PDL 2.038

[Henning Glawe]

Moin Ed,

I am also explicitly including Bas Couwenberg <sebastic at debian.org>, as he
was doing the latest uploads of PDL to Debian-Exprimental and can provide
more insight...

On Tue, Apr 20, 2021 at 07:27:19PM +0000, Ed . wrote:
> Since PDL is vastly unlikely to ever have any security issues that need fixes,
> thank you for dealing with that issue. Also, I actually completely support such
> a stability policy, since I am also a consumer of such distributions. If
> Bullseye is already frozen, so be it.

You never know when it comes to security issues... especially, when someone
runs your code in the context of a website ;)

> https://tracker.debian.org/pkg/pdl implies experimental is tracking the latest
> version, so that seems Ok. If the Debian Perl Group can help me help them (if
> such is needed), I’ll be pleased to assist.

But note that "Experimental" means "not even yet in Unstable/Sid"; so the
more recent PDL versions (which include a "proper" complex datatype) may only
available to the users in more than 2 years.


> I hope numpy is serving your needs well in your current work! Can you either
> give me a very brief idea of what makes it preferable to PDL, or maybe just
> point me at a page that already has such information?

For myself, switching to python/numpy was less a matter of love than of
necessity.

I wrote significant parts of the high-level evalution code im my
PhD in Perl/PDL.

But nowadays, python and numpy are more common in the physics community than
perl/pdl, so the former is also more frequently taught to students than
the latter.

And we have to adapt... at least it is not Matlab ;)


-- 
c u
henning