Hi,

I have known the basics of programming for many years. My problem is that I don't know what to code. When I made my first payroll-program in c++ I remember it was so fun and kind of aqcuired a taste for programming that stuck with me through all this time.
I have these giant ideas for a game which I kind of build upon and draw inspiration from by doing sketches on the drawboard and some 3d-modeling (landscapes, characters and machinery...it's supposed to be sci-fi).
Now the thing is...I have some friends who are in the game industry as programmers and they are pretty experienced. So I ask them for advice - you know - "what the hell should I code? I love it, but..."
They firmly believe that I should start making small games that I could finish in 1-2 days. Probably Unity btw.
My question to you is if this is good advice. Isn't the time-frame they're suggesting way too short? And how are you gonna come up with something so compact and still with a decent gameloop in that time? I mean the design...
I guess that's about it. Thanks in advance.

I wouldn't focus that much on a 1-2 day deadline. Just keep the game small. Start with pong maybe? And move up from there?

Yes, I think it can be a valuable way to practice, especially if your difficulties are with project scope as described.

Maybe take part in some game jams.  The social aspect can be encouraging and is valuable networking, and you might pick up some tips from others. Also a great opportunity to prototype ideas that you might expand on further later if they work out well.

For a card game you dont need a game loop, just events.

For a gambling machine you can add timer to randomize graphics.

Then its time for tetris, can be still done without game loop, here you can add gameloop for smooth block falling and accurate keyinput.

Now make bomberman, try online gameplay.

1/2 days you know how it goes : that is what everybody says until they find out its always like : i would make that in 2 days, yeah right.

They not even trying, nor have they finished such crappy games, learn from it for 2 days and proceed if you see the light.

Dont ever finish a game, proceed learning.

As @Kuurde suggested, don't focus so much on deadlines other developers give you. Some of the best developers I've met are absolutely terrible at projecting ETA, and time on task. 

However, if you keep the project small, i.e Pong, you can certainly achieve that in 1-2 days if you allocate yourself enough development time, and have the proper tool set from the start.  The 1-2 days things is more analogous to, "Keep the iteration time short" to churn out a project. This has the benefit for beginners since churning out smaller, more compact projects gives the rush of completing the project, gives a needed ego boost for taking on more difficult projects and, of course, having something to show for it within a reasonable time-frame. 

Depending on your language of choice, and if you know how to code, jumping right into an engine is certainly an option, and it wouldn't be an inherently bad idea, but can easily prove overwhelming for someone new, imho. Whereas sticking to more simple "game-geared" framework for your language of choice will likely have a slightly less learning curve, and can give you a slight idea of the effort involved in interactive software.  

Somebody has downvoted me without saying why, that is rude.