On Monday, November 3, 2014 4:30:11 AM UTC-8, Sergey Kirpichev wrote:
>
> On Sun, Nov 02, 2014 at 06:56:31PM -0800, Richard Fateman wrote: 
> >    The Cathedral and the Bazaar essay 
> >    doesn't work if bugs do not become shallow with enough eyes. 
>
> I'm giving up.  Probably, I newer can understand why.  Esp., if there 
> is no difference for commercial support and it's obvious that "eyes" can 
> be of very different qualification. 
>
> >      Yes, but in first - you can do anything that can do a client of 
> >      closed-source vendors.  And much more (look to the code yourself, 
> pay 
> >      someone else for support, etc) 
> > 
> >    In principle. 
>
> No, as a matter of fact.  There is a lot of FOSS companies around. 
>
> >    Practically, are you willing to pay someone to become an 
> >    expert on (say) Maxima so that sympy can learn from it? 
>
> Practically, it could be much more easy for Maxima experts to 
> learn SymPy.  But yes, I don't see why the quoted above is 
> impossible: i.e. sponsored work (we have this, at least with 
> GSoC) or a commercial project, that rely on SymPy (at least I 
> have some invitations). 
>
> The bad news with SymPy - is that it's derivative could 
> be closed-source.  But I don't think such fork will be useful. 
>
> >      >      Apparently, Spain universities doesn't matter for the 
> >      >      Wolfram Research. 
> >      > 
> >      >    That sounds reasonable to me.  
> > 
> >      Are you sure that any academic institution does matter for Wolfram? 
> > 
> >    I don;t know what their market looks like, but I am quite sure that 
> the 
> >    people buying Mathematica for Univ. Calif ask the users at UC, and 
> for 
> >    the most part they don't are about bug reports enough to cancel 
> >    the re-licensing... 
>
> An how much these users do care about the Mathematica bugs?  I see 
> a lot of people from Cal, who use something different in their work, 
> incl. the Python and scipy.org stack. 
>

There is a faculty member at UC  (F. Perez) pushing scipy; he invented
ipython, I understand.  I am not sure how to characterize "a lot of people
at Cal".   There are 35,000 students. Maybe 1,400 faculty.  Perez is
a researcher in Brain Imaging, not in computer science or engineering or 
math.
So it may seem like a lot from where you sit.  I think it may not seem
like so much from where I sit.  I do agree with the general premise that
it would be nice to improve the quality of scientific computing support
from all perspectives, including programming languages.


> >    i agree.  It might be a problem with (say) addition of certain  big 
> >    numbers. 
>
> Or with caching, or... 
>
> >    This convention was certainly in wide use before 1961. 
>
> Better source? 
>

Maybe check with Bill Gosper, who has done lots of thinking about
summation.  This is really an "analytic continuation" of the simple indexed 
sum..
 

I checked with Gradshteyn and Rhyzik (1960, revised various times later), 
and they
define sum if n<m to be zero.  Somewhat to my surprise, but in fairness, I 
figured
I should report it.

Summation of Series,  edited by Jolley,  also 1960-ish (Dover) does not 
consider this issue.

nist's  dlmf web site doesn't seem to be concerned with finite summation.
disappointing.   I think the Knuth etc  Concrete Mathematics may have 
something to say,
but I'm not near a copy.  Knuth tends to talk about sum over a set, and not 
the
order of indexing.

The usual formal definition of a a sum with upper and lower limits, 
recursively, looks like this:

sum(f(i),i,a,a) = f(a)
sum(f(i),i,a,b) = f(b) + sum(f(i),i,a,b-1)    at least for b>a. Why not 
allow b>=a??

and  therefore  sum(f(i),i,a,b) =  sum(f(i),i,a,c)+ sum(f(i),i,c+1,b).

Now this latter formula can work for c<a.  Should work, in my opinion.


-- 
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/dd72c14e-207e-47b5-9512-cf04ad4a4a85%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to