FpML Issues Tracker

806: use of Xpath vs English [exists() vs “exists” / not(exists()) vs “does not exist”]

August 29, 2008

closed

Minor

Always

Validation Rules

Admin

None

Summary

The valwg agreed (8/26/2008) to implement exists (and others) tests as Xpath expressions within Conditions (see issue 613, closed)

e.g., [exists(xyz)] vs "xyz exists" [not(exists(xyz))] vs "xyz does not exist" [count(xzy)>1] vs "at least 2 xzy elements exist"

There are a number of new issues raised to expand the implementation throughout rule definitions. If we agree on this issue, we can close all related issues.

While most rule definitions can fit into the pattern "exists(xyz)", some rule definitions are slightly more complicated. e.g., fx-25 "Two or more fxSingleLeg elements must exist." ird-34 "Either unadjustedPaymentDate or adjustedPaymentDate must exist." cd-7 "effectiveDate/dateAdjustments or effectiveDate/dateAdju... must exist." cd-22 "The following elements must not exist: protectionTerms/creditEvents/bankruptcy, protectionTerms/creditEvents/failureToPay, protectionTerms/creditEvents/repudiationMoratorium, protectionTerms/creditEvents/obligationDefault, protectionTerms/creditEvents/obligationAcceleration." cd-24 "The following elements must exist: protectionTerms/creditEvents/creditEventNotice, protectionTerms/obligations, generalTerms/referenceInformation/referencePrice." cd-36 "Either standardPublicSources or at least one publicSource element must exist. "

Implementing Xpath expressions will certainly result in more precise rules (though more verbose when there are enumerations). That said, this still seems to conflict with one mandate to produce plain English rules or stylized plain English that can also be read by business analysts.

Ref: (4.8 "The working group collaborates with selected experts from product working group to establish an initial set of rules, expressed in plain English.") (4.6 "Validation rules should be written in a stylized plain English which is both precise and complete, but not necessarily formal")

Are the validation rules still accessible to non-programmers? E.g., sum(xyz) vs "the sum of all xyz"

Is there agreement to proceed with implementation?

If we agree on this issue, we can rule for all related issues (and many more not reported yet)

Notes:

  • lyteck

    09/16/08 3:32 pm

    The ValWG (2008-09-16) disagreed to extend the use of Xpath functions to rule definitions. The group prefers for the rule definitions to be expressed in natural/plain english language, easily accessible to business analysts and non-programmers.

    As a result of the decision, related issues requesting a rewrite using Xpath functions are marked as resolved.

  • lyteck

    09/16/08 3:35 pm

    The ValWG (2008-09-16) disagreed to extend the use of Xpath functions to rule definitions. The group prefers for the rule definitions to be expressed in natural/plain english language, easily accessible to business analysts and non-programmers.

    As a result of the decision, related issues requesting a rewrite using Xpath functions are marked as resolved.

  • Leave an update

    You must be logged in to post an update.