Hi all – I would love to hear comments on a set of talks I am presenting to NITheCS.
Here’s brief description of the audience: they are an online community of South African scientists (see https://nithecs.ac.za/). So I expect scientific sophistication but not too much knowledge of Julia.
And here is the outline:
TITLE: The Julia language: easy, generic, fast (but no free lunch)
DESCRIPTION: These four talks on Julia form a general and high-level introduction to Julia as a language for scientific computing.
Talk 1: A showcase of Julia code, development interfaces and achievements, and introducing the claims that Julia solves the two-language problem, and also solves the expression problem.
Talk 2: Julia’s design principles and tradeoffs (covering things like multiple dispatch, the type system, the object model, modularity, scope rules, JIT compilation, code introspection, parallellisation by design, and perhaps others)
Talk 3: The DifferentialEquations organisation (which claims to be by far the best general-purpose system of packages for differential equations).
Talk 4: Rapid development of performant code, massive code re-use: does Julia really solve the two-language problem and the expression problem? And what might this mean for HPC?