Announcement: Diofant 0.8.0 release (SymPy's fork)

[Francesco Bonazzi]

Hi everyone,

Sergey Kirpichev has asked me to publish the announcement for the Diofant 0.8.0 release. Here it is:

Diofant is a Python library for symbolic mathematics, it can run on any

computer with Python 3.4 or above.  Installation instructions and
release notes (with detailed list of closed issues) are available from
the online documentation (http://diofant.rtfd.io/en/latest/).  The
highlights for this release: assumption system improvements, the Gruntz
algorithm, better testing with pytest, flake8 and coverage tools.

This is a beta release.  Now I'll focus on the documentation, adding tests -
and I appreciate any feedback, esp. on this point. I can't (and I'll not,
anyway) answer you here, so please use issue tracker on the project
website (https://github.com/diofant/diofant) if you expect some comments
from me (don't use my email - the Diofant project will be public).

I thank the SymPy development team for their work: several PRs with
backported bugfixes were merged after this fork was started.  Also, I
thank Aaron Meurer, Colin B Macdonald and Raoul Bourquin for permissions
to merge their work, that is not in SymPy yet.  I thank Francesco Bonazzi,
who posted this announcement to the SymPy maillist on my behalf, I hope
that both projects could benefit from this.

Few notes on why this fork was started at all.  On a first look, the reason
is just social.  (In my view,) SymPy's owners care too much about the
attraction of newcomers (bigger number - better), but they are unable to
keep old developers in the project.  On the other hand, this policy has
direct technical consequences and SymPy's development principle "merge
now - improve later" makes this situation worse.  As SymPy has only few
people who know the old codebase well - related bugreports mostly stay
unfixed, even with the "wrong result" label.  Also, it's not surprising
to observe that SymPy is now overbloated with unmaintained code,
inconsistent interfaces, etc... - that no one could improve.

I hope that this situation will get better in my project if contributions
are judged by their technical merit.  And it is already a little better
right now, as you can see from the numerous bugfixes on series, limits and
assumptions (now it's a well tested and consistent subsystem).  Next
release will be centered on solvers (and, perhaps, sets module) with
goal to provide uniform and expressive system, that can represent
solutions of generic algebraic equations.