Ahh, I understand your position better now. But am still against it. I
think the names t. etc. are more valuable for use as new primitives than
for backward compatibility.
Backward compatibility is important. But we need to draw lines from time to
time.
On Sat, Apr 2, 2022 at 12:49 PM Raul Miller <rauldmil...@gmail.com> wrote:
> On Sat, Apr 2, 2022 at 10:32 AM Eric Iverson <eric.b.iver...@gmail.com>
> wrote:
> > I am quite sure that we will not revisit the decision on those
> primitives.
> > They, and similar, are much better dealt with as library routines. and
> > that is where effort should go. If they are worth having, then they
> warrant
> > the effort of a library implementation.
>
> I was not suggesting that we remove the library routines.
>
> I was suggesting that the reinstated primitives call the library routines.
>
> In other words, the d. implementation would look something like this
> (notional, untested code):
>
> DF2(jtddot) {
> A ddotconj;
> ASSERT(ddotconj=unquote(a,w,jtfindnameinscript(jt,"~addons/math/calculus/calculus.ijs","deriv_jcalculus_",CONJ),EVNONCE));
> FAV(ddotconj)->lc=FAV(self)->id;
> R ddotconj;
> }
>
> Thanks,
>
> --
> Raul
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
