meta Meta language released

[zbyti]

In #344 was incorrect Rebol forum link to FizzBuzz - fixed. It may have looked intentional but it was not.

[Kaj de Vos]

That forum has a permalink on top of the threads. The normal links change when new posts appear on the forum.

[zbyti]

Not everything works as expected in the Rebol world  

[zbyti]

Which example is more readable and "human-friendly"?

 

Meta [
	Title:   {FizzBuzz math "game"}
	Author:  "Kaj de Vos"
	Rights:  "Copyright (c) 2021 Kaj de Vos"
	License: {
		PD/CC0
		http://creativecommons.org/publicdomain/zero/1.0/
	}
	Notes: {
		https://wiki.c2.com/?FizzBuzzTest
		https://www.rosettacode.org/wiki/FizzBuzz
	}
]

For counter 100 [
	Third?: unless modulo counter 3 [write "Fizz"]

	Any [
		unless modulo counter 5 [write "Buzz"]
		third?
		write counter
	]
	Write " "
]
Write new-line

or

For number 100 [
  p: 0
	
  unless modulo number 3 [p: 1 write "Fizz"]
  unless modulo number 5 [p: 1 write "Buzz"]

  if p = 0 [write number]

  Write " "
]
Write new-line

1, 2 or none?  

 

maybe this one?

For number 100 [
  p: 0
	
  unless modulo number 3 [p: 1 write "Fizz"]
  unless modulo number 5 [p: 1 write "Buzz"]

  unless p [write number]

  Write " "
]
Write new-line

 

Edited December 14, 2021 by zbyti

[ilmenit]

It's probably just a personal preference but I'd say "none"  Lets compare with the Python solution:

for number in range(1, 100):
    if number % 3 == 0 and number % 5 == 0:
        print('FizzBuzz')
    elif number % 3 == 0:
        print('Fizz')
    elif number % 5 == 0:
        print('Buzz')
    else:
        print(number)

 

[zbyti]

[ilmenit]

?

[Kaj de Vos]

I think Zbyti's version follows the Principle of Least Surprise. I optimise my programs heavily with REBOL constructs that are elegant and that I know to generate efficient code, but not all of them are available in other languages. I want the examples to show off Meta's strengths, but in beginners' guides I will start with simpler constructs.

 

Many REBOL examples are different because they disregard efficiency. I think they make too much use of the dynamic features of the language. I inherited that attitude from Atari.

[zbyti]

1 minute ago, ilmenit said:

?

[Kaj de Vos]

11 hours ago, zbyti said:

Not everything works as expected in the Rebol world  

Yep, you get free thinkers with original solutions.

[ilmenit]

38 minutes ago, Kaj de Vos said:

Many REBOL examples are different because they disregard efficiency. I think they make too much use of the dynamic features of the language. I inherited that attitude from Atari.

Many programmers like to feel smart by writing very dense code and like in their eyes "elegant solutions". In many software houses such approach is considered a "bad habit" leading to "write-only code". Such code is hard to read for the ones who didn't write it, require compilation of the code in the head, is hard to modify and makes life harder for new programmers. While languages nowadays are very powerful and flexible it also takes experience not to use all the fancy features the language provides. 

[zbyti]

I think my last example shows a problem in trying to talk to the computer with too many words:

 

[unless modulo number "works" do something] in other words: do something if something happens.

 

unless p I'm reading - do something if something didn't happen.

Edited December 14, 2021 by zbyti

[TGB1718]

I think what we are missing is comments !!!

yes, that age old thing that many programmers just don't do.

 

As this is a new language, adding a comment alongside each line of code will help explain not

just what the program is doing, but also give insights on how the language works

[Kaj de Vos]

In the end, these problems are caused because our brains are limited. A language that can express more concepts in the same space is said to be more expressive. Thus, bigger challenges fit in your head when the programs are smaller.

