Proto for Boost mailing list page: Re: [proto] talking proto

Subject: Re: [proto] talking proto
From: OvermindDL1 (overminddl1_at_[hidden])
Date: 2010-07-18 00:15:30

On Fri, Jul 16, 2010 at 10:03 PM, Eric Niebler <eric_at_[hidden]> wrote:
> On 7/16/2010 11:15 PM, Marcin Zalewski wrote:
>> On Fri, Jul 16, 2010 at 09:19, Eric Niebler <eric-xT6NqnoQrPdWk0Htik3J/w_at_[hidden]> wrote:
>>>
>>> I am not familiar with other technologies used to implement DSEL in
>>> other languages. Are you? Can you point me to some information or
>>> describe the other efforts in this area? I used little more than my own
>>> needs and my knowledge of compiler construction toolkits to guide the
>>> design of proto. Maybe there are other good ideas out there we can steal.
>>
>> Haskell people have Template Haskell [1]. It allows one to freely
>> generate syntax trees. One can have DSLs that "hijack" Haskell's
>> syntax, assigning it non-standard semantics. Oleg Kiselyov has a very
>> simplified example of tuning an interpreter into a compiler [2].
>>
>> It would be great if we could write DSLs in C++ at compile time, as
>> one can use Template Haskell to write DSLs in Haskell (or Lisp DSLs in
>> Lisp, and so on). But, until then, we have proto. :)
>>
>> Cheers,
>> Marcin
>>
>> [1] http://www.haskell.org/haskellwiki/Template_Haskell
>> [2] http://okmij.org/ftp/tagless-final/#tc-GADT
>
> Thanks for the links. Some good bedtime reading for me. But obviously if
> C++ is the host language, then a DSEL's syntax is necessarily
> constrained to that of C++. That rules out nifty tricks like
> meta-programming C++ in C++ a-la Lisp or Haskell. You can't make that
> pig fly. Maybe the answer is to just use Haskell, but that's not
> something I can tell proto's users. ;-)
>
> I guess my question is: what ideas from these languages/DSEL toolkits do
> we have a reasonable chance of cherry-picking in proto? Proto has
> grammars and semantic actions (transforms) because I've found them
> useful in Spirit and Antlr. Someone on the dev list brought up Van
> Wijngaarden grammars. I haven't tried to implement them yet (mostly
> because I don't understand them). There are probably lots of other good
> ideas out there.

Proto already feels a lot like LISP, no doubt there is a lot to pull
from it.  I have not made any Proto-based DSEL's yet, but no doubt I
could have a lot of input based on my LISP experience once I do...