Early 2007, The Economist ran an article on the logic of privacy--researcher efforts to formalize the logic and instincts most people have for the privacy of their personal information. This seems a commendable approach to bring all companies to a decent level of respect for privacy. However, such efforts seem unlikely to formalize the unpredictable emotional responses people may have in unusual circumstances. As such, privacy modeling software probably will never do a great job of predicting customer responses to new privacy situations.

But the article brings up the interesting discipline of Linear Temporal Logic which uses temporal words like "previously" and "always" as operators. Software, such as Spin has been written or adapted to process LTL. An NYU researcher even developed the Templar language to specify LTL directly, though it appears no compiler exists for this language. The Economist mentions professor John Mitchell at Stanford especially, for his research on privacy and contextual integrity. Mitchell notes at the end of one paper, however, that his work lacks a means for expressings violations of the logic. Legislation usually requires customer notification in this case: obviously, automating such notifications would be of considerable value to a company.

This all seems like fascinating, yet utterly practical work, that awaits a useful software implementation in a web server language such as PHP or JSP.