1. The reason the C2SIM standard must address compliance is that government organizations will need to let contracts for C2SIM-compliant systems and they can't do that without a specification of compliance. I don't know that there is a SISO definition for compliance. I've looked into HLA - they expect a compliant system to meet the specifications for the HLA Rules, Interface Specification and Object Model Template (OMT). Having an actual compliance test is a different step that requires somebody to implement testing software (this has been done for HLA). I don't believe we need to do that; only to state what we mean by compliance.

2. By "implementing" an ontology class, I mean that the system that implements it must actually have an internal value for the data described by that class and be able to make use of the data in a significant way.

3. I'll illustrate here a range of compliance descriptions we might consider:

a. We could say that every single class in the C2SIM ontologies and all of their properties *must* be implemented for full compliance. (If we don't define C2SIM compliance in the standard, I guess this would be the default.) But this definition this could work against adoption of C2SIM because there could be classes in the ontologies that define data which may not be meaningful to all systems. For example, in yesterday's meeting we discussed EntityType which could be APP6-SIDC or DISEntityType and concluded they should both be in the ontology; but DISEntityType is unlikely to be used in a C2 system while APP6-SIDC might be of no interest to a simulation. 

b. We could, as I suggested in previous message, identify some classes without which we consider interoperation to be meaningless and identify them in the ontology. This would allow compliance at multiple levels: Core, SMX and LOX.

c. Or, as I also suggested in previous message, provide a minimalist description where to be compliant a system only has to accept messages that match the ontology/schema but not necessarily do anything with them. This seems to be so weak that it is not a satisfactory definition of compliance.

Description b above is the best solution I've been able to think of; but perhaps PDG members will have other good ideas.  It would be very helpful to see some quick feedback from the PDG endorsing one of the above or suggesting another alternative.

Mark

On 10/9/19 3:17 PM, Douglas Reece wrote:
[log in to unmask]">
What is the requirement for defining compliance in the standard? Is this supposed to be a test or suite of tests?
Is there an example in another SISO standard?

Mark, I don't understand what "implementing" a class means here. If one writes a piece of code with a definition of a class called Actor, is that implementing the class?

Doug Reece

On Tue, Oct 8, 2019 at 11:06 PM Mark Pullen <[log in to unmask]> wrote:
The C2SIM standard needs to include a definition of compliance.

Some ideas that have been put forward include (1) require all Core data classes to be implemented (2) choose mandatory classes and include for them an ontology property "mustBeImplemented" (3) require that compliant systems are able to receive Core classes without exception, but not that the systems be able to use that data.

If you like any of these or have another idea to recommend, please respond to this email.

Mark

########################################################################

To unsubscribe from the SAC-PDG-PSG-C2SIM list, click the following link:
https://discussions.sisostds.org/index.htm?SUBED1=SAC-PDG-PSG-C2SIM&A=1


--

Dr. Douglas Reece |  Principal Engineer
VT MÄK  | 150 Cambridge Park Drive, Third Floor, Cambridge, MA 02140
T: +1.857.209.3483  | 
[log in to unmask]  |  www.mak.com

  



To unsubscribe from the SAC-PDG-PSG-C2SIM list, click the following link:
https://discussions.sisostds.org/index.htm?SUBED1=SAC-PDG-PSG-C2SIM&A=1




To unsubscribe from the SAC-PDG-PSG-C2SIM list, click the following link:
https://discussions.sisostds.org/index.htm?SUBED1=SAC-PDG-PSG-C2SIM&A=1