Great to read from someone who has startup ideas and is prepared to learn to program. My two bits worth: It's like a bar stool, best to have three legs.

1. Choose the startup idea that you feel most passionate about. There will be others as you progress, but stick to the first one (unless your customers vote with their money to go for a different one).

2. Pick a widely used, well documented language. My suggestion would be Python as it supports most paradigms and comes with a comprehensive collection of standard and additional libraries. Buy at least strongly recommended book. I recommend the ones from the O'Reilly series.

3. Be prepared to go off and explore particular issues with the programming language as well as your idea. You will probably end up with a lot of small point solutions. I frequently end up with two or more solutions to specific problems before selecting the one to integrate into my key project.

Other considerations:

* It's a good idea to establish test driven approach and incorporate your tests from point 3 in your code libraries.

* Use version control, I prefer git, but others are good as well. It's not so much which one you choose, but to use it consistently.

* If your idea is very web centric, then you might have to seriously look at JavaScript and Node.JS. But JavaScript is a difficult first language.