New report feature for EuroVO Registry

[Markus Demleitner]

Hi Costas,

On Tue, Mar 14, 2017 at 06:33:50PM +0200, Kostandin Vrioni wrote:
> I understand that some of these "anomalies" are not actually anomalies.. and we are thinking to not remove but to consider them as "warnings".
> 
> 
> I have grouped them as follow:
> 
> 
> 
> Registry Anomalies Report
> 
> 
> RoR registry resource anomalies
> ROR-01 :: Common managed authority
> ROR-02 :: Common managed authority prefix

That should be a warning only -- there may be legitimate reasons for
doing this.

> ROR-03 :: Identical critical registry info
> ROR-04 :: Same (case-insensitive) registry IVOA identifier
> ROR-05 :: Same (case-insensitive) registry title
> ROR-06 :: Same (case-insensitive) registry short name
> ROR-07 :: Malformed OAI-PMH access URL
> ROR-08 :: Same (case-insensitive) registry access URL
> ROR-09 :: Same (URI-based check) registry access URL
> ROR-11 :: A registry is not declared as a harvestable registry

Again, there are legitimate reasons for that, so that should be a
warning at worst.

> ROR-16 :: Duplicate managed authority in RoR
> 
> 
> RoR registry resource warnings
> -> ROR-10 :: A registry doesn't contain a 'shortName' field
> -> ROR-12 :: XSD-invalid RoR registry resource

That I'd promote to a "full anomaly", as that can have potentially
very bad consequences.

> -> ROR-13 :: Identifier in OAI-PMH namespace
> -> ROR-14 :: Title in OAI-PMH namespace
> -> ROR-15 :: Short name in OAI-PMH namespace

I suppose all of these are the consequence of the same thing (binding
the oai namespace URI to the emtpy namespace).  I'd predict these
always turn up together, so I'd say it's ok to just test ROR-13 and
just remove the other two tests.

> Harvested registry resource anomalies
> HRV-01 :: Common managed authority
> HRV-03 :: Identical critical registry info
> HRV-04 :: Same (case-insensitive) registry IVOA identifier
> HRV-05 :: Same (case-insensitive) registry title
> HRV-06 :: Same (case-insensitive) registry short name
> HRV-07 :: Malformed OAI-PMH access URL
> HRV-08 :: Same (case-insensitive) registry access URL
> HRV-09 :: Same (URI-based check) registry access URL
> 
> 
> Harvested registry resource warnings
> -> HRV-02 :: Common managed authority prefix
> -> HRV-10 :: A registry doesn't contain a 'shortName' field
> -> HRV-11 :: A registry is not declared as a harvestable registry

Agreed.

> Do you agree on this approach? Any suggestion will be welcomed!

Well, I'd like to work out an operational definition for "anomaly
[sc. message]" and "warning".  How about:

:anomaly:
  indicates that some element of the Registry will probably
  fail for some subset of records

:warning:
  indicates that a Registry operator may have inadvertantly expressed
  something that, while by itself perfectly legal, may not match what
  they have intended.

Written like this, I'm not sure I like the term "anomaly" a lot and
would perhaps like to upgrade it to something that is obviously
stronger than "warning".  Hm.

But would you agree with the two basic categories, regardless of
terminology?

           -- Markus