FpML-VAL Tomorrow's Val WG meeting

Pam, I would like to add these points for discussion to the agenda.

1. Syntax for multiple preconditions

a. [if firstPaymentDate and lastRegularPaymentDate both exist]

b. [if both firstPaymentDate and lastRegularPaymentDate exist]

c. [if paymentDates/firstPaymentDate exists and calculationPeriodDates/effectiveDate exists]

d. [if firstPaymentDate exists] [if lastRegularPaymentDate exists] ß preferred?

2. Terminology

a. Use of mathematical expressions: “>=” vs. “greater than or equal to” (eqd-9,cd-37 vs. eqd-13,cd-1b)

b. “If and only if” vs. function iif()?

ird-1,7, shared-1,15 http://www.fpml.org/issues/view.php?id=740

c. “If… then… else” ird-10,11 http://www.fpml.org/issues/view.php?id=697

3. Component ordering ln-7,8,9 http://www.fpml.org/issues/view.php?id=743

- E.g., [n]th and [n+1]th element http://www.fpml.org/issues/view.php?id=766

4. Should functions be shared across asset classes? (i.e., can we streamline definitions and maintain reusable functions that can be reused in multiple validation rules)

- E.g., SameCurrency in Loan, FX, CD, EQD http://www.fpml.org/issues/view.php?id=744

Thanks for everyone’s earlier feedback. There maybe key participant apologies so I don’t think we can through all the items but I think we should be able to find consensus on the easier items.

Lyteck

From: valwg@xxxxxxxx [mailto:valwg@xxxxxxxx] On Behalf Of Peter Geraghty
Sent: Friday, August 01, 2008 5:26 AM
To: valwg@xxxxxxxx
Subject: RE: FpML-VAL validation rules questions

You could require that “exists” was specified for each node in question, so that the word “and” appeared between proper conditions e.g.,

if paymentDates/firstPaymentDate exists and calculationPeriodDates/effectiveDate exists

Alternatively, “all of” would be a more appropriate term than “several” for cases where more than 2 nodes were involved.

Pete

From: valwg@xxxxxxxx [mailto:valwg@xxxxxxxx] On Behalf Of mark.a.addison@xxxxxxxxxxxx
Sent: 01 August 2008 09:18
To: valwg@xxxxxxxx
Subject: RE: FpML-VAL validation rules questions

These changes are all improving the rule in terms of making them more consistent and precise.

However, on the point about using "both," I really do not think it adds anything and is superfluous.
Simply listing each node as a separate condition is sufficient and expresses that both must exist (as Lyteck suggests.)
Consider cases where three, four or more nodes must exist together - we surely would not write several to express the related existence of these nodes?

Regards,

Mark.

"Lyteck Lynhiavu" <LLynhiavu@xxxxxxxx>
Sent by: valwg@xxxxxxxx

31/07/2008 21:07

Please respond to
valwg@xxxxxxxx

To	<valwg@xxxxxxxx>
cc
Subject	RE: FpML-VAL validation rules questions

Thanks Daniel and Christian for the feedback.

See in red below (attached zip contains updated IRD rules)

From: valwg@xxxxxxxx [mailto:valwg@xxxxxxxx] On Behalf Of Daniel.Dui@xxxxxxxxxxxxxxxxxxx
Sent: Monday, July 28, 2008 7:19 AM
To: valwg@xxxxxxxx
Subject: RE: FpML-VAL validation rules questions

Lyteck

Yes, “only” is a word to look for in rules. I think that ird-23 and ird-24 should be reworded to use “if-and-only-if”. [Lyteck: ird-23, ird-24 changed “should only exist if” to “exists if and only if”]

A few other points:

1. Rule ird-9 oddly uses “can exist”. It should probably be turned around to:

Context: InterestRateStream (complex type)
[if calculationPeriodAmount/calculation/compoundingMethod exists]
cashSettlement/cashSettlementPaymentDate must exist

[Lyteck: yes, it makes more sense turned around.

