Re: [oslc-ccm] Review notes config spec, issue 29

["Nicholas Crossley" <nick_crossley@us.ibm.com>]

Martin,

1. I will work on a diagram - it will probably go into the Primer, which I need to write soon!
   
2. I did not write a shape for oslc_config:Configuration because I do not expect any configuration to have just that type; all real configurations would have one of the subtypes (possibly as well as rdf:type = oslc_config:Configuration).
   
3. Fixed - there was a typo that resulted in three missing shapes in the rendered HTML - they were in the .ttl file. I have also now added subsections for each shape so they appear in the index.
   
4. Fixed.

On the points for issue 29, I have applied
all your suggested changes. For the last question, that entire section
is non-normative, so I have reworded that sentence to avoid use of 'must'.

Nick.

From:      
 "Sarabura, Martin"
<msarabura@ptc.com>
To:      
 "OSLC CCM TC (oslc-ccm@lists.oasis-open.org)"
<oslc-ccm@lists.oasis-open.org>
Date:      
 03/09/2017 10:31 AM
Subject:    
   [oslc-ccm] Review
notes config spec, issue 29
Sent by:    
   <oslc-ccm@lists.oasis-open.org>

---

I’m going through the config spec and
had some notes specifically related to issue 29 but also some other notes.
 
General notes:
1.      I really could use
a diagram to illustrate how these resources are conceptually related to
each other. It need not be a complete UML diagram showing all the metadata
– that would be too busy. But for sure something showing the relationship
between component, configuration, baseline, stream, etc.
2.      Where is the shape
for config#Configuration?
3.      The oslc_config:selections
property references oslc_config:ChangeSetSelections which is not specified
4.      https://tools.oasis-open.org/version-control/browse/wsvn/oslc-ccm/trunk/specs/config-mgt/config-resources.html#citemoperations
PUT: …. Might or might not … should be
MAY …
 
 
Issue 29, https://tools.oasis-open.org/version-control/browse/wsvn/oslc-ccm/trunk/specs/config-mgt/config-resources.html#changesets
This issue is resolved by introducing the
concept of change sets and the oslc_config:overrides property, however
the verbiage could use some tightening up – along with points 2 and 3
above.
The configuration defined by a change set
thus identifies both the 'before' state (the base configuration that is
overridden by this change set), …
 
Do you really mean prov:wasDerivedFrom
here? Isn’t this oslc_config:overrides?
When a change set is based on a stream
(the prov:wasDerivedFrom property on an instance of an oslc_config:ChangeSet…
Providers can avoid this lack of precision
by ensuring either that all changes made to a stream are eithercaptured….
… or that the selections
resource of the stream stream’s oslc_config:selections
property …
The bottom half of the paragraph is OK
but it is really tough to parse. I think my recommendations may be enough.
 
Is this normative?
Systems that allow a single change set
to be applied to multiple base configurations must represent
that internal change set by …

Dr. Martin Sarabura
Technical Fellow, Office of the CTO

+1 519.502.4819
msarabura@ptc.com