Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Roadmap to v2 #130

Open
oxinabox opened this issue Oct 29, 2019 · 0 comments
Open

Roadmap to v2 #130

oxinabox opened this issue Oct 29, 2019 · 0 comments
Milestone

Comments

@oxinabox
Copy link
Member

Many repos have a roadmap to v1.
This plan goes a little further.

The idea is:

ChainRules/ChainRulesCore v1:

  • It pragmatically works.
  • It is completely usable, and is in use by multiple AutoDiff systems by the time we tag v1.
  • But it might play fast and lose with the math, like it will probably keep something like the current defintion of AbstractDifferential, which conflates the idea of scaling and adding.
  • It has some edge cases that we just bail out on and don't handle representing. This may include complex deriviatives, and it probably will include mutation/nonpure functions.
  • It will support Julia 1.0

ChainRules/ChainRulesCore v2:

  • The math will be tighter
  • After a while of use in public we will have identified the actual edge cases that matter of things we can't represent. and it will be enhanced to support those
  • It may not support julia 1.0, e.g. might depend on features added in 1.3

This roadmap needs some refining,
but this was the big picture @willtebbutt and I were talking about

@oxinabox oxinabox pinned this issue Oct 29, 2019
@nickrobinson251 nickrobinson251 added this to the v2 milestone Oct 29, 2019
@oxinabox oxinabox modified the milestones: v2, v1 Dec 19, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants