An Abstract Domain to Infer Types over Zones in Spreadsheets - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2012

An Abstract Domain to Infer Types over Zones in Spreadsheets

Abstract

Spreadsheet languages are very commonly used, by large user bases, yet they are error prone. However, many semantic issues and errors could be avoided by enforcing a stricter type discipline. As declaring and specifying type information would represent a prohibitive amount of work for users, we propose an abstract interpretation based static analysis for spreadsheet programs that infers type constraints over zones of spreadsheets, viewed as two-dimensional arrays. Our abstract domain consists in a cardinal power from a numerical abstraction describing zones in a spreadsheet to an abstraction of cell values, including type properties. We formalize this abstract domain and its operators (transfer functions, join, widening and reduction) as well as a static analysis for a simplified spreadsheet language. Last, we propose a representation for abstract values and present an implementation of our analysis.
Fichier principal
Vignette du fichier
sas12.pdf (205.83 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-00760424 , version 1 (03-12-2012)

Identifiers

Cite

Tie Cheng, Xavier Rival. An Abstract Domain to Infer Types over Zones in Spreadsheets. SAS'12 - 19th International Static Analysis Symposium, Sep 2012, Deauville, France. pp.94-110, ⟨10.1007/978-3-642-33125-1_9⟩. ⟨hal-00760424⟩
168 View
183 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More