Man-Computer Symbiosis

posted Feb 13, 2012, 11:08 PM by Nick Bader   [ updated Feb 13, 2012, 11:08 PM ]
The fundamental vision of Licklider's Man-Computer Symbiosis is of humans and computers complementing one another's weaknesses, forming a whole greater than the sum of its parts.  Licklider focuses on technical hurdles that must be overcome in order to usher in this new era.  I am surprised to note, as I look over this list of hurdles, that we have had remarkably variable progress in dealing with these issues since 1960.  To wit:

Proposed: A machine should be able to deal with incomplete specifications of a problem.
Status: Computers must still be told exactly what to do.  Programmers must still spell out their instructions in unambiguous code.  But maybe this is unfair: perhaps we need to look for improvements at a higher level of organization, like the Bayesian learning algorithms that can be trained to sort your junk mail.

Proposed: Clever tricks must be employed to overcome the inevitable and eternal limitations of computer memory.  (Here Licklider is is contrast to Vannevar Bush's As We May Think, where the limitless storage seemed inevitable, albeit on microfilm.)
Status: Advantage Bush.  Memory is cheap.  Today my colleague handed me a capacious memory stick barely large enough for its own USB jack.

My favorite theme in Man-Computer Symbiosis is the need to find a common ground for communication with machines.  Here we still find a tremendous gulf between human and computer languages.  The computer language Python, lauded for its readable code, is to a first approximation very similar to the Fortran code described in this article.  Human languages are still evolutionary and illogical.  In our last meeting I mentioned a constructed human language (Lojban) designed to be logically unambiguous, but it remains a small project.  Progress in communication with our computers may be the most significant impediment to the symbiosis Licklider envisions.