Re: [sympy] Blog post about the the future of SymPy

2023-08-16 Thread Aaron Meurer
On Wed, Aug 16, 2023 at 12:08 PM Aaron Meurer  wrote:
>
> Thank you for the blog post Oscar. I submitted it to Hacker News as
> well https://news.ycombinator.com/item?id=37150867
>
> I agree with all the things you wrote. I agree with the idea of having
> computation subsystems that are targeted and fast. This concept is
> related to what James Powell calls a "restricted computational
> domain". See https://twitter.com/dontusethiscode/status/1362809888360005636
> (unfortunately you need to be signed into Twitter to read that whole
> thread. If you don't have an account you can read it here
> https://web.archive.org/web/20210219170345/https://twitter.com/dontusethiscode/status/1362809888360005636).

By the way, this talk in particular by James is very relevant to your
work https://www.youtube.com/watch?v=Ix04KpZiUA8. I'd say your
approach to making SymPy faster exactly lines up with the general
framework for improving Python performance that he outlines in that
talk.

Aaron Meurer

>
> One thing I like about the polys module is that everything is
> organized in layers, where the lower layers operate directly on data
> structures and the top layer is designed for user interaction. The top
> layer has things that are a little slower but more user friendly, like
> type checking, which the bottom layers do not have.
>
> One of the reasons SymPy ended up the way it is is that there's really
> only one symbolic system. This has made things very simple and makes
> the whole thing quite easy for developers to work on, but it also
> means that there's no distinction between lower level fast data
> structures and higher level data structures. And in particular, it
> means that slower abstractions like automatic evaluation have leaked
> in which really should only ever be part of the top level layers.
>
> I look forward to reading the followup blog posts.
>
> Aaron Meurer
>
> On Wed, Aug 16, 2023 at 9:14 AM Oscar Benjamin
>  wrote:
> >
> > Hi all,
> >
> > I have written a blog post about recent work on SymPy and what I think
> > should be future development plans:
> >
> > https://oscarbenjamin.github.io/blog/czi/post1.html
> >
> > This is the first in a series of posts and in this one I describe what
> > I think are the core parts of SymPy and outline from a high level what
> > I think should be done going forwards. Subsequent posts will describe
> > specific components in more detail.
> >
> > Oscar
> >
> > --
> > You received this message because you are subscribed to the Google Groups 
> > "sympy" group.
> > To unsubscribe from this group and stop receiving emails from it, send an 
> > email to sympy+unsubscr...@googlegroups.com.
> > To view this discussion on the web visit 
> > https://groups.google.com/d/msgid/sympy/CAHVvXxTNyu8eXUZM_SesY7DAkUbpeFd--55wzkF4opfZ1jYeUg%40mail.gmail.com.

-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sympy+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sympy/CAKgW%3D6Jyu1tK16aA5cuL%3Dh08hH4pCefDv8hmVf9-Zo6TiyiovA%40mail.gmail.com.


Re: [sympy] Blog post about the the future of SymPy

2023-08-16 Thread Aaron Meurer
Thank you for the blog post Oscar. I submitted it to Hacker News as
well https://news.ycombinator.com/item?id=37150867

I agree with all the things you wrote. I agree with the idea of having
computation subsystems that are targeted and fast. This concept is
related to what James Powell calls a "restricted computational
domain". See https://twitter.com/dontusethiscode/status/1362809888360005636
(unfortunately you need to be signed into Twitter to read that whole
thread. If you don't have an account you can read it here
https://web.archive.org/web/20210219170345/https://twitter.com/dontusethiscode/status/1362809888360005636).

One thing I like about the polys module is that everything is
organized in layers, where the lower layers operate directly on data
structures and the top layer is designed for user interaction. The top
layer has things that are a little slower but more user friendly, like
type checking, which the bottom layers do not have.

One of the reasons SymPy ended up the way it is is that there's really
only one symbolic system. This has made things very simple and makes
the whole thing quite easy for developers to work on, but it also
means that there's no distinction between lower level fast data
structures and higher level data structures. And in particular, it
means that slower abstractions like automatic evaluation have leaked
in which really should only ever be part of the top level layers.

I look forward to reading the followup blog posts.

Aaron Meurer

On Wed, Aug 16, 2023 at 9:14 AM Oscar Benjamin
 wrote:
>
> Hi all,
>
> I have written a blog post about recent work on SymPy and what I think
> should be future development plans:
>
> https://oscarbenjamin.github.io/blog/czi/post1.html
>
> This is the first in a series of posts and in this one I describe what
> I think are the core parts of SymPy and outline from a high level what
> I think should be done going forwards. Subsequent posts will describe
> specific components in more detail.
>
> Oscar
>
> --
> You received this message because you are subscribed to the Google Groups 
> "sympy" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to sympy+unsubscr...@googlegroups.com.
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/sympy/CAHVvXxTNyu8eXUZM_SesY7DAkUbpeFd--55wzkF4opfZ1jYeUg%40mail.gmail.com.

-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sympy+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sympy/CAKgW%3D6KXWRdsSYU9EmUmWTnOhCxhzB%3D_x38uJoE9ZpJjfJBUhQ%40mail.gmail.com.


[sympy] Blog post about the the future of SymPy

2023-08-16 Thread Oscar Benjamin
Hi all,

I have written a blog post about recent work on SymPy and what I think
should be future development plans:

https://oscarbenjamin.github.io/blog/czi/post1.html

This is the first in a series of posts and in this one I describe what
I think are the core parts of SymPy and outline from a high level what
I think should be done going forwards. Subsequent posts will describe
specific components in more detail.

Oscar

-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sympy+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sympy/CAHVvXxTNyu8eXUZM_SesY7DAkUbpeFd--55wzkF4opfZ1jYeUg%40mail.gmail.com.