 

It seems to be a problem of communication between human and machine, but the machine actually concurs. It, too, is limited, and smaller programs perform better.

[Kaj de Vos]

2 minutes ago, TGB1718 said:

I think what we are missing is comments !!!

yes, that age old thing that many programmers just don't do.

 

As this is a new language, adding a comment alongside each line of code will help explain not

just what the program is doing, but also give insights on how the language works

This example is ten lines and is compared with equivalent versions in hundreds of other languages. That should be enough for such an example.

 

As I said, in beginners' guides I will build it up as simple as possible, keeping the feedback here in mind.

[zbyti]

I think that symbolic languages are more exact and does not get the brain involved in understanding words but only problems.

[Kaj de Vos]

22 minutes ago, ilmenit said:

Many programmers like to feel smart by writing very dense code and like in their eyes "elegant solutions". In many software houses such approach is considered a "bad habit" leading to "write-only code". Such code is hard to read for the ones who didn't write it, require compilation of the code in the head, is hard to modify and makes life harder for new programmers. While languages nowadays are very powerful and flexible it also takes experience not to use all the fancy features the language provides. 

Meta is not APL:

https://www.rosettacode.org/wiki/FizzBuzz#APL

or K:

https://www.rosettacode.org/wiki/FizzBuzz#K

 

Meta is deliberately designed to be more verbose, literally. That's why I claim it to be closer to natural language. It is designed to work at a similar verbosity as human language.

[Kaj de Vos]

4 minutes ago, zbyti said:

I think that symbolic languages are more exact and does not get the brain involved in understanding words but only problems.

Meta is just as exact. It's Lisp with types added. But it tries to use real English words instead of weird abbreviations, and only the math symbols that many people are familiar with.

[zbyti]

@Kaj de Vos you do your job developing META I do mine using it. I have my preferences like K65 over 6502 mnemonics but I'm always open to something new. You demand feedback - I provide one  

Edited December 14, 2021 by zbyti

[Kaj de Vos]

Because Meta is word based, you will be able to equate existing methods to any weird Unicode symbol you want. Pi would be useful, for example.

[ilmenit]

1 minute ago, zbyti said:

I think that symbolic languages are more exact and does not get the brain involved in understanding words but only problems.

If I'm correct, adults usually do not read single letters in words but read familiar words as a symbols, leading to these tasks like "count the Fs" or https://www.lookhuman.com/design/58406-if-you-can-read-this-you-are-smart/tshirt (while in case of children just learning it's an issue). 

 

[Kaj de Vos]

https://www.lookhuman.com/design/58406-if-you-can-read-this-you-are-smart/tshirt

I think John Lennon and Einstein would be able to read that.

[zbyti]

28 minutes ago, Kaj de Vos said:

In the end, these problems are caused because our brains are limited. A language that can express more concepts in the same space is said to be more expressive. Thus, bigger challenges fit in your head when the programs are smaller.

 

It seems to be a problem of communication between human and machine, but the machine actually concurs. It, too, is limited, and smaller programs perform better.

I read the biography of Stanisław Ulam and there was a discussion about what language is better and mor suitable for certain things. The gentlemen knew 3 languages then, maybe four (Latin?).

 

If I'm not mistaken they found German more suitable for mathematics than Polish, I don't remember what they said about the French language

[zbyti]

9 minutes ago, ilmenit said:

If I'm correct, adults usually do not read single letters in words but read familiar words as a symbols, leading to these tasks like "count the Fs" or https://www.lookhuman.com/design/58406-if-you-can-read-this-you-are-smart/tshirt (while in case of children just learning it's an issue). 

 

yes, I think this only happens if the words are relatively small in number and are used frequently - such as key words in programming languages.

[zbyti]

7 minutes ago, Kaj de Vos said:

https://www.lookhuman.com/design/58406-if-you-can-read-this-you-are-smart/tshirt

I think John Lennon and Einstein would be able to read that.

I can read that, I often have a bigger problem with what I write myself