On Tuesday, 17 March 2020 at 00:48:24 UTC, ahmat wrote:
> Hi everyone,
> I use mainly Python for scientific computing and I want to switch to D but I can't find good libraries as replacement for pandas, matplotlib, scipy, ...
> Are there plans to make D better in this area?

Do you mean writing D libraries that do these things? Probably not. That would take a lot of resources and would duplicate work already done.

I've done a fair amount of this type of work myself, and IMO D is great if you don't mind wrapping C libraries like GSL. That's really all Python did in the beginning - it was just a glue language. I don't find it all that time consuming due to D's great interoperability with other languages, and I'm a lot more productive in D than in those other languages. For me, the cost-benefit analysis works out in favor of D.

If you want something polished, something that "just works", you're better off using Julia. If you want to write libraries to make D as convenient to use as Python, it will be welcome. Don't hold your breath waiting for others to deliver something. I'd do it if I had the time...

On Tuesday, 17 March 2020 at 18:16:17 UTC, bachmeier wrote:
> On Tuesday, 17 March 2020 at 00:48:24 UTC, ahmat wrote:
>> Hi everyone,
>> I use mainly Python for scientific computing and I want to switch to D but I can't find good libraries as replacement for pandas, matplotlib, scipy, ...
>> Are there plans to make D better in this area?
>
> Do you mean writing D libraries that do these things? Probably not. That would take a lot of resources and would duplicate work already done.
>
> I've done a fair amount of this type of work myself, and IMO D is great if you don't mind wrapping C libraries like GSL. That's really all Python did in the beginning - it was just a glue language. I don't find it all that time consuming due to D's great interoperability with other languages, and I'm a lot more productive in D than in those other languages. For me, the cost-benefit analysis works out in favor of D.
>
> If you want something polished, something that "just works", you're better off using Julia. If you want to write libraries to make D as convenient to use as Python, it will be welcome. Don't hold your breath waiting for others to deliver something. I'd do it if I had the time...

Python just work and that’s what a beginner need from a language. You can just run jupyter notebook, import pandas and load a csv dataset to a pandas DataFrame. But if you want to write high performance code or use advanced things it’s better to use another language like C++ and maybe D.
D is great for wrapping C libraries but many python libraries are binding of C++. How is the status of interoperability between D and C++ ? I am available to help if someone is interested in writing some useful libraries in D.

I think Julia is fast and one doesn’t need numpy and others libraries to write high performance. It’s gaining users in Academia and research even in deep learning.

On Tuesday, 17 March 2020 at 11:45:20 UTC, maarten van damme wrote:
> Have you considered julia? I like D and even tried to sneak it into as many uni assignments as possible but the package ecosystem isn't really there. You will spend a considerable amount of time on non-scientific-computing tasks such as writing interfaces for other c libraries.
>
> Op di 17 mrt. 2020 om 01:50 schreef ahmat via Digitalmars-d < digitalmars-d@puremagic.com>:
>
>> Hi everyone,
>> I use mainly Python for scientific computing and I want to switch
>> to D but I can't find good libraries as replacement for pandas,
>> matplotlib, scipy, ...
>> Are there plans to make D better in this area?

I have tried Julia and it’s great for machine learning (Flux.jl) and automatic differentiation. I am and undergrad but I will consider using it in the future for R&D.

On Tuesday, 17 March 2020 at 18:57:33 UTC, Ahmat wrote:
> [snip]
>
> Python just work and that’s what a beginner need from a language. You can just run jupyter notebook, import pandas and load a csv dataset to a pandas DataFrame.

I generally find R's data frames easier to use than pandas.

On Tuesday, 17 March 2020 at 18:57:33 UTC, Ahmat wrote:

> D is great for wrapping C libraries but many python libraries are binding of C++. How is the status of interoperability between D and C++ ?

Getting better, but I don't use it. There's actually a "backdoor" to C++ numerical computing libraries that I use. Many C++ libraries can be called from R, but the R interface is nothing more than a C interface that can be called directly from D, and doesn't actually involve R in any way. So for scientific computing, I'd say it's pretty darn good.

> I think Julia is fast and one doesn’t need numpy and others libraries to write high performance. It’s gaining users in Academia and research even in deep learning.

Well, it has something D doesn't: manpower. Although I prefer D, there's certainly nothing wrong with Julia. I plan to work on calling Julia from D when I have a reason to do so.

On Tuesday, 17 March 2020 at 18:57:33 UTC, Ahmat wrote:

[...]

One other thing I'll add is that my bigger goal is to move us away from encouraging people to choose a particular language. D's strength is that any code you write can be called from any other language. For me, that's far more important than the short-term goal of writing code that does what I need it to do this week. I've done something useful when I write code in D and others use it from their language of choice without ever having heard of D. If someone wants to use Python, R, Matlab, Octave, Fortran, or Cobol, I still want them to be able to take advantage of the code I've already written.

On Tuesday, 17 March 2020 at 19:36:42 UTC, bachmeier wrote:
> [snip]
>
> Well, it has something D doesn't: manpower. Although I prefer D, there's certainly nothing wrong with Julia. I plan to work on calling Julia from D when I have a reason to do so.

In addition to dpp, symmetry has done some work on this here
https://github.com/symmetryinvestments/juliad