Re: [Help-smalltalk] Unicode character special syntax

[Paolo Bonzini]

Mike Anderson wrote:
> I know I should have spoken up earlier, but I am a bit disappointed
> about the special syntax for Unicode character constants.
>   
Nothing's set in stone.
> My feeling is that much of the beauty of Smalltalk comes from the fact
> that it has very little in the way of syntax. Also, is it compatible
> with other Smalltalk dialects?
No, but it took me about twenty minutes to implement it in the RBParser 
which is what many other Smalltalk dialects also use.  I will post a 
message to comp.lang.smalltalk asking for advice.

I don't know the status of Unicode implementation in VW, Squeak, etc.  
At least for Squeak, they have a non-Unicode implementation of 
multilingualization, because it was written by a Japanese guy and they 
are usually unhappy about Han unification (sharing the same code points 
between Chinese, Japanese and Korean characters).
> "258 asCharacter" would be, at least potentially.
>   
Yes, you are right.  And indeed I was tempted to have "258 asCharacter" 
resolved at compile-time instead of having $<258>.  On the other hand I 
think that the problem in Smalltalk comes from the fact that it has *a 
lot* in the way of terseness (paraphrasing your sentence above): I liked 
the similarity between $<16r102> and <U+0102> (which is how the Unicode 
guys refer to the same character), and I think it is innatural to refer 
to a character as a conversion.  In almost every other language you can 
write something like '\u0102': it happens rarely, but it can be very useful.

I must say that I've often found "$ " to be completely unreadable, to 
the point that I'd write "Character space" instead.  So one could repeat 
the same discussion and say that the whole syntax for character literals 
is useless, because you could write "'X' first" or possibly "'X' 
asCharacter" instead.  In any way, character literal syntax leaves 
something to be desired.

Paolo