1623   GC_CON_range condition not appropriate for rangeC

This tissue has following status: green

Created: 27 Mar 2018

Links: #926 Presence Conditions within RangeConfig

Page: 33

Clause: 7.4.2

Paragraph: Table 30

Category: Issue for edition 2 of this part

Issue: GC_CON_attr is defined like this: "A configuration data attribute shall only be present, if the (optional) specific data attribute
(attr) to which this configuration relates is also present."

Some conformance testers logically interpret the "shall only be present" to mean that "rangeC" must NOT be present if "range" is not present. But apparently many devices include "rangeC" without "range". They need "rangeC" to perform deadband calculations but they do not want to add "range".

There is also basically the same problem for "rangeAngC" and "rangeAng".

Proposal: Change the condition for "rangeC" to Optional.
OR
Change the text for GC_CON_attr from "shall only be present" to "must be present", and add "If (attr) is not present, this attribute is optional". This way you guarantee "rangeC" is present when "range" is present, because it is needed to compute "range". But it may also be present for other reasons.

Discussion Created Status
?
Ballot until Editor
21 Oct 18 green
The attribute rangeC has two components:
- the configuration of the range values
- the configuration of the min and max value

min and max values are not only used for the range. They may be used as well e.g. for deadband. That is, why within rangeC, the attributes for the limits have been made optional (see TISSUE #926).

So apparently, the correct interpretation of GC_CON_attr is, that it "must be present" if attr is there and it is optional otherwise. This has been applied accordingly in Amendment 1, CDV (Condition MO(attr), not MF(attr)).

However, mandating rangeC if range is present while the attributes for the limits are optional within rangeC is meaningless. Therefore, we should use the same condition for the attributes for the limits within the rangeC.

So final proposal:
- Presence condition of rangeC has to be interpreted as already specified in Amendment 1 CDV (i.e. MO(range))
- change presence condition of the limit attributes from AllOrNonePerGroup to MORange
26 Apr 18 final proposal 26.5.18
Change the condition for "rangeC" to Optional is the preferred solution as this is most consistent with Ed1 and Ed2 Amendment 1.


In practice, rangeC is required to implement deadbanding, and db is required to implement range.
27 Mar 18 red

 

Privacy | Contact | Disclaimer