On Tuesday, 24 March 2015 22:25:04 UTC+5:30, Aaron Meurer wrote:
>
> I think you guys are perhaps talking past each other. To Ondrej, 
> "series" means "series approximation", whereas to "Sartaj" it menas 
> "formal power series". Both are mathematically related, but from a CAS 
> point of view they are quite different.
>
 
Oh, I see. I am inclined towards infinite representation of Formal Power 
Series. But it is easy to represent it in truncated form.
  

> Formal power series is something that would be nice for SymPy to have, 
> but my understanding is that the summation algorithms will need to 
> improve a lot for you to be able to compute anything nontrivial (like 
> Cauchy products). 
>
 
Yes, you are correct summation algorithm need to be improved, some formulas 
are quite involved. But it will be nice to use ring_series for various 
operations for truncated series. When summations algorithm improve, 
non-trivial cases can be implemented for infinite series.

Sartaj Singh

Aaron Meurer 
>
> On Mon, Mar 23, 2015 at 7:12 PM, Ondřej Čertík <[email protected] 
> <javascript:>> wrote: 
> > On Mon, Mar 23, 2015 at 1:41 AM, Sartaj Singh <[email protected] 
> <javascript:>> wrote: 
> >> It looks like ring_series, works only for finite cases, can it be 
> extended 
> >> for infinite case? SeriesData will just zip two sequences together. It 
> is 
> >> more general as variables can even be of the form sin(n*x), and not 
> only x. 
> >> In the main FormalPowerSeries class, I plan to provide an infinite 
> >> representation of series. 
> > 
> > I think the ring series can only be used for n terms, i.e. if you have 
> > some expression, like sin(x)*cos(x), then you first expand sin(x) and 
> > cos(x) into n-terms, and then you multiply them out, and cut the 
> > result appropriately (at n-terms in this case). 
> > 
> > But you can use any "n". How do you store infinite number of terms in 
> > your approach? 
> > 
> >> Ring_series is undoubtedly fast, so it is a plus to use it in my 
> >> implementation. I have glanced over the code of ring_series, and it may 
> be 
> >> modeled into the structure I propose, i.e based on sequences. The 
> algorithms 
> >> in ring_series can be used for manipulating truncated form of the 
> series. 
> >> 
> >> In summary, I think SeriesData should be left as a higher abstraction, 
> and 
> >> operations for truncated series can be added in SeriesX or in 
> >> FormalPowerSeries. Let me know what you think. 
> > 
> > That would work. 
> > 
> > Ondrej 
> > 
> > -- 
> > 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 [email protected] <javascript:>. 
> > To post to this group, send email to [email protected] 
> <javascript:>. 
> > Visit this group at http://groups.google.com/group/sympy. 
> > To view this discussion on the web visit 
> https://groups.google.com/d/msgid/sympy/CADDwiVBTqTBBFCUUu_pgXNyvAibPU2ObX3_JMVQac8YAk1MtiQ%40mail.gmail.com.
>  
>
> > For more options, visit https://groups.google.com/d/optout. 
>

-- 
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 [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sympy.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sympy/0fadbdc4-41ea-4735-a9f5-83e18395d746%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to