DocumentCode :
3079363
Title :
Cleanroom: Edit-Time Error Detection with the Uniqueness Heuristic
Author :
Ko, Andrew J. ; Wobbrock, Jacob O.
Author_Institution :
DUB Group, Univ. of Washington, Seattle, WA, USA
fYear :
2010
fDate :
21-25 Sept. 2010
Firstpage :
7
Lastpage :
14
Abstract :
Many dynamic programming language features, such as implicit declaration, reflection, and code generation, make it difficult to verify the existence of identifiers through standard program analysis. We present an alternative verification, which, rather than analyzing the semantics of code, highlights any name or pair of names that appear only once across a program´s source files. This uniqueness heuristic is implemented for HTML, CSS, and JavaScript, in an interactive editor called Cleanroom, which highlights lone identifiers after each keystroke. Through an online experiment, we show that Cleanroom detects real errors, that it helps developers find these errors more quickly than developers can find them on their own, and that this helps developers avoid costly debugging effort by reducing how many times a program is executed with potential errors. The simplicity and power of Cleanroom´s heuristic may generalize well to other dynamic languages with little support for edit-time name verification.
Keywords :
error detection; program debugging; program diagnostics; programming languages; software engineering; Cleanroom; dynamic programming language features; edit-time error detection; standard program analysis; uniqueness heuristic; Calculators; Dictionaries; HTML; IEEE Potentials; Reflection; Semantics; Syntactics; bugs; dynamic languages; error-detection;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Visual Languages and Human-Centric Computing (VL/HCC), 2010 IEEE Symposium on
Conference_Location :
Leganes
ISSN :
1943-6092
Print_ISBN :
978-1-4244-8485-0
Type :
conf
DOI :
10.1109/VLHCC.2010.11
Filename :
5635293
Link To Document :
بازگشت