[scikit-learn] Scikit-learn porting strategy

Laurent Julliard laurent at moldus.org
Tue Feb 5 03:19:43 EST 2019


Hi Andreas,

The person.year input is very valuable. This is a also the kind of
information I was looking for. The language would be Ruby. Now, it's true
that Ruby can already benefit from Scikit-learn through the PyCall
extension...

The point in my first question was also around the porting strategy: is it
like you can start small and get there step by step or you cannot make
anything work until you have completed say 50% of the code or more.

Eljay

On Mon, Feb 4, 2019 at 4:43 PM Andreas Mueller <t3kcit at gmail.com> wrote:

> Hi Eljay.
> Which language? And you want to reimplement it? How many full-time
> developers do you have for how many year? ;) Openhub estimates scikit-learn
> took 39 person-years:
> https://www.openhub.net/p/scikit-learn/estimated_cost
>
> I'm asking about the language because there are similar projects already
> existing in other languages, like Julia.
>
> Cheers,
> Andy
>
>
>
>
> On 2/4/19 10:28 AM, Laurent Julliard wrote:
>
> Hi everyone,
>
> If one were to start porting scikit-learn to another language what would
> be the plan to follow? I'm looking for directions that would say something
> like
> a) start with foundational components (e.g. numpy I guess)
> b) then port module A for a quick win,
> c) follow with modules B, C...
> d) the address the scipy dependent modules...
>
> Thank you for your help and advice.
>
> Eljay
> --
>
> Laurent Julliard
>
> _______________________________________________
> scikit-learn mailing listscikit-learn at python.orghttps://mail.python.org/mailman/listinfo/scikit-learn
>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
-- 

Laurent Julliard
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20190205/fe2abdd1/attachment.html>


More information about the scikit-learn mailing list