Closed nigelmegitt closed 8 months ago
In preparation for normatively referencing IMSC-HRM from IMSC,
I am not convinced IMSC should reference IMSC-HRM.
I am not convinced IMSC should reference IMSC-HRM.
That was the basis on which I thought we were doing the work. What alternative(s) do you have in mind?
I am thinking IMSC focuses on semantics and syntax, and IMSC HRM focuses on presentation performance constraints. Organization can choose to adopt only the first without the second, although, evidently, we would recommend they adopt both and could have an informative note in IMSC suggesting as such.
This probably warrants a discussion!
The effect of removing HRM from current IMSC would be that "IMSC conformance" would no longer carry any sense that the document has limited presentation complexity, which is both a significant dilution and an opportunity for future development. We might also remove the constraint on the maximum number of regions in IMSC. I'd like more industry feedback before making such a big change.
The reality is that, until recently, very few actually enforced the HRM, despite being part of conformance, so the HRM being part of the core IMSC recommendation does not guarantee its adoption.
We also have the constraints that existing IMSC editions have their own versions of the HRM.
Finally, as you point out, it is not clear how to avoid circular references if the HRM, which requires IMSC, is referenced from IMSC.
We could always split conformance as a separate document, which is not unusual.
The Timed Text Working Group just discussed Plan to reference IMSC-HRM from future revisions of IMSC?
, and agreed to the following:
SUMMARY: Keep the references circular for the time being, and check that they can all be resolved successfully.
SUMMARY: Nigel to verify that the current text will not cause any issues.
@nigelmegitt See https://github.com/w3c/imsc-hrm/pull/60 which defines what conformance to the the HRM means. This definition can be referenced by other organizations and/or standards as they see fit, without having to normatively reference this specification from IMSC -- an undated informative note from future versions of IMSC might be useful.
I did a pass through the current spec and any references back to IMSC seem fine. Closing with no change needed.
The IMSC-HRM uses terms defined normatively in IMSC, for example "presented region". In preparation for normatively referencing IMSC-HRM from IMSC, we should ensure that there are no circular references that cannot be resolved.