Re: some questions about pragmas, and the arguments as I understand them

Thanks for Bethan and Tom and Dave for helping make me see something
more clearly.

I think BTW is correct that we can usefully distinguish grammatical
realizations in which nonterminals 'comment' and 'pragma' are each
specializations of some more general notion, from realizations in which
they are distinct and independent.  

I think DP is also correct that the choice between 5.b/c and 5.d is
intermingled with decisions on whether comments and pragmas can appear
in the same places or not.  Answer d to question 5 entails allowing them
in the same places.

Michael


Bethan Tovey-Walsh writes:

> Ah; okay. I think what I’m trying to convey with this suggestion is
> the sense that comments and pragmata are related in a sibling-like
> fashion, not hierarchically. I understand 5c to be saying that there’s
> a thing called "comment" and a thing called “pragma”, and they’re
> independent. I’m suggesting instead that we say that there’s a thing
> called “out-of-band information”, and that it has two subtypes:
> “comment” and “pragma”.

> But I may be making an unreasonable distinction here; I’m sorry if so.

>> On 3 Feb 2022, at 09:26, Tom Hillman <tom@expertml.com> wrote:

>> I think that's what I understood 5.c. to mean (i.e. that comments
>> aren't simply any inline out of band information, but inline out of
>> band information that is NOT for the processor).

>> On 3 Feb 2022, 09:24 +0000, Bethan Tovey-Walsh <accounts@bethan.wales>, wrote:

>>> Is it reasonable to posit a 5d. - that comments and pragmata are
>>> both subtypes of inline out-of-band information? This could lead to
>>> a representation in the grammar along these lines:

>>> oob : “{“, S*, (comment; pragma), S*, “}”.
>>> comment : whatever.
>>> pragma : “[“, S*, pragma-name, pragma-data?, S*, “]”.

>>> (Apologies for any syntactic errors - I’m writing on my phone, which
>>> is not ideal!)

>>> I think I’d prefer this, conceptually, because it leaves aside the
>>> question of the relationship between pragmata and comments. But I
>>> could absolutely live with 5b.

>>>> On 3 Feb 2022, at 09:09, Tom Hillman <tom@expertml.com> wrote:

>>>> Thanks ..

>>>> In particular, I can now see how, if a comment is “inline out of
>>>> band information”, and a pragma is “inline out of band information
>>>> for processors”, then pragma are, indeed, a subtype of comment, at
>>>> least conceptually, and I can see how that is an argument for
>>>> 5.a. or 5.b.

>>>> ...

-- 
C. M. Sperberg-McQueen
Black Mesa Technologies LLC
http://blackmesatech.com

Received on Thursday, 3 February 2022 15:02:18 UTC