Re: Is Systems Software Research Irrelevant?

[Sundar Narasimhan <address@hidden>]

   Adam Bosworth says more about asynchronous communication and defined 
   choreography
   in this interview:

   http://www.xmlmag.com/upload/free/features/xml/2001/06jun01/ab0103/ab0103.asp

Hi, Dan: I couldn't get to this link, but know of Adam's ideas --
although I don't know if anything real has come of them -- not
counting the BEA acquisition. (He's usually too high level for me to
find actually useful, but then we don't hack XML/HTTP/SOAP here for
everything :)

To relate this thread back to stuff that might be of interest to ll
members, I wonder if the "code->semantics" movement (i.e. UML,
choreography files etc. that seek to say stuff about what the
programmer/application intent is/or should be :) can be merged with
the "code->machine" movement (i.e. how can we do GC, VM mapping such
as the AWE extensions in Windows, optimizing compilers - hotspot,
better JITs etc). It would seem to me that we are in a funny place
right now.. the code that we write -- attempts to abstract away from
what the machine can do (because language designers find the machine
level too horrible to think about or expose to users of their language
:), and yet the code is not at a level where we can do useful stuff
with it that business or consumers care about (so we continue to
invent these other crutches such as UML, XML schemas, WSDL files or
whatever). No wonder some of the problems mentioned in the list exist
(such as students not understanding tail recursion right, or fresh
programmers not understanding key business goals right).