Note: the original rule was “calculationPeriodAmount/calculation/compoundingMethod can only exist if a resetDates element exists.” It looks like I made a copy/paste error in the condition, rule should be:

Context: InterestRateStream (complex type)
[if calculationPeriodAmount/calculation/compoundingMethod exists]
resetDates must exist

ird-9 corrected in SVN]

2. For consistency, I’d reword the definition of the condition “isParametric” to:

Condition: isParametric
(context: InterestRateStream) If cashflows does not exist, or cashflows/cashflowsMatchParameters ~~contains~~ is true.

[Lyteck: ok – updated in SVN]

3. I think also the word “both” also be used in a standard way. I.e. “both X and Y…”.

I suggest rewording Rule-6 as:

ird-6 (Mandatory)
Context: InterestRateStream (complex type)
[isParametric] [hasInitialStub] [if both paymentDates/firstPaymentDate and calculationPeriodDates/effectiveDate exist]
paymentDates/firstPaymentDate > calculationPeriodDates/effectiveDate/unadjustedDate.

[Lyteck: Your suggestion is consistent with some rules (e.g., cd-31, cd-33, ird-36) but inconsistent with other rules (cd-5, fx-3, fx-8, fx14). I think it makes sense to use “both” before listing the nodes.

That said, should we avoid “both” altogether and move toward separating sentences using separate condition tags as it is done for ird-35 (discussed in #4 below) and a few others (e.g., eqd-19, eqd-25)?]

4. What happened to IRD-35?

ird-35 (Mandatory)
Context: PaymentDates (complex type)
[if firstPaymentDate exists] [if lastRegularPaymentDate exists]
firstPaymentDate < lastRegularPaymentDate.
Test cases: [Valid] [Invalid] [Invalid]

Should it not be as follows?

ird-35 (Mandatory)
Context: PaymentDates (complex type)
[if both firstPaymentDate and lastRegularPaymentDate exist]
firstPaymentDate < lastRegularPaymentDate.

Regards

-daniel

From: valwg@xxxxxxxx [mailto:valwg@xxxxxxxx] On Behalf Of Lyteck Lynhiavu
Sent: 25 July 2008 21:33
To: valwg@xxxxxxxx
Subject: FpML-VAL validation rules questions

We would like to add these topics for discussion to the next agenda:

1. Terminology
a. “If and only if” ird-1,7, shared-1,15 http://www.fpml.org/issues/view.php?id=740
o Isn’t this the same as “if” from an implementation perspective?
o Is function iif() bringing clarity

b. “Should only exist if” ird-23,24
o “Should” vs. “Must”
o Isn’t this the same as “if and only if”?

c. Use of mathematical expressions
o “>=” vs. “greater than or equal to”

d. “If… then… else” ird-10,11 http://www.fpml.org/issues/view.php?id=697

2. Should functions be shared across asset classes?
- E.g., SameCurrency, same-currency() http://www.fpml.org/issues/view.php?id=744

3. Component ordering ln-7,8,9 http://www.fpml.org/issues/view.php?id=743
- E.g., [n]th and [n+1]th element http://www.fpml.org/issues/view.php?id=766

Please find attached HTML renditions of the validation rules* in \trunk\src\validation (4.5). Feel free to share any thoughts through this forum before the next meeting.

Thanks, Lyteck

* FYI - Below is a summary of the validation rules that we upgraded to conform to the new specifications. They fall into two categories: (a) validation rules tracked as issues in Mantis or (b) waiting to be noticed and upgraded

a. Logged as issues: The following 34 issues below (related to the implementation of the new validation specs) have been fixed in \trunk (FpML 4.5):

