Recently, a developer showed me a desktop and a mobile interface for the same application. To my surprise, I liked the mobile interface much more than the desktop interface.
After some introspection, it occurred to me that the mobile interface was built in terms of “screens” showing actionable icons while the desktop interface was built with a toolbox and a canvas – I could actually imagine the state machine the mobile app was using, while I could not do that with the toolbox-based interface. I now suspect that user interfaces can be thought in terms of LANGSEC – different interaction paradigms requiring different computational power from users, more computational power in the system being harmful to accessability (this cerrtainly is a no-brainer – have you checked your computational privilege toyay?), usability and security. We may know the adverse effects of “user interface computational dark energy” already, but I am not aware of analysis from a computational linguistics perspective. Mystery meat navigation goes against “full recognition before processing”. Phishing can be considered exploiting in-band signaling (as user interfaces show no distinction between object language and meta language). Side note: Yesterday, I tried to explain the concept “URL” to my mother. She searches almost everything on the web using google – I found that she repeatedly executes “programs” like “search, go to page 3, click second link” instead of bookmarking URLs. She told me that a more complex interface would actually make her more powerful than me who prefers less powerful interfaces – reminding me of arguments I already heard in defense of turing complete template languages like JavaScript. After having written some of the text above I read an article by Olia Lialina about “the turing-complete user”. As I understand it, Lialina does not advocate minimizing computing power of user interfaces: <http://contemporary-home-computing.org/turing-complete-user/> -- Nils Dagsson Moskopp // erlehmann <http://dieweltistgarnichtso.net> _______________________________________________ langsec-discuss mailing list langsec-discuss@mail.langsec.org https://mail.langsec.org/cgi-bin/mailman/listinfo/langsec-discuss
