>
> http://ashishagarwal.org/2011/10/04/pdf-type-theory/ 
>
> Basically, the author describes a DSL for describing statistical 
> distributions, and programming techniques for turning those 
> descriptions into random number generators.
>
 
 Hmm, this is *interesting*. I look forward to a prototype of the language.

Back to the statistics thing. The grammar is tiny, but involves 
> Unicode. I'm tempted to provide "US Typeable" equivalents for the 
> Unicode bits, which are mostly "Γ", "ϒ" and "⋅" as far as I can see at 
> first glance. I trust nobody would object too loudly if I used 
> "GAMMA", "UPSILON" and maybe "." for them? Maybe just "G", "Y", and 
> "."? They also casually kick "π" around, and I'd assume "pi" is quite 
> sufficient for "US Typeable" there. 
>
 
Providing ASCII aliases is a sane solution, but I'd keep the Unicode 
versions around as well. Prior art:

   - ALGOL with different representations for printing and actual 
   programming<https://en.wikipedia.org/wiki/ALGOL_68#Program_representation>. 
   This tradition has been carried on to a certain degree with academic Lisp 
   and Haskell.
   - C with digraphs and trigraphs for non-ASCII systems.

 Note also that "<" (less-than) and "〈" (left angled bracket) are different 
operators. For the latter, the digraph "<(" might be more sensible. Using 
"<" in both meanings would be rather confusing.

Aside from me, does anyone else feel like teaming up on the coding of 
> the semantics of the DSL as actual Perl code? I think I can handle the 
> writing up of the DSL as a SLIF grammar, but the mathematics used in 
> the rest of the document is right at the borders of my understanding. 
> I know there are some _seriously_ smart cookies on this list, so maybe 
> if I had one or two people prepared to act as "math coach" if / when 
> the going gets tough, that might be enough. 
>

 I'm going to follow this project closely, but I don't have time to 
actively contribute code. I have a working knowledge of statistics 
(probably on a sufficient level for this), but I haven't (yet) implemented 
a type system. I encourage you to go forward with this project, and just 
drop somebody a line when something isn't clear to you. Together, we should 
be able to figure this stuff out.

-- 
You received this message because you are subscribed to the Google Groups 
"marpa parser" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to