MATLAB programming best practices -


i'm looking resources on how structure medium- large-scale matlab projects, ones involve several independent modules. how manage global configuration variables, how structure project folders, how manage couplings between modules, etc.

is there kind of standard text on subject? looks if matlab textbooks have been written scientists or engineers. i'm looking for, guess, matlab textbook written software engineer. :-)

matlab unusual choice large-scale projects , suited such task assembler, cobol or sql. if still choose matlab @ least automatically test code! kind of tests - integration tests, unit tests, load tests! , of course use version control system.

as said, matlab not created large projects in mind therefore best practice regarding project structure, modules, coupling common sense.

if taking on existing large matlab project sorry you, refactoring nightmare. if going start new large project matlab crazy - there better alternatives matlab not bad regarding numeric performance. large project implies code business logic, not numerics, therefore why god's sake matlab?

large project implies structured components, implies oo, the weak point of matlab because sacrifices heap performance numeric performance degree of unusability.

my experience:

  1. i spent years in in half-million loc matlab project.
  2. i have seen painless transition of multiple large matlab projects c#.

with matlab still have use large amounts of java decent looking gui, c/c++ mex fast not numeric parts imports, maybe sql, etc. java (or better c#) free numeric library have 1 language suited need in large project.

i not saying matlab bad - rules rapid prototyping , numerics! , simulink has no alternatives (but can compiled , used everywhere).


Comments

Popular posts from this blog

apache - Add omitted ? to URLs -

redirect - bbPress Forum - rewrite to wwww.mysite prohibits login -

php - How can I stop spam on my custom forum/blog? -