Re: [Gretl-users] New plot for arma models
On Sun, 26 Mar 2017, Riccardo (Jack) Lucchetti wrote:
On Sun, 26 Mar 2017, Clive Nicholas wrote:
> Confirmed on my Linux Mint 18.3 system after updating to 2017a.
>
> Very good work, Jack!
Thanks, on behalf of Allin too, who actually wrote most of the code.
This may be more than people want know, but I'm tempted to expand a
little on the (IMO) excellent division of labour that stands behind
gretl development, and of which this is a not untypical example.
1) Jack proposes a very good idea, initially in the form of a
hansl-scripted "prototype".
2) I can see it's a good idea, but I have trouble following Jack's
hansl script -- not because it's obfuscated (= deliberately obscure)
but just because Jack can see 3 or 4 moves ahead on the chess-board
where most of us (well, I speak for myself) can only see 1 or 2
moves at a time.
3) Jack produces C-code to expand on the hansl prototype. Since
coding in C forces you to go relatively step-by-step I now "get it"
more definitely.
4) There then arise practical questions of how exactly the required
information can be got into the right place at the right time, in
the context of gretl's C code. Since I know that code like the back
of my hand, I can contribute something useful at this point.
5) We then go back and forth a few times, trying to iron out any
problems that may arise (e.g. special cases not handled).
6) If all goes well we end up with something that implements Jack's
original idea and is reasonably robust across a wide range of
possible cases.
Allin