Question:
Recently (ok, it's been a good few months now) I've been learning about the F# language , which is Microsoft's answer to offering a working programming language on the .NET platform.
However, it was not clear to me the real cases and advantages of using this paradigm over "traditional" object-oriented languages that are already well consolidated (in this case, C# or VB.NET).
Therefore:
- In which cases does functional programming enable solutions that object-oriented programming is faulty / time-consuming / impossible?
- Are there any use cases of this language/paradigm in known (or open-source ) projects?
- From a professional point of view, for the demands of our (Brazilian) market, is it worth deepening the knowledge in this paradigm?
Answer:
Although the Imperative programming paradigm is the most popular among [professional] programmers, it is only one of several ways to "give orders" to a computer. In addition to it and Functional , we also have Logic Programming (ex.: Prolog), Dataflow (ex.: the internal engine of a spreadsheet), Function-Level (programs do not manipulate data, but other programs), etc. And, of course, the more conceptual/theoretical Turing Machine – to which all previous paradigms are computationally equivalent.
Each of these paradigms has its advantages and disadvantages, and it is common as programming languages "evolve" that they incorporate aspects of the other paradigms. The main reason why one paradigm or another is used is its expressiveness : in principle, we could all be programming in Turing Machines, but in addition to the difficulty of reading/writing/understanding such programs, they would also be too extensive in relation to what make of useful. Likewise, different paradigms offer a greater degree of expressiveness for specific domains (even if the language as a whole is Turing-complete).
As for F# in particular, I'll quote an answer to a similar question on Stack Overflow in English:
My expectation is that F# will be used for parts/parts of some specialized systems – the parts that involve complex threading / math / financial calculations / modeling / etc, where F# fits well. In most other areas (UI, Data Access Layer, etc), a general purpose language like C# seems (in my opinion) preferable.
One of the advantages of F# is that (in theory) you can prove that the code is working, rather than just testing it. Threading support (thanks to immutability and the asynchronous use of
!) is also good (although PLINQ can compete on threading ).
(Comment: I disagree with the author of this answer in the sense that it is possible to prove that an imperative program is correct – it just takes more work…)
I can't comment on our market, since I don't participate in the Microsoft "ecosystem" (I prefer to develop in free software), but in general, opening the minds of Brazilian programmers to functional programming I think would be very welcome!