Tools aimed at helping users safely navigate the web and safeguard themselves against potential online predators have become reasonably common. Currently there are two families of tools; heuristics analysis tools that test websites directly using automated scripts and programs, and community based tools where users rate websites and write reviews for the benefit of others. In this paper we examine the relative strengths and weaknesses of each technique, whether these techniques are compatible, and how community feedback can be combined with heuristic-based evaluations. In order to do this we conduct a large-scale comparison of the ratings of heuristic and community based tools, and explore novel methods for abstracting key information from user comments, which could be used to add context and nuance to heuristic based ratings. We find that heuristic and community based ratings are highly complementary, and can be combined to potentially guide users to make more informed decisions.

Users and programmers frequently need to move information between applications, including desktop and web applications. Transferring data often involves reformatting strings such as phone numbers or extracting parts from them, but actually performing these transformations typically requires tedious, error-prone operations. For example, programmers must write messy code to parse and reformat strings passed between web services, while ordinary end users must manually reformat strings that they want to copy-paste between applications. In this paper, we show at an architectural level how topes can be used to smooth the flow of data between applications by automatically transforming strings on demand. We have demonstrated the generality and usefulness of this approach by using topes to automatically transform data moving between applications, web sites and web services, thereby showing how topes can make it simpler for both end users and programmers to transfer information between applications.