Creating Formal Specifications with Analogical Reasoning


O’Donoghue, D.P. and Monahan, Rosemary and Grijincu, D. and Pitu, M. and Halim, F. and Rahman, F. and Abgaz, Y. and Hurley, D. (2014) Creating Formal Specifications with Analogical Reasoning. PICS - Publication Series of the Institute of Cognitive Science, 1-2014. ISSN 1610-5389

[img]
Preview
Download (339kB) | Preview



Add this article to your Mendeley library


Abstract

We describe the Arís (Analogical Reasoning for Implementations and Specifications) system that uses analogical reasoning to create formal specifications for a given implementation. Arís is built on the hypothesis that structurally similar implementations often represent similar functionality. It leverages this similarity to create new specifications, by analogy to a retrieved similar example. Of course some similarly structured implementations provide different functionality, so a major focus of Arís is to discriminate between analogous and dis-analogous pairs of code. Examples are used to highlight Arís’ ability to create specifications, across a range of similar implementations and even similar algorithms. Results are presented on Arís ability to create verified specifications for a sample of ten textbook problems. We argue that Arís both emulates and supports the workaday little-c creativity of formal software developers.

Item Type: Article
Keywords: Creating Formal Specifications; Analogical Reasoning; ARIS; Analogical Reasoning for Implementations and Specifications;
Academic Unit: Faculty of Science and Engineering > Computer Science
Item ID: 5624
Identification Number: http://ikw.uni-osnabrueck.de/en/ikw/pics
Depositing User: Dr. Diarmuid O'Donoghue
Date Deposited: 16 Dec 2014 17:09
Journal or Publication Title: PICS - Publication Series of the Institute of Cognitive Science
Publisher: Institute of Cognitive Science
Refereed: Yes
URI:

    Repository Staff Only(login required)

    View Item Item control page

    Document Downloads

    More statistics for this item...