FpML Issues Tracker
closed
Minor
Always
Validation Rules
Admin
None
Summary
ird-58 is missing an else clause as currently expressed. The problem is that there is no result if the "if" clause isn't satisfied.
The rule today: " Context: CalculationPeriodDates ird-58 (Mandatory) Preconditions: isParametric If rollConvention is a day of the week ( MON, TUE, WED, THU, FRI, SAT or SUN) then the period must be W. "
An alternative corrected expression is below. This has the problem the rule fires even though it isn't really in context: " Context: CalculationPeriodDates ird-58 (Mandatory) Preconditions: isParametric If rollConvention is a day of the week ( MON, TUE, WED, THU, FRI, SAT or SUN) then the period must be W, else true(). "
The corrected rule, which is the proposed solution: " Context: CalculationPeriodDates (complex type)[not (rollConvention = ("NONE", "SFE", "MON", "TUE", "WED", "THU", "FRI", "SAT", SUN"))] ird-58 (Mandatory) Preconditions: isParametric The period must be W. "
Notes:
matthewdr
06/03/08 1:56 pm
”
Context: CalculationPeriodDates (complex type)[not (rollConvention = (“NONE”, “SFE”, “MON”, “TUE”, “WED”, “THU”, “FRI”, “SAT”, SUN”))]
ird-58 (Mandatory)
Preconditions: isParametric
period = “W”.
”
matthewdr
06/03/08 1:57 pm
Agreed at VWG. ISDA to implement in the correct form.
lyteck
07/07/08 2:49 pm
fixed as proposed (using new specs format)
h_mcallister
07/07/08 4:49 pm
The new expression inverts the original sense of the rule.
Formerly we had:
If rollConvention is a day-of-the-week, then the period must be W.
Now we have:
If rollConvention is none of {NONE, SFE, days-of-the-week}, then the period must be W.
(1) This is not equivalent to the original (perfectly intuitive) expression of the rule.
(2) The inclusion of NONE & SFE in the predicate seems to be arbitrary; what is the rationale?
(3) rollConvention and period are direct children of CalculationPeriodFrequency – shouldn’t the context of the rule be CalculationPeriod*Frequency*, rather than CalculationPeriod*Dates*?
lyteck
07/09/08 8:05 pm
– changed context from CalculationPeriodDates to CalculationPeriodFrequency (3)
– expressed condition in English rather than full Xpath
– corrected the “inversion” of the original sense (removed negation on the condition)
mgratacos
07/09/08 10:06 pm
If the rollConvention is TBILL, shouldn’t the period be W too?
lyteck
12/10/08 7:48 pm
ird-58 resolved with issue 838.
last question (re: TBILL) should be addressed in new issue, if needed.