Most wanted, 2019 edition

[JJ]

It's 2019 and things have changed. Are there any modules out there that can be considered most wanted? Which modules would you want to see in Perl 6?
Let's discuss here and eventually add them to the Most Wanted list.

[JJ]

Protocol buffers are becoming incredibly active lately. It might be convenient to work with them.

[bduggan]

I've started a grammar :) https://github.com/bduggan/p6-protobuf

[Altai-man]

NumPy/SciPy analogs are missing and I often see newcomers asking about it.

[Xliff]

NumPy/SciPy are great ideas, but are Shaped arrays mature enough to implement it?

[JJ]

El sáb., 6 abr. 2019 a las 16:33, Xliff (<[email protected]>) escribió:

NumPy/SciPy are great ideas, but are Shaped arrays mature enough to implement it?

What will you need? I think the main thing with NumPy is fast arithmetics. Shapes work, although they are kinda slower.

[taboege]

@bduggan : Have you seen this Protobuf module stub? Maybe you can borrow some bits from it.

[AlexDaniel]

Yes, I'd love to see a numpy alternative. Especially in a combination with OpenCV.

are Shaped arrays mature enough to implement it

I don't think we'll be able to find and fix the limitations until someone actually starts a module that heavily relies on shaped arrays.

[bduggan]

Thanks @taboege -- no, I hadn't seen that!

[titsuki]

Hi,
My most wanted list in 2019 is here:

• Bindings for MxNet: MxNet is a deep learning framework. Comparing the other framework (e.g., tensorflow, chainer), it seems relatively easier to implement bindings since this framework already supports Perl5.
• Swig: is a glue code generator for bindings. I have seen the github issue that requests adding support for Perl 6 (c.f., Add support for Raku swig/swig#1245)
• Bindings for Stan: Stan is a state-of-the-art platform for statistical modeling and high-performance statistical computation.

Cheers,