s0apgun Wrote:I've been considering learning a programming language for a while now as I haven't really been happy with being an engineer. I don't want to go back to school for a CS degree, so do you think it would be practical to become skilled at a programming language this late in the game?

I've known lots of great programmers that came from all sorts of random education backgrounds including engineering, math, and various applied sciences. A very solid coworker of mine actually had a Doctorate in English Literature and had learned programming on the side.

dizmox Wrote:Though I think you'd need learn a range of different languages and technologies and build up some experience to have much chance of being hired

Being a generalist early in your career is a deathtrap. I know a great C++ programmer than can't break out of being an underling because he spent too much time in his 20s floating around in various tasks and never excelling in one. On the other hand jack-of-all-trades is great if you want to be in pure management or if you want to start a company.

sholum Wrote:That's why I'd learn something more straightforward yet still popular, like Python or JavaScript.

Anyone starting now should definitely focus on one of the web languages. I wouldn't recommend JavaScript as a very first language, but anything like Ruby, Python, or Java would be a great start. Also become familiar with the popular web framework of your language of choice. (Rails, Django, Swing)

If you have aspirations for living in Japan it's probably best to focus on Ruby and Java.

s0apgun Wrote:I've been considering learning a programming language for a while now...

Writing software is not that hard and you can learn that skill relatively quickly. But it's one thing to be able to write the software and another to be a software engineer. My observation (and opinion at the same time) is that it's best to combine the ability of writing applications with expertise in some other field.

For example:
- mathematics: stock exchange are/were paying big money for mathematicians who would write trading software
- electronics/automation: all the embedded stuff
- physicists: somehow they naturally pick up programming
- language learning: look at the stuff that cb is developing – knowing what the problems in language learning are allows him to write things like sub2srs, rikaisama, and others

@tokyostyle: I stand corrected

tokyostyle Wrote:Being a generalist early in your career is a deathtrap. I know a great C++ programmer than can't break out of being an underling because he spent too much time in his 20s floating around in various tasks and never excelling in one. On the other hand jack-of-all-trades is great if you want to be in pure management or if you want to start a company.

The flipside of this, of course, is that if you become too specific, you can sometimes run into problems finding a job, but that's usually an issue later in your career. (And brought on by not continually learning)

Inny Jan Wrote:- physicists: somehow they naturally pick up programming

Physicists usually get exposure to programming when they are in school. Simulations have to be written in a programming language (FORTRAN is popular for that).

dizmox Wrote:@tokyostyle: I stand corrected

Ultimately it's still just my opinion and I probably could have used less definitive wording. Digging deep into things at least once will broaden your knowledge in ways that transcend your specialty. Also vix86 brought up a really important point that is equally relevant.

vix86 Wrote:(And brought on by not continually learning)

Yes, continual learning is imperative. The Pragmatic Programmer, which is a much more coherent and thorough discussion of this topic, recommends that you learn a new language every year. In fact maybe that's really overall better advice. You should strive for a balance of both deep knowledge and broad knowledge, forsaking one or the other would be a big mistake.

If anyone's wondering I ended up finding that this course was the best way for me to learn the basic concepts of machine learning:

http://work.caltech.edu/telecourse (based on the book "Learning from Data")

I like this series a lot, get to practice python a bit while doing the examples too.

Aside from Japanese I'm self-teaching myself chemistry and physics right now.
If it were up to me I'd spend all that time doing Japanese, but because I have exams coming up in physics and chemistry I have to focus on those especially.

This is actually a pretty good set-up, because you have a mixture of subjects you HAVE to do (like the chem/phys) and subjects you do for your own pleasure (like the Japanese). Anyway, I found it hard to spend adequate time on each subject. Some nights I'd study Japanese for 5 hours and barely spend anytime on phys/chem. Or I'd get stuck on one page of phys/chem for several hours and get nowhere. Despite telling myself "to only spend one hour on each subject", somehow it always spiraled out of control.


So eventually- as an experiment- I just set the timer on 10 minutes for each subject. So I would study chemistry for 10 minutes and as soon as the timer went off I'd immediately switch to physics and do that for 10 minutes. As soon as the timer went off after another 10 minutes I'd switch to Japanese. This would go on for a few hours.

It turned into a rotation and I didn't get stuck as much. I actually got a LOT more done. I'm increasing the time-slots for each subject to 20 minutes but anything above 40 minutes per subject doesn't work for me as I get distracted and loose track of time.