WAFFELL

WAFFELL is the WYSIWYT and Analysis Framework For Excel-Like Languages. The goal is to provide an infrastructure for developing spreadsheet dependability mechanisms for Microsoft Excel and similar spreadsheet environments (current plans include support for Gnumeric and OpenOffice also). Currently there are three versions of WAFFELL, all implemented in Java and all designed to work with the WYSIWYT front-end for Excel implemented by the Forms/3 group at Oregon State University.

JavaWYSIWYT/WAFFELL 1

This version of WAFFELL was implemented for a course project while I was Oregon State University. It provides support for WYSIWYT and Fault Localization. It supports region inference using three techniques, as well as working without regions (although it only supports updating for the no-regions and discontiguous regions). This system uses an extended version of the CRG model that represents complex cell formulas using multiple formula graphs with dependencies between these formula graphs.

TODO

WAFFELL 1 Development is frozen.

WAFFELL 1.5

Implementation on WAFFELL 1.5 began after implementation on WAFFELL 2 stalled. It is meant as an intermediary step in terms of technology between WAFFELL 1 and WAFFELL 2. It includes support for testing and fault localization, and support for test case generation and test reuse is planned for the new future. It will also support regions (Discontiguous and Contiguous) as well as working without regions. The primary technical advantage of this system over WAFFELL 1 is that the underlying model was changed to a more generic system using sources and targets that can be readily applied to a wider range of formulas that the CRG model.

WAFFELL 2

At this point WAFFELL 2 is mostly a plan. It will support a highly modular pluggable architecture that will hopefully simplify the process of developing new dependability mechanisms. At its core it will include modules for performing basic common analyses, such as data analysis and formula parsing, to which extensions can be added to support a wider variety of tasks, for example interval-based testing or unit inference.