[PRL] legacy code and AOP

[Matthias Felleisen]

Karl, Mitch, and Felix, but everyone else, too:

Felix and Mitch and I had a conversation about the need of engineers to 
train their students in the maintenance of legacy systems, usually 
written in C.

Larry and I had a conversation with the engineering school about the 
same topic.

In a recent exchange with Karl, I came across this statement:

> The empirical hypothesis beneath AOP is  that these constructs enable 
> the effective modularization of crosscutting concerns. The  ability to 
> pattern match across an entire code base, and to extend and modify it 
> in ways  that are not possible with inheritance is at the heart of the 
> matter.

If this is true and AOP is also about change huge code bases to do the 
right thing _after the fact_, then has anyone thought of developing an 
AOP system for C so that engineers can re-engineer concerns in piles of 
C code and help programmers navigate ill-defined, useful, and 
to-be-maintained systems? If not, why is nobody working on this? Is CS 
really guilty of always solving engineering problems that nobody has 
created yet?

-- Matthias