain separate versions for a data API
    Best Practice 28: Assess dataset coverage
    Best Practice 30: Update the status of identifiers
    Best Practice 31: Gather feedback from data consumers
    Best Practice 32: Provide information about feedback
    Best Practice 33: Enrich data by generating new metadata.

We currently have two groups of BPs to guide the publisher.

Maybe we could, from this two groups, make an exercise to define a more
fine grained set of groups to, in some sense, assert some "quality"
(mature) to a published dataset.

What do you think about this?

Cheers,
Laufer

-- 
.  .  .  .. .  .
.        .   . ..
.     ..       .