Should vs must in some odrl:LeftOperand

[larsgsvensson]

The usage notes on some LeftOperands are very inconsistent regarding the use of must vs should

In count the note says

Should be a positive integer

When would it be OK to use a negative integer?

In dateTime the note says

Right operand should be a xsd:date or a xsd:dateTime, Value must conform to [iso8601] as represented in [xmlschema11-2]. The use of Timezone information is strongly recommended.

In what cases would it be OK not to use an xsd:date or an xsd:dateTime? And can I use any datatype as long as the representation conforms to xsd?

In delayPeriod the note says

Value must conform to [iso8601] as represented in [xmlschema11-2] as a xsd:duration. This maybe used with the odrl:andSequence operator.

Why is the use of xsd mandatory here but only recommended in dateTime? (And the text should read "This may be used ..."; a common typo throughout the document)

In meteredTime the note starts with

The value must conform to ...

whereas in payAmount the note starts with

The datatype of the right operand...

It would increase readability if all instances where values must conform to certain datatypes were phrased the same way.

[nitmws]

Quick note: as editor of leftOperands I have copied and pasted notes from the ODRL specifications of corresponding/equivalent leftOperands (constraints called there).

Questions 1: were the must, should, could etc applied to the note in the RFC 2119 (https://www.ietf.org/rfc/rfc2119.txt) context?

Question 2: what level of agreement is required to modify the notes: by the Editors of the draft document of by the WG?

[riannella]

A1: Should be the same ;-)
A2: Editors

[riannella]

A1: MUST be the same ;-))))

[riannella]

1. count - removed the note.
2. dateTime - xsd is now a MUST (same for all date/time constraints)
3. fixed wording

commit: a82f282

[larsgsvensson]

Yes, that commit resolves my issues.

Thanks,

Lars

[riannella]

Thanks