[Css-csts] Procedure and operation version numbers - continued

[John Pietras]

Members of the CSTSWG ---
Last week I sent an email (below) requesting confirmation of the intent
(documented in the Heppenheim MoM) to include version numbers for both
procedures and operations, including derived procedures and operations. 

In thinking a bit more about this, I find myself confused as to how this
would work an whether it's useful, in particular for derived operations.
For example, the Cyclic Report (CR) procedure extends the Unbuffered
Data Delivery procedure, and also extends the baseline START operation.
According to the Heppenheim MoM, the CR START should get it's own
operation version number. Does this mean that the CR START can be used
to extend a procedure or in the creation of a new service-specific
procedure? If so, then perhaps the "CR START" should be pulled into the
Operations Section (4) of the CSTS Framework specification, and in that
case perhaps the "new" START should be given its own name and operation
type (much the same as we do for derived procedures) and not treat it as
a different version of the START.

If it's not the intent to be able to independently use a derived
operation, then each derived "version" of an operation is tied to the
procedure for which that operation is derived, and it doesn't seem
necessary to have a separate version number - it's simply (for example)
the START operation, as derived for the CR procedure. In fact, since
every operation invocation carries the procedure type identifier as part
of the procedure-instance-identifier parameter, it essentially serves as
the operation version identifier.

Please let me know if I am missing some additional meaning for or
usefulness of having each derived operation have its own version number.

Thank you.

Best regards,
John


John Pietras
Global Science & Technology, Inc. (GST)
7855 Walker Drive
Suite 200
Greenbelt, Maryland, USA
20770-3239
Direct:   +1-240-542-1155
GST Main: +1-301-474-9696
Fax:      +1-301-474-5970

-----Original Message-----
From: css-csts-bounces at mailman.ccsds.org
[mailto:css-csts-bounces at mailman.ccsds.org] On Behalf Of John Pietras
Sent: Thursday, April 03, 2008 5:07 PM
To: css-csts at mailman.ccsds.org
Subject: [Css-csts] Procedure and operation version numbers

Members of the CSTSWG --
In Heppenheim, it was agreed that all procedures would have version
numbers assigned to them. References to this decision in the MoM from
Heppenheim are as follows:

- 3 Concepts, Section 4.4: "The procedures should as well have a version
number."

- 4 Guidelines: " A Service shall identify its profile so that with its
version number, one knows the versions of the used procedures and used
operations. Derived/refined procedures and extended operations shall
have their own version number."

- 5 Procedures: "Every procedure (including extended procedures) and
every operations (including extended operations) shall have a version
number."

- 6 Recommendation: " All procedures shall have a version number
documented in table 6-1. A specific statement is required for the
Association procedure. All operations shall have a version number
documented in table 4-1."

At least with respect to the Minutes of Meetings and telecons, this is
the lat time that versioning of CSTS procedures and operations was
mentioned. As of the v0.12 Procedures Definition, the only versions
number for any operation is the one for the BIND operation, and that
definition ties it to the version of the Recommendation in which it
appears ("The version-number parameter shall identify the version number
of the service specification that is to govern this association if the
BIND succeeds"), not an operation-specific version number. There is no
mention of procedure version numbers in the v0.12 Procedures Definition.


Can I assume that it is still the intent to add the operation and
procedure version numbers in a future (draft) version of the Procedures
Definition/Framework? Also, would it make sense to change the definition
of the instance-number parameter of the BIND operation to decouple it
from the version of the Procedures Definition/Framework and make it it's
own version, especially since the BIND operation is very unlikely to
change as the Procedures definitions evolve? (Perhaps this is what the
sentence "A specific statement is required for the Association
procedure." in the Heppenheim MoM item on procedures definition was
alluding to).

Best regards,
John