thanks for the explanation, and yes, I agree that there is a problem - I 
have no good idea to fix it, though.  In fact, looking at my code with this 
in mind, it looks particularly stupid: already the first call to "order" 
should be avoided, I should simply check whether g has a constant term.

After that, we are probably out of luck, because, as you said, recognising 
the zero power series is not possible.

On the other hand, if g is zero, calling code should not expect divide to 
succeed anyway.

I admit I am rather on the "immediate progress" side

Martin

Am Mittwoch, 23. Januar 2019 17:50:51 UTC+1 schrieb Waldek Hebisch:
>
> Martin R wrote: 
> > 
> > Dear all! 
> > 
> > the question fits into the title :-) 
> > 
>
> Does not fit: it is truncated by many viewers.  Fixed: 
>
> > shouldn't UTS(F, x, 0)  export EuclideanDomain if F is 
> > a Field? 
>
> That is a tricky question.  I see that you gave an 
> implementation, but there is a core problem: 
> EuclideanDomain implies BasicType and this already 
> is a lie.  That is equality for power series is 
> noncomputable.  Which means that generic algorithms 
> may have very poor performance or fail.  To explain 
> it better, consider Gaussian elimination applied 
> to invertible matrix of power series.  Generic algorithm 
> will struggle trying to find out if an element is suitable 
> pivot.  Algorithm aware of power series will try increasing orders 
> and checking all elements of a column for given order. 
> In this way it will find pivot of best order, without 
> wasting time examining many entries of zero or nearly 
> zero series. 
>
> So question is if after lying several times we want 
> to keep lying, or if we want to clean up things. 
> Pragmatically lying may look attractive as it gives 
> immediate rewards, while punishment is delayed. 
> OTOH most had to fix bugs comes from original developers 
> ignoring problems and opting for immediate "progress". 
>
> -- 
>                               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].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/fricas-devel/5329f69d-9b98-4333-a70e-acc438c25f69%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to