Walter Stiers from the Academic Relations Team at Microsoft wrote that Microsoft Research is accepting proposals for Semantic Computing. His post caught my eye, because of what I am working on, which can basically be described as semantic computing.
I like the phrase used, "semantic computing." It reminded me of the another phrase coined by the Data Access Team "conceptual layer" to refer to the higher level of abstraction offered by LINQ and expression trees.
My main concern was whether Microsoft invading my turf and whether I would have to put up a fight. It was just a false alarm as Microsoft seems preoccupied with understanding the web to aid in search requests: The program is entitled "Beyond Search" and has two tracks, Semantic Computing and Internet Economics. Microsoft's intent is laid out explicitly in the actual program description, which is to improve search and then milk the advertising revenue: "The total spent by Internet advertisers in 2005 is estimated at $8.3 billion, a growth of 13.3% from 2004."
I noticed that the industry's natural language and other "semantic" efforts seem focused on narrow range of uses such as search engines and command-and-control. Search seems to be in Microsoft's vision because it is demonstrably monetizable and currently dominated by mortal enemy Google. The company seems wary appropriately of putting money on risky black hole projects with poor business prospects.
It's going to be a while before we see deep understanding in every application. Well, it already takes forever for new features to be added to existing applications, where fifty developers add less than fifty features over a couple years.