Detecting Cross-browser Issues in Web Applications

Shauvik Roy Choudhary
Georgia Institute of Technology, USA

Cross-browser issues are prevalent in web applications. However, existing tools require considerable manual effort from developers to detect such issues. Our technique and prototype tool - WEBDIFF detects such issues automatically and reports them to the developer. Along with each issue reported, the tool also provides details about the affected HTML element, thereby helping the developer to fix the issue. WEBDIFF is the first technique to apply concepts from computer vision and graph theory to identify cross-browser issues in web applications. Our results show that W EB D IFF is practical and can find issues in real world web applications.