On Mon, Feb 13, 2023 at 1:08 AM Frederick Virchanza Gotham wrote:
>
> We already do this. We already have language features in C++ that hide
> extra data inside each object.


I really feel that I need to make a point here -- and not a point
pertaining solely to the proposal in question.

A few people here on the mailing list are very caught up in evaluating
the uniqueness of any new idea, and once the uniqueness has been
pointed out, other people then start trying to find precedents in
order to justify the new idea. This really is a rotten way of
thinking. A new idea can be totally new and unique, without any
precedents, and still be a good idea.

I have been able to give the example on this occasion of adding a
virtual destructor to a class, the effect being that it causes the
size of each individual object to increase. In some people's minds,
this is seen as precedent that goes toward justifying my new idea. But
why should I need a precedent? Even if no other language feature
caused an opaque increase in the size of each object, why can't I
suggest a new features that does?

Like with my suggestion last year of 'continuity methods', people were
really adverse to it quite simply doing something that hadn't been
done before. Resistance to change can at times be resistance to
progress.

Sometimes I really think that this mailing list is really missing out
on fluidic constructive creative conversation. People are being
careful and conservative, and we're missing out on the juicy stuff
because of it.