Hi Nicolas
Nicolas Goaziou writes:
> Even though Org syntax partly is, I don't think parameterizable syntax
> is a way to go. I'd rather have less variables controlling it. (I'm
> looking at you `org-list-allow-alphabetical', and
> `org-plain-list-ordered-item-terminator'.)
Thanks for your explanations. I understand that the use of variables
here is perhaps not the happiest solution. Naturally, I speak of
'solution' assuming there is a 'problem', the choice of the comma as an
argument separator; but I also assume that what is a problem for me, it
doesn't have to be for other users, as here everything is subjective.
Anyway, I couldn't think of another way to respect the original comma
(and backward compatibility), but leaving the user some freedom to
choose a different separator. I also understand that implementing that
can lead to confusion.
>> Rationale for this patch: There are many contexts where the comma character
>> can be
>> inappropriate as an argument separator, since it has to be escaped
>> many times.
>
> That's true. But I wonder if you're hitting a limit of replacement
> macros use case. IMO, macros are good for short text. For bigger ones,
> you may want to use some Babel code, in the language of your choice.
> WDYT?
I think replacement macros have huge potential, that I often explore in
my daily work and (I admit) wearing the things to the limit, which means
struggling with resistance to the material ;-) Perhaps that emerging
potential was not originally foreseen ... Of course, the scenario is
always in short texts, I agree. But even in those scenarios, the comma
as an argument separator is somewhat uncomfortable and confusing. That
is, macros have great potential (IMHO) in contrast to an "ugly" syntax.
But at this point it has a difficult solution...
A typical use case for macros (for me) can be to export to LaTeX with
the command \foreignlanguage{lang}{short text} in multilingual documents
(see this screenshot:
https://gnutas.juanmanuelmacias.com/images/macros-sample.png). I can't
think of another way to do (easily) something like that (with the plus
of being able to evaluate some Elisp in there and add a conditional for
the backend, etc.).
Best regards,
Juan Manuel
