Re: [Chicken-users] srfi-19 'date?'

[Matt Gushee]

Hmm ...

On Sun, Oct 13, 2013 at 11:11 AM, John J Foerch <address@hidden> wrote:

> I noticed that srfi-19-date (version 3.3.4) does not export 'date?'.  I
> think that perhaps it should.

On my system:

csi> (use srfi-19)
; loading /usr/lib/chicken/6/srfi-19.import.so ...
; loading /usr/lib/chicken/6/srfi-19-timezone.import.so ...
; loading /usr/lib/chicken/6/miscmacros.import.so ...
... and so on

csi> date?
#<procedure>

Note that I loaded 'srfi-19', not 'srfi-19-date'. So there is not an
SRFI-conformance issue. But if your point is about having a logically
consistent API,  ... yes, I see that srfi-19-time exports time?, and
srfi-19-period exports time-period?, so you seem to have a valid
point.

> Also, 'date?' is not documented in the
> egg docs (http://wiki.call-cc.org/eggref/4/srfi-19).

If you read the intro to the egg doc, it states that "This document
only describes the extensions [to the spec]."

I'm not claiming this is not a problem. The way Scheme is modularized
leads to a situation where documentation for closely-related functions
can be scattered all over the four corners of the internet. I have
long believed that this is a serious obstacle for anyone trying to
master the language, and I haven't seen anyone deny that ... but I
don't think anyone knows how to fix the problem. I have envisioned a
sort of 'smart documentation aggregator' that would enable you to view
in one place all the documents that logically go together. But that
would require either (a) an insanely good search engine; (b) a
Scheme-community-wide agreement to add certain kinds of metadata to
all the docs everywhere (which would not be technically difficult, but
would require an awful lot of work, and ... sounds like a herding-cats
problem to me). Or you could screen-scrape everything, but then the
system would break whenever anyone decided to change their
documentation format.

--
Matt Gushee