lua-users home
lua-l archive

[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index] [Thread Index] 



Le 23/11/2016 à 23:38, Mikoláš Štrajt a écrit :

Hi Mathieu,

as man interested both in lua and esperanto I am impressed by your project(s).

In fact I was thinking about easily pronounceable programming language too. My version was named Parola and it was based on verses for structure and Esperanto for vocabulary. I never described it yet but a I can write some description if you are interested.
Sure I am interested, I would be delighted to receive more information about that. 
	
I have few points to your projects:

  - it definitely needs some form of english documentation in repositories. Link to your project introduction in lua-l archive is good start.
Well, for Lua-i18n, so far it's all in English, but adding the link might indeed be an interesting idea.
For Mallupa, possibly I'll do welcome documentation in misc. languages. The project is already taking the shape of hosting misc. dialects to standard Lua.
For Lupa, currently there is currently almost no documentation. So far the "dokumentaro" directory is just the English documentation of Lua. There's the README and the wiki page, which contains some instructions. I don't plan to translate that in English or any other language for now, but contributions are welcome. 


  - mathematical operators symbols are more readable than keywords IMHO. Anyway when program is read aloud then "z = x + y" becomes "z igi x plus y".
Well, give yourself criteria for measuring readability, and then you might evaluate it. The problem with introducing extra-alphabetical symbols, is that you implicitly encourage people to go this way, and you end up with a bunch of esoteric formula that are convenient to type once you mastered it, often hard to decode even for great guru of the obfuscation, and impossible to grab in any way for those who didn't passed the mystic initiation ritual.
It's a well known psycho-linguistic theory that people don't read characters but read entire words anyway (but some articles moderate this theory[1]). So whether you use a single glyph or a set of them that readers can grab straight forwardly as a reading unit doesn't make much difference per se on the lexical recognition.
However it can make a huge difference for the reader. If "z" is depth, then just get rid of the cryptic tradition and name it depth. It's not only more meaningful: it also far easier to search in the code. Juxtaposition of single glyphs may ease your work on a paper pencil scribble that no one need to browse and edit, that won't help much in a source code where people have to team up.
On the other hand it looks like occult symbols have a huge social attracting force. 


[1] https://www.microsoft.com/typography/ctfonts/WordRecognition.aspx

  -  when I was designing my language I found there is no easy way how to express nested structures in human languages. Did you came across this or similar issue?
That's because most human languages grow for human communication needs, and that most of the time it goes with implicit assumption that the interlocutor is smart enough to infer the biggest part from context.
Nested structure like the `for` loops you give as an example aren't often necessary when you talk to human. In a you couple in love one may say "Please my dear sweaty honey, will you help me with making the dinner and prepare the chips?", or in the same couple a few years later "Chips!". You don't tell the other how to prepare chips (well, maybe the first time), and certainly you don't explain that in a way where you explicitly enumerate each potato.
And what you rarely explicitly saying in a given language, will often be longer to express. 



E.g. There is no easy way how to express this structure using sentences from human language:
Well, there is often no short way to express it, but there definitely exist ways to express it. 


```
for i=1,5 do 
   for j=1,i do
     io.write("*")
   end
   io.write("\n")
end
``` 


(example taken from Rosetta code)

Here is a quick more or less Esperanto sentence congruent version:

por ero valorata ekde unu ĝis kvin faru ke
    por aĵo valorata ekde unu ĝis unu faru ke
      skribu "*"
    sekve
    skribu "\n"
fine
Not that here I just made relexicalisation, if you authorize some syntax change, then you can make things look more regular, and probably more compact. For example you may write `fine skribu "\n"`. From a syntactic analyze point of view, you just tell that the next expression is the last one of the current block, rather than saying that you are closing this block right now (or that the previous expression was the last of this block if you prefer). I'm not sure that placing "fine" in the end of the proposition is even possibly an error in Esperanto grammar. In Lupa I opted for "hop" which is according to PIV an interjection used to urge a jump. It's also a word related to jump in several languages. Also, for the "shorter is better" freaks, it's as short as "end". 
Finally:

In the later discussion, there were mentioned emoji-based language. I think you should look into HearthForth (https://github.com/neilk/HeartForth/).
Well, that's funny. I can't much sense of having a hearths. Well, the broken hearth do make some sense to me, but it's definitely pass through arbitrary cultural references. Even dropping to garbage is so (🚮[2]). And ⚖[3] or 𐄷 for equality maybe. At least it seems, to my mind, closer to what was suggested in the thread. 

[2] http://unicode-table.com/en/1F6AE/
[3] http://unicode-table.com/en/2696/

Regards