Interesting take on the PLO - page 4

 
Maxim Dmitrievsky:

there's talk that he's gone to build a strong AI

with his enthusiasm, if he gets it right, I think he will be very good at it.

another 2 or 3 years before morale is boosted)

 
Maxim Dmitrievsky:

rumour has it that he has gone to create a strong AI

I wonder if he has read about the amount of data to be trained? It's a good idea to have a small cloud of your own just to start trying things out.
 
TheXpert:
i wonder if he has read about the amount of data to learn? it's a good idea to have a small cloud of your own just to start trying things out

well, yes, a small one, like a google or sber cluster)

 
Mikhail Mishanin:

With the advent of OOP, it really scared me off with encapsulations and inheritance and other morphisms, which are popular to start an introduction to OOP. They would write simply that data together with functions in one block - a class. But then I got used to this fashion - to invent new terminology each time. When I need OOP - no problem, but since it, in my humble opinion, is twice as voracious and slower than the procedural variant, I do with the procedural one at most.

And the question arose, because I wanted to hear opinions about FP and other variants, other than OOP and procedural, maybe someone has a super positive experience. And I suppose it's wise to use some kind of functioning mix and not to be a fanatic.

Try not to use all of those offered by OOP at once, start with the simplest and imho most "tasty" - with encapsulation, and then maybe other OOP features will come in handy.

encapsulation allows you to get a new user-defined data type, imho, it is convenient - you can use a new data type in procedural style


SZZ: the first example that comes to mind - matrices and operations with matrices, after creating a class "matrix", you can add operations with matrices - methods, resulting in a ready-made implementation, which will work as in mathematical programs Matlab, R etc.

Matrix A,B,C;
//.... заполнение данными матриц
C = A + B;
 

What's the point, exactly? )))

matrix addition functions and in general python and Er can be implemented in functions with references, and just name rules, directory or formation name logic ...))))

 
Valeriy Yastremskiy:

What's the point, exactly? )))

matrix addition functions and in general python and Er can be implemented in functions with references, and just name rules, directory or formation name logic ...))))

By the way an example of solid OOP is python. There, rather no one knows that anything other than OOP exists

 
Valeriy Yastremskiy:

What's the point, exactly? )))

matrix addition functions and in general python and Er can be implemented in functions with references, and just name rules, directory or formation name logic ...))))

if the question is to my example - at least you hide the implementation ( you hide even from yourself) - i.e. you write only the logic of calculations, it's convenient, it's readable, it avoids logical errors


SZZY: with respect to trading - I wrote and still do experiments with grids of orders, where I write the logic of placing orders as A + B - C , where A, B, and C - these are orders with predefined parameters, very convenient to use for genetic algorithms- interesting topic

 
Igor Makanu:

Try not to use all the OOP features at once, start with the simplest and, imho, the most "tasty" - with encapsulation, and then other OOP features may come in handy

encapsulation allows you to get a new user-defined data type, imho, it is convenient - you can use a new data type in procedural style


SZZ: the first example that comes to mind - matrices and operations with matrices, after creating a class "matrix", you can add operations with matrices - methods, resulting in a ready-made implementation, which will work as in mathematical programs Matlab, R etc.

The impression is that everyone reads between the lines and everyone has something different) Where it's more convenient to use OOP, I use OOP; where you need to save memory and time, and to code for yourself, I stay with the procedural. Just came across an article, I wanted to hear an opinion, where / what is better). As a result, I have heard a lot of different things in my address, and not about programming) Everything is as usual.

 
Mikhail Mishanin:

It seems that everyone reads between the lines and everyone has something different) Where OOP is more convenient, I use OOP, where I need to save memory and time, and code for myself - I stay in procedural. Just came across an article, I wanted to hear an opinion, where / what is better). Result - I've heard a lot of different things in my address, not about programming) Everything is as usual.

It's not Peter, he's much younger in mind.

 
Maxim Dmitrievsky:

By the way, an example of continuous OOP is Python. There, rather, no one knows that anything other than OOP exists.

Encapsulation gives freedom of names. And if this problem is solved by name logic. This is of course costly. then python can be written in functions. but it will not be a marketable solution. BUT it is POSSIBLE.

Reason: