Re: Why context-free?

Matthias Blume <find@my.address.elsewhere> wrote:
>nmm1@cus.cam.ac.uk (Nick Maclaren) writes:
>
>> However, my experience is that most people think in a context-full
>> fashion, and so there is a strong argument for designing a language
>> where controlled aspects of the context are first-class parts of the
>> syntax.
>
>I agree with the premise, but I don't understand how you derive the
>conclusion from it. Context-sensitive aspects of programming
>languages are routinely handled by subsequent passes (type checking,
>elaboration, etc.) There is no particularly good reason to move this
>stuff into the parser, and there are several good reasons (which
>people here have already enumerated) for not doing so.


The fact that there is a strong argument in favour of A doesn't mean
that there aren't even stronger arguments against it! All is means is
that A shouldn't be rejected without good reason.


My logic is that it is often necessary to play revolting games in
conventional languages to get around restrictions imposed by the
syntax, often in ways that could quite easily be relaxed. Most such
relaxations are best done in a context-free fashion, I agree (why
complicate things?), but some aren't.


Something that has been forgotten by some of the language community is
that the conventional approach is just that. It has become
conventional because it has a good theory behind it that maps fairly
well into practice, but that doesn't mean that it is the only
worthwhile approach. Designing languages and writing compilers is
like writing tribal lays - see Kipling :-)


Regards,
Nick Maclaren.

Return to the comp.compilers page.
Search the comp.compilers archives again.