On 17 March 2016 at 22:46, Juergen Sauermann <juergen.sauerm...@t-online.de> wrote:
> Hi Elias, > > this is quite feasible technically. What I don't like, though is the > incompatibility > that it creates for the source code. > > Suppose I write a function Z←AؠB. Who else except myself can decipher > what it is supposed to mean? I strongly believe that restriction names > to almost only ASCII letters (and writing the names in English) is a very > good convention that we should not easily give up. > Well, you used an arabic letter in your example, which of course is quite indecipherable to me. If you had named the function KASHMIRI_YEH, it would be equally indecipherable. However, allowing arbitrary Unicode would allow you to name a function ≈, for example. Another useful example would be a physics application using the symbol ℏ. As you can see, only allowing english letters doesn't guarantee readability. And allowing Unicode doesn't imply hard-to-understand. Anyone who cares about incompatibility simply wouldn't use these characters. You could even add a flag --disable-extensions to ensure this. Regards, Elias
