Category: Issue may impact interoperability of implementations of Edition 2

Issue

In behalf of the WG10 Task Force UML-modeling:
The presence conditions as stated in the current 7-4 edition don't consider the case of derived statistical calculation. Actually depending on many conditions (whether statistical calculation can be apply, or not, or is forbidden) the original presence conditions will have to change in case of instances of LN derived for statistical calculation purpose

Proposal

Rule 1) Group L LNs are excluded (statistics never apply to them). Possibly challenged.
Rule 2) CDCs that may be used in the context of statistics are:
INS, BCR, MV, CMV, SAV, WYE, DEL, SEQ, HMV, HWYE, HDEL, INC, BSC, ISC, APC, BAC.
Add stereotype <<statistics>> to these CDC in UML (on abstract, where applicable) and in the printed doc,
<<statistics>> stereotype is inheritable (we want to print it for every potential subclass).
Add to 7-3 template in the intro somewhere the definition of this property.
Rule 3) in ControllableCDC, add description: "When used in the context of a statistical logical node, ControllableCDC.ctlModel must have value 'status-only'.
Rule 4) Administrative DOs are all the DOs defined in abstract classes DomainLN, StatisticsLN.
Add stereotype <<admin>>: abstract logical node class whose data objects' presence condition:
- does not change between derived and non-derived statistical contexts; and
- becomes not applicable in the derived statistical context in non-statistical subclass.
Agreed to have stereotype <<admin>> for DomainLN and StatisticsLN, not inheritable (we print it for those classes only).
For <<admin>> DOs: all presence conditions remain the same if derived statistic except :
***StatisticsLN.ClcMth with new presence condition MOStat (M if derived statistic, O otherwise)
***StatisticsLN.ClcSrc with new presence condition MFStat (M if derived statistic, F otherwise).
Rule 5) Remaining DOs:
5.a) its CDC is in the list 2) (i.e <<statistic>> stereotyped), derived statistical presence condition is 'O'.
5.b) its CDC is not in the list 2) derived statistical presence condition is "F".
Rule 6) Document printing specifics for presence conditions
Printing in Word: M/O/C -> M-O-C nds / ds (presence condition; non-derived statistical; derived statistical), one column. Update 7-4 template, also with definition of stereotype.
Rule 2) In word: <<abstract, statistics>>;

Discussion

Created

Status

The solution o this Interop tissue will be published in revision 2.1 of part 7-4.

20 Apr 17

In Force (green)

The description of the tissue resolution doesn't reflect the agreed specification (within the TF UML) of the presence conditions of ClcMod, ClcIntvTyp and ClcIntvPer. In edition 2 part 7-4 the presence condition was conditional (what was not correct).
It is necessary to add the new complete presence conditions to the tissue resolution :
ClcMod nds=optional ds=optional
ClcIntvTyp nds=optional ds=optional
ClcIntvPer nds=optional ds=optional
(nds- non derived statistical, ds- derived statistical)

21 Jun 13

In Force (green)

05 Dec 12

In Force (green)

Here is the wording from the UML model, what we want to use in the next version of part 7-4 :
"Definitions and rules for deducing M-O-C ds (derived statistic) in LN tables - from [wg10-0491]:
Administrative DOs are all the DOs defined in abstract classes DomainLN and StatisticsLN (classes have stereotype "admin").
LNs not inheriting from StatisticsLN are not applicable (group L).
CDCs that may be used in the context of statistics are: INS, BCR, MV, CMV, SAV, WYE, DEL, SEQ, HMV, HWYE, HDEL, INC, BSC, ISC, APC, BAC (stereotype "statistics").

4) Rules:
4.a) DOs of "admin" classes (from 1) all maintain the same nds and ds condition.
4.b) DOs of LNs not inheriting from StatisticsLN (from 2): stats condition is 'F'.
4.c) DO's CDC is not "statistics" (from 3), stats presence condition is "F".
4.d) DO's CDC is "statistics" (from 3), stats presence condition is 'O'.

10 Oct 12

Discussion (red)

I do not understand the second bullet of point 4. <<admin>> DOs should be applicable everwhere in the same way. Possibly you mean that the appearance in the derived statistics LN is independent from the appearance in the base LN (which might be statistics or not).

21 Jun 12

Ballot Period

Agreed. Will be considered in the next publications.