Browse Prior Art Database

Combining combinatorial models while maintaining full coverage

IP.com Disclosure Number: IPCOM000237570D
Publication Date: 2014-Jun-25
Document File: 4 page(s) / 31K

Publishing Venue

The IP.com Prior Art Database

Abstract

A common situation in combinatorial models is one in which a model consists of several sub-models, combined into one. When combining such models, it is often the case that one wishes to separately compute Combinatorial Test Design (CTD) test suitesfor some of the sub-models, and use that result for the test suite of the large model. In such cases, full interaction coverage is not guaranteed unless restrictions between the sub-models are taken into account. This publication presents a method for guaranteeing full coverage when performing CTD in several phases.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 42% of the total text.

Page 01 of 4

Combining combinatorial models while maintaining full coverage

Combinatorial models consist of attributes, values for them and restrictions on combinations of values that may or may not appear together. An assignment of one value to each attribute, that complies with the restrictions represents an abstract representation of a test. The main usage of combinatorial models is for Combinatorial Test Design (CTD) - an effective test planning technique that chooses a small subset of the test space derived from the model that satisfies certain interaction coverage requirements (e.g. pairwise, which requires that each pair of values appears at least once in the test suite).

    A common situation is one in which a model consists of several sub-models, combined into one. For example, for a healthcare claims system, one could create a model for describing patients, and another model for describing treatments. The joint model, which contains the two sub-models is the complete model for testing the system - each test would consist of a patient claiming expenses on a treatment.

When combining such models, it is often the case that one wishes to separately compute a CTD test suite for some of the sub-models, and use that result for the test suite of the large model. In such cases, full interaction coverage is not guaranteed unless restrictions between the sub-models are taken into account.

    One common case of combining sub-models is for data generation. In this scenario, some of the model attributes represent data while others represent functionality. In such cases, it is often required to separately generate a set of data
(i.e. a test plan consisting only of the data attributes), in order to minimize data creation effort, as well as facilitate data reuse between models. For example, one may wish to generate patients from the patient sub-model, and then reuse these patients for generating complete tests for several different sub-systems, e.g. the claiming subsystem, the report generation subsystem and the billing subsystem.

Prior Art
[1] In a previous patent (filed as docket IL920120028US1), we present a method for reusing parts of existing tests. That patent discusses how to take into account a test suite for a sub-model when generating one for the complete model. However, the method by which the suite for the sub-model is considered is not discussed there.

    In this invention, we propose a method to guarantee full interaction coverage when performing CTD in several phases. We show how to take into account restrictions between the


Page 02 of 4

sub-models, as well as interaction coverage requirements for the

combined models in order to ensure 100% coverage of (valid) requirements when combining sub-models. This is achieved by one of the following approaches:
a) computing a set of coverage requirements to be met by the separately computed sub-models such that if they are satisfied, all valid requirements for the complete model can be satisfied; or
b) g...