Is it worth developing a new programming language?

Question

Programming languages

Is it worth developing a new programming language?
Actually, I have repeatedly asked myself about the redundancy
and excessive universalization of existing solutions ... The fact is that there are NO
optimal solutions in the field of compilers
and programming languages ​​today . And that's why:
- optimization for a specific processor architecture (CPU or GPU) is a rather abstract thing and is solved not at all by experienced systems and not even by heuristics, but by a simple DFA (if ... then) due to possible heavy loads. It would seem that this solution is quite logical and optimal, but with the advent of clouds, clusters and other parallel goodies, it would be possible to find some kind of “faster” integrated solution, and it still exists ... but who needs to invent turbocharged bicycles?
- a person solves the problems of application abstraction in most cases with an imperative
"black box" and a pack of templates (ideally 2-3), this is how zombie templates and other underprojects are born. “O_o I discovered the world of erlang'a,” Yura said. "And we don't care." team said.
And this has already happened to me more than once - in one case, OOP in another functional ... in the third brainfuck and a site on fasm'e. who cares? .. no one cares - "as long as it works" no one thinks about speed and ephemeral "optimality". So I will say that people have too much freedom! Abstraction for that and abstraction to say this is a ball. but let the language itself decide what is inside, if methane is needed, there will be methane, if silver is needed, there will be silver.
- 0_o we all like STL it's so soft and fluffy... there are so many goodies in it =)
why do we need to learn those fucking algorithms? why bother with the data structure?
It can be redundant? - but what are you talking about ... it works well, even if slowly ...
but it does work. And so every day ... people often try to solve the problem of calculating the optimality of algorithms for the final implementation, but almost everyone is always waiting for a complete epic fail.
As Uncle Knuth said, "Low-level optimization is the root of all problems."
And here, too, people have too much freedom ...
Actually, what am I talking about? .. and the moral is quite simple:
the existing solutions are somewhat complex, but due to hardware loads, the key points are generalized and “universalized”, which leads to the loss of “optimality” of the output.
That is, Universal solutions cannot be optimal, only complex ones ...
And a huge problem in our time is precisely optimization, no matter what multi-core, multi-threading, load distribution, low-level ...
Let us have such an ephemeral torrent compiler a la SkyNet for this.
We don't care about hardware time or memory size.
We need a "no-compromise" compiler and i386-time paradigms.
Actually, the software has never shined with performance - and the hardware platform is already far ahead.
That's all, I think the reasons for the "decline of compilers" are clear to everyone.
Interested in the standard question: "is the game worth the candle?"
I took up the development of such a "miracle", while (unfortunately) for 1.5 years the mat has been perfectly honed. part.
I just don't want to repeat google wave.
I don't want to be bought by someone.
I want to find like-minded people.
PSI apologize in advance for possible mistakes - Russian is not my native language.
I also apologize for the excessive confusion of the presentation.
I hope for your understanding ... It's just that "that's all" is already an order of magnitude.

Answer 1

If there is a burning desire, then write your own programming language. And do not listen to those who tell you "it's not worth it, no one needs it."

Answer 2

I'm sorry...
When reading the comments, the bloody star from the BolgenOS logo flares up brighter in front of my eyes.

Answer 3

no. )

Answer 4

If you want to develop for your own pleasure - definitely worth it.
If you think that your development will solve a whole layer of abstractly described problems and become necessary everywhere - hammer it. There will be no such thing.

Answer 5

Is it possible to see something specific by language? Read your articles, blog posts, etc.?

Answer 6

Fuck dude, what are you up to? I tried a lot of things, but what would the wing be like ... Envy-envy-envy.

Answer 7

Formulate the goals of the project, the tasks that the new language will solve, more clearly. Up to the formal detailed formulation.
As soon as you succeed, an understanding should come whether it is needed, and how much the game costs, and how much the dressing.

Answer 8

You can write programs in a language tailored to a single piece of hardware, and it will run faster than programs on "universal" computers. But in order to use the capabilities of iron by 100%, you need to write your own program for each piece of iron in your own programming language.
Another approach is what you wrote about, "universality". Double-edged sword. Market requirements dictate the need for rapid software development - this is how all sorts of high-level programming languages ​​appear. But once pure C was considered impermissibly "bloated" in comparison with the same asm. But the "market" said - "we need it to work here and now, and so that specialists can easily learn new languages ​​and write software for our hardware." Performance had to be sacrificed - this is called a "compromise".
Summarizing, and answering your question: if you want to practice developing original technologies, learn how to step on a rake, and notice them from afar in the future, create your own PL.

Answer 9

Of course it’s worth it, you have your whole life ahead of you, then there will be something to laugh at , something to remember.
In general, you would be on rsdn in philosophy, to a more targeted audience.

Answer 10

If you ask yourself this question, the answer is no :)
+ I liked the answer from Christmas

Answer 11

How will the new language differ from the known ones? What tasks will it solve, what are the benefits of using it? Here is what the developer should answer before creating the brainchild.

Answer 12

IMHO, you need to create a language only if you are related to hardware ...
Write an assembler if you are the creator of your own processor ...
And write high-level languages ​​​​like LISP, ADA, PASCAL. Only for the purpose of teaching students or as a hobby (Python).
PS Try to think of the world as a supply-demand world... And the question of the need for programming languages ​​will disappear by itself...

Answer 13

I think this is a utopia. If you stick to the saying - Every programmer must write his own programming language, then maybe it's worth it to satisfy his impulses. But as a serious programming language, I think it’s not worth it, this is a colossal work, and it will be very difficult for one, even a small team, to complete, and even then to support. If you do an open-source project, then if only it didn’t work out like with all kinds of PHP, when release 5.2.0 was released, in which the passing of parameters to functions by reference was killed.

Answer 14

Programmers' time is worth more than hardware. Therefore, they optimize writing time (easier to write, existing libraries) than execution time (any optimizations).
According to this parameter, existing languages ​​(already known to programmers) are better than new ones (which need to be learned)

Answer 15

Wow, how did you get stuck! You're more careful with the dosage of any golucinogens!

Answer 16

prologue strikes back \(*o*)/

Answer 17

Along the way, the world is full of languages ​​today no worse than the Tower of Babel. Write a better new processor.