Computational Physics: Future's Call ?

Using computers for the physics research has become very commonplace these days. While the software development in commercial arena often involve large teams and is well planned, it not not unusual to find physics packages developed and written by a single author. And software engineering is not a part of usual physics curriculum. So a vast body of the code done by practising physicists usually smack of inefficiency and non-reusability. Codes written without compliance to the sound principles of software engineering are usually hard to read, and debug. There usability is often restricted to the original author,
sometimes even the original context. In this backdrop it becomes imperative for the physics community to learn and start practising the disciplined way of writing the programs...and something of what is called software engineering.