Currently most things planned for next release is done.
However, there are several problems with integration
of algebraic functions:
- at least two bugs with handling algebraic constants
during integration (I have fixes for two)
- the possibleOrder routine sometimes returns wrong order
(this shows for Masser and Zannier example included
in recent Davenport papar)
- possibleOrder goes info infinite loop searching for
good reduction (Euler example posted on sci.math.symbolic)
- lack of support for restart after division by zero during
reduction (we can limit effect by using more randomness)
Note: they lead to looping, internal errors or wrong
results so are much worse than unimplemented cases.
There is also unfinished integration code that I wanted to
finish and include.
I would like to fix bugs before adding new code to integrator.
However, I am afraid that fix will take substantial time.
First, it requires rather time-consuming checks at various
stages to find out what went wrong. Second, fixing problem
in one place does not mean that overall result is better:
sometimes two bugs (partially) cancel and fixing only one
causes regressions, so one have to look out at possible
undesired consequnces.
ATM I am looking at the bugs, but I consder shipping next
release without fixing them. More precisely, randomness
workaround is rather safe so we could ship it. But
properly handling other issues may delay release too
much.
--
Waldek Hebisch
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/fricas-devel.
For more options, visit https://groups.google.com/d/optout.
