Nelson, Clark wrote:

> On the one hand, they don't want to have to include all the standard
> library headers to figure out how to configure for a library. On the
> surface of it, that seems reasonable.

When I raised this a year ago the only argument you seemed to have was that
the person adding the feature to the stdlib should not have to open another
file in their editor to add a define there.

It's honestly one of the most shocking arguments I've ever read on the
internet. It completely floored me. I fully did not expect to see an
argument like it raised in a forum like this.

However, as you're the one who decides, and no one else seemed to have a
position, I effectively left it at that. I stopped following this group
because the 'open as few files as possible!' argument makes absolute-zero
(as in zero Kelvin) sense to me but appears to be considered sensible as a
winning argument to SG 10.

If that's the winning argument (or an argument which is relevant in any way
at all), and particularly if it's the argument raised by the person
responsible for the content of SD-6, then clearly my logical understanding
of the world does not fit SG 10 logic and I have nothing further to bring
to the group :).

Anyway, now you have a second independent source telling you the exact same
problem exists with the stdlib features, so I thought I'd notify.

Here are two of many available options:

* Going back to the drawing board to find out what options are available to
solve the problem.
* Trying to make the document more clear that the current specification is
already what it should be so that we understand that there is no problem.

Thanks,

Steve.