On 11/27/2012 04:19 PM, John Cowan wrote:
> Per Bothner scripsit:
>
>> The difference isn't always clear, but it does seem a useful
>> distinction especially for "document"-like applications.
>
> On reflection, I'm inclined to agree that initial-expressions are useful,
> but I still think they belong in an "advanced quasi-literal" SRFI, not
> in this one.  Granted, without them the distinction in bracket types is
> harder to motivate, but I can live with that.

I think a useful common mapping for
   #&cname{pre-exp ...}[abc&{infix-exp1}def&{infix-exp2}...xyz]
to translate into the equivalent of:
   (cname pre-exp ... #&[abc&{infix-exp1}def&{infix-exp2}...xyz])

I.e. a function (or macro) application whose arguments are the
initial expressions (often keyword-argument pairs),
followed the "literal" arguments converted to a string.

My plan is to provide a library macro to make this easy.

I already have 3 SRFIs to juggle, and it may be reasonable to
split out support for internationalized strings from SRFI-109
into a separate SRFI.  So I'm reluctant to create a separate
"advanced quasi-literal" SRFI.  Maybe if we end up with more
"advanced" features.
--
	--Per Bothner
xxxxxx@bothner.com   http://per.bothner.com/