Re: Syntax

Stephane Bortzmeyer schrieb:
> On Sun, Jan 07, 2007 at 06:12:09PM +0100,
>  Julian Reschke <julian.reschke@gmx.de> wrote 
>  a message of 31 lines which said:
> 
>> 1) Although the language is designed to be used in IDs and RFCs,
>> restricting it to ASCII here is IMHO a very bad idea. After all, you
>> may want to use it for other specifications, and the IETF may lift
>> the current restrictions at some point of time. I would suggest to
>> require a specific text encoding such as UTF-8,
> 
> Any idea about the support of UTF-8 in typical languages *and* parsing
> tools? For instance, with C and Yacc, I assume it is quite
> difficult. With Haskell, I'm not sure :-)

Well, I've been living in Java world for a long time, so I really can't 
say anything useful about other languages anymore.

> The problem of UTF-8 is that many assumptions no longer hold:
> 
> * case insensitivity becomes a problem,
> * enumeration of "reasonable" characters become a complex task.

That's a problem with Unicode (the character set), not UTF-8 (the encoding).

Yes, case insensitivity is harder, but is this relevant for cosmogol, if 
  everything stays case-sensitive (which I think is the right thing to do)?

Choosing characters for identifiers: again, just borrow from somewhere 
else, such as <http://www.w3.org/TR/REC-xml/#NT-Name>).

> I'm a big fan of internationalization and Unicode but, since Cosmogol
> is intended for a technical and limited use, is it reasonable? 

I think inventing a new format, but not taking I18N is very hard to 
defend. As far as I can tell, there's no real chance to get it published.

> Anyway, I recorded the point as a TODO (we do not have a formal issue
> tracker) in the draft source. Other advices?

Let's leave it at this for now.

Best regards, Julian

_______________________________________________
Cosmogol mailing list
Cosmogol@ietf.org
https://www1.ietf.org/mailman/listinfo/cosmogol