Am 25.07.2018 um 01:28 schrieb Allin Cottrell:
On Tue, 24 Jul 2018, Allin Cottrell wrote:
> On Tue, 24 Jul 2018, Sven Schreiber wrote:
>> OK, but doesn't this consitute a counterexample to the claim of full
>> backward compatibility?
That would not be computationally costly, and it would be maximally
backward-compatible, but I think my preference is to break
compatibility in that one respect, because what we're doing now is not
Yes, I tend to agree. (Remembering earlier discussions about the NA/nan
situation, I'd say better late than never, hehe.)
I'm wondering whether I should extract from the function packages a
mailing list of package authors for such occasions, to alert them about
architecture changes like this.
One point I'd forgotten is that comparable software doesn't
"NA * 0 = 0". Not that we want to slavishly follow R, but NA * 0 = NA
I think the main point is that this NA * 0 isn't really marketed or
mentioned in the gretl documentation, is it?
Apart from that, what's the status of stuff like table 16.1 in the
guide? It mentions "NA" for matrices instead of NaN, has this been wrong