I'm not sure if it is a question for Stack Overflow or there is another, more appropriate site.

I need to ask this, but I don't even know how to ask this question.

Is there a name for code that is explicitly coded line by line? (compared, for example to a neural network (NN), that is trained and is not clear what it does) And there is a name for trainable code (NN) whose final data processing depends on the data and evaluation criteria it is feed?

Obviously, the code that runs the neural network is coded line by line, but that code is a meta-code which runs the weights of the neural network, which define a different data processing program.

I expect an answer like "classic code is named procedural programs, and NN are empiric programs"

  • 3
    @πάνταῥεῖ This is a legitimate site recommendation, not a blatantly off-topic question. Oct 6, 2018 at 17:17
  • 2
    Sorry for the above comment. We routinely get users who try to ask their (blatantly off-topic) programming questions here and try to get their answer here, so the community tends to act quickly on what appears to be those at first glance. I've edited your question's title to make it more clear that this is an on-topic site-recommendation. Sorry for our initial confusion. Oct 6, 2018 at 17:19

2 Answers 2


There are a few sites that come to mind in terms of scope. I'm not sure though the question fits the SE model well.

At least on the site I moderate, and the ones I'm active in, terminology questions are typically not a good fit. They're rarely interesting, practical problems of broad usefulness. If an answer is one line its either answering a question of monumental simplicity or is a work of mindblowing genius.

I'd take a careful look at the AI site, and see if they have similar questions before asking, but based off what I see, I don't feel that its a good fit in its current form.

  • Terminology is essential for people which needs to learn to fish instead of asking for fishes. I had deal many times with problems which were difficult because I didn't knew that they belonged to an already known, classified, studied and solved TYPE of problems. You waste a lot of time and effort until somebody tells you something like "hey, you need a SEARCH ALGORITHM". That one line is way more helpful than explaining how to find something step by step. There is something very wrong with stackexchange if it doesn't help with that.
    – jonuko
    Oct 7, 2018 at 17:26
  • Cool. You might want to suggest a site for op where he can ask Oct 7, 2018 at 23:14

It's easier for me to answer your question than to find another place to tell you to ask it. In problem domains where machine learning is used, the distinction is between:

  • Rule-based systems
  • Model-based, or machine-learned systems

As a further distinction, some rule-based systems are actually implemented with a system that stores a database of rules, while others are just procedural code.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .