NSF/TCPP Curriculum Planning Workshop Andrew Lumsdaine Indiana University.

NSF/TCPPCurriculum Planning Workshop

Andrew Lumsdaine

Indiana University

Mechanism and Processes (How?)

Andrew Lumsdaine

Indiana University

It won’t happen all at once

I’m Looking for a Controversy

I remember my first parallel program

I am thinking of a parallel programming paradigm It is already used by millions

of programmers It is what our students will be

doing when they graduate It has a clean mental model So intuitive that no one calls it

parallel programming We rarely teach it (if we do,

we probably don’t teach it well)

And the answer is Client-server

programming

I am thinking of another parallel programming paradigm It is already used by millions

of programmers It is what our students will be

doing when they graduate It has a clean mental model So intuitive that no one calls

it parallel programming We do teach it, but not with

view of concurrency

And the answer is Object-oriented

programming

Real World Programmers Do It Concurrently

For the Skeptics

Mechanism and Process Leverage what we already know how to do

Or what industry already knows how to do Solve the problems that need solving

Avoid gratuitous concurrency Absolutely have to modernize curriculum

It is long past time to say goodbye to “Hello World!\n” Our students spend their first 18 years in a highly

computationally concurrent world (and non-deterministic) Absolutely have to strengthen ties to industry

Not so we can train better workers Innovation is coming from industry

I’m Looking for a Controversy