W3C home > Mailing lists > Public > public-rif-wg@w3.org > January 2006

[RIF][UCR] Suggestion for new Abstract UC

From: John Hall <john.hall@modelsys.com>
Date: Mon, 23 Jan 2006 21:03:32 +0000
Message-ID: <43D544A4.9030006@modelsys.com>
To: W3C RIF WG <public-rif-wg@w3.org>
Hello all,

I'd like to propose a new category of use case: "Interchange of 
Human-oriented Business Rules". A RIF is needed for interchange between 
tools that support rules of this kind.

An example of a human-oriented business rule is (from the EU-Rent case 
study used in SBVR):

    "A EU-Rent car must not be handed over to a driver who appears to be 
intoxicated."

This rule can be expressed formally, but will not be automatically 
applied (at least, not with the technology currently in use in car 
rental companies). It will be applied by people who work in EU-Rent.

There are two important aspects of this. First, people who apply rules 
need decision support. For example, EU-Rent staff would need criteria 
for judging whether a driver is intoxicated; SBVR (and, I think, 
REWERSE) would categorize these as structural rules.

Second, people don't always do what they are supposed to, so enforcement 
will be needed. This may have two levels:

-     Penalties applied to the organization if it breaks rules. If 
EU-Rent hands cars over to drunks, will it be fined, closed down, 
charged as an accessory to criminal offences? We could call this 
'external enforcement'.

-     Penalties applied within the organization to the responsible 
person. If a person in EU-Rent hands a car over to a drunk, will he/she 
be reprimanded, be fired, become ineligible for legal support from 
EU-Rent in the event of criminal charges? We could call this 'internal 
enforcement'.

For both types of enforcement, there may be rules that escalate the 
penalty for repeated offences.

[I know - my example rule has a deontic operator and enforcement needs 
higher-order logic. But I'd like to postpone those discussions until 
after I've convinced the WG that this category of use case is needed.]

So, why is a RIF needed?

First, tools are needed to store, organize and quality-assure 
human-oriented business rules. Most businesses have lots of these rules, 
originating in different parts of the business, and many of them are 
inter-related. My example rule affects EU-Rent's terms and conditions 
for car rental, its employment contracts, its insurance contracts, and 
its regulatory compliance policies, to name just the obvious ones. 
Businesses need tools (better than Word and Excel) to integrate the 
rules, to categorize them, and to support analysis of overlap, 
redundancy and conflict - which requires their interpretability in 
formal logic.

The tools themselves are outside our scope. The industry is developing 
them in response to SBVR (on which, along with Donald Chapin, I 
represent the OMG in the RIF WG). But the WG needs to ensure that the 
rules and the required metadata can be represented in the RIF.

Second, businesses are increasingly reliant on external sources. For 
example:

-    EU-Rent could buy or license a standard car rental contract 
(including a "no drunk drivers" rule) from an industry association, and 
adapt/extend it to meet EU-Rent's specific requirements.

-    For judging intoxication, EU-Rent could adopt criteria from a 
recognized authority. These criteria would support a defence stronger 
than one based on criteria that EU-Rent had made up for itself, if an 
aggrieved customer were to sue EU-Rent for being over-zealous or if the 
police were to prosecute EU-Rent for being under-zealous.

-    EU-Rent does not have to research the external enforcement rules 
(which may vary between countries and states) for itself. It could: 
share the costs and effort by participating in a trade association; buy 
the rules from an expert source; obtain them directly from regulators.

-    When EU-Rent has decided on internal enforcement rules, it could 
check them out with trade associations (for comparison with peer 
practice), employee unions, regulators and other interested parties.

This is where a RIF is needed - for interchange of rules between the 
tools. The interchanges may be between an organization and external 
bodies, or between different units within a distributed organization.

Regulatory compliance (on which, along with Said Tabet, I represent the 
OMG in the RIF WG) is currently a high-profile area for human-oriented 
business rules. But contracts (sales, purchasing, services, employment, 
leases, partnerships) and operations (organization roles and 
responsibilities, workflow) are also important. There is a lot of scope 
for business application.

There are also two things to think about for the bigger picture. One is 
that the business intent expressed in the human-oriented business rules 
is the source of many of the automated rules that are more 'mainstream' 
to the RIF WG.

The other is that many human-oriented business rules can be more-or-less 
directly automated. For example:

    "If a rental car is dropped off at a location other than the agreed 
return location, a penalty charge must be added to the rental cost"

There is the basis here for a continuum from business motivation to 
deployment of automated rules, but we should probably postpone 
discussion of it until after we have established the basic requirements 
for supporting human-oriented rules.

Regards,

John
Received on Monday, 23 January 2006 21:03:32 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:33:26 GMT