That is interesting, but I believe that many self-taught programmers never intend to make it a career, but rather a skill to augment their career that also makes an interesting hobby. This is particularly true for pure scientists and engineers. You'd be hard pressed to find one who has not muddled through at least one scripting or programming language. The trick is that it is the principles of the utility of computer software in science are common, and regardless of what language or platform the application runs on, the "user" is usually forced to learn both in great detail to make effective use of the tool, particularly as the complexity of the subect of study grows.
These thoughts tie in to usability and intuitive design. A good scientist knows how things "should" work, and well written software and their asscociated scripting environments take this into account. A scientific user with well written software should only rarely need to consult documentation. Data aquisition rates and data sets grow ever larger in modern science and the software interfaces to that data must also evolve, yet retain intuitive principles of operation.
I recently realised that I have a gift when it comes to software usability and should consider making use of it more actively.
Here's another spin on it: every time I get some great (in my mind) new idea, nineteen times out of twenty someone else has already thought of it or started on it. I look at using software from a design perspective. So the commonality of good ideas guides my software usage behaviour and learning curve. Good software should be completely intuitive to use (considering one's audience of course).
Data exploration and modeling are always what have turned my crank. It is what computers are for. But not just for supercomputer batch jobs written in arcane code for obscure tasks anymore...
On the airy-fairy side of it I have the view that it is all part of the man-machine-matter connection. We are composed of the the very same matter and energy that we manipulate, how can knowledge of all it's properties not be inate? (hmmm, new sig?)
Topic: Great post about autodidacticism and programming (Read 1340 times)