- http://www.fpml.org/issues/view.php?id=616 cd-1 (update)
- http://www.fpml.org/issues/view.php?id=698 eqd-3 (updated)
- http://www.fpml.org/issues/view.php?id=672 eqd-2, eqd-4, eqd-12, eqd-13, eqd-14 (update)
http://www.fpml.org/issues/view.php?id=672 eqd-2b, eqd-4b, eqd-12b, eqd-13b, eqd-14b (new)
- http://www.fpml.org/issues/view.php?id=699 eqd-6 (updated)
- http://www.fpml.org/issues/view.php?id=700 eqd-15 (updated)
- http://www.fpml.org/issues/view.php?id=683 eqd-19 (updated)
- http://www.fpml.org/issues/view.php?id=684 eqd-20 (updated)
- http://www.fpml.org/issues/view.php?id=685 eqd-23 (updated), eqd-30 (new), (pending: eqd-31, 32)
- http://www.fpml.org/issues/view.php?id=686 eqd-25 (updated)
- http://www.fpml.org/issues/view.php?id=701 eqd-28 (updated)
- http://www.fpml.org/issues/view.php?id=702 eqd-29 (updated)
- http://www.fpml.org/issues/view.php?id=759 shared-4 (updated)
- http://www.fpml.org/issues/view.php?id=559 shared-5 (updated)
- http://www.fpml.org/issues/view.php?id=758 shared-6 (updated)
- http://www.fpml.org/issues/view.php?id=757 shared-7 (updated)
- http://www.fpml.org/issues/view.php?id=756 shared-9 (updated)
- http://www.fpml.org/issues/view.php?id=716 shared-18 (new)
- http://www.fpml.org/issues/view.php?id=717 shared-19 (new)
- http://www.fpml.org/issues/view.php?id=719 shared-20 (new)
- http://www.fpml.org/issues/view.php?id=720 shared-21 (new)
- http://www.fpml.org/issues/view.php?id=721 shared-22 (new)
- http://www.fpml.org/issues/view.php?id=722 shared-23 (new)
- http://www.fpml.org/issues/view.php?id=723 shared-24 (new)
- http://www.fpml.org/issues/view.php?id=724 shared-25 (new)
- http://www.fpml.org/issues/view.php?id=613 ird-5 (update)
- http://www.fpml.org/issues/view.php?id=614 ird-6 (update)
- http://www.fpml.org/issues/view.php?id=695 ird-25 (new)
- http://www.fpml.org/issues/view.php?id=696 ird-29 (new)
- http://www.fpml.org/issues/view.php?id=715 ird-48 (new)
- http://www.fpml.org/issues/view.php?id=689 ird-57 (new) (reopened by Harry/ TBILL question)
- http://www.fpml.org/issues/view.php?id=690 ird-58 (new) (reopened by Harry/ TBILL question)
- http://www.fpml.org/issues/view.php?id=734 ln-2 (update)
- http://www.fpml.org/issues/view.php?id=745 ln-7, ln-8, ln-9 (update)
- http://www.fpml.org/issues/view.php?id=585 all (precondition -> condition)

b. Not logged as issues: A second wave of changes was applied as a result of reviewing all the other rules (not logged as issues but requiring refactoring to conform to the new specs. The following 70 rules contained "if" statements now refactored using local <condition>s:

- CD: cd-1b, cd-2, cd-5, cd-8, cd-9, cd-10, cd-12, cd-13, cd-14, cd-15, cd-16, cd-17, cd-18, cd-25, cd-26, cd-27, cd-28, cd-29, cd-30, cd-31, cd-32, cd-33, cd-34, cd-39, cd-41, cd-42, cd-43
- IRD: ird-9, ird-10, ird-11, ird-35, ird-36, ird-46, ird-47
- LOAN: ln-4, ln-5, ln-10, ln-11, ln-12, ln-14, ln-15, ln-16, ln-17, ln-18, ln-19, ln-22, ln-24, ln-26, ln-27, ln-28, ln-30, ln-31
- FX: fx-2, fx-3, fx-8, fx-9, fx-12, fx-13, fx-14, fx-15, fx-16, fx-20, fx-21, fx-22, fx-26, fx-29, fx-30, fx-44, fx-45
- REPO: repo-1