Re: Bug in Sum causes index variable to remain set?

*To*: mathgroup at smc.vnet.net*Subject*: [mg33896] Re: [mg33838] Bug in Sum causes index variable to remain set?*From*: Andrzej Kozlowski <andrzej at platon.c.u-tokyo.ac.jp>*Date*: Sun, 21 Apr 2002 06:14:40 -0400 (EDT)*Sender*: owner-wri-mathgroup at wolfram.com

I don't know whether this is a bug or not, and it seems to me that this is just the sort of situation where we lack a clear definition of what constitutes a bug (Behaviour undesirable to some users? Unintended undesirable behaviour? Easily avoidable undesirable behaviour?). However, I can point out that this behaviour is a universal feature of Mathematica design in analogous cases. The same thing will happen if you replace Sum by Product, Table or Do. Integrate is not a good example because it does not assign values to the variable of integration (it just finds the antiderivative and evaluate limits). NIntegrate would have been a better example, but it actually also does not make assignments to an iterator. So bug or not it seems to be a feature (feature!) of Mathematica"s design, more precisely of the way Matheamtica uses iterators. Andrzej Kozlowski Toyama International University JAPAN http://platon.c.u-tokyo.ac.jp/andrzej/ On Friday, April 19, 2002, at 03:28 PM, Martin Jenkins wrote: > > Hello, I have found what appears to be a bug in Sum in Mathematica 4.0. > I stumbled across is trying to sum an expression with subscripts. > Anyway: > > In[1]:= n = 1; > > In[2]:= Sum[1, {f[x, 1], 1, 3}] > Out[2]= 3 > In[3]:= ?f > In[4]:= Sum[1, {f[x, n], 1, 3}] > Out[4]= 3 > In[5]:= ?f > > In[6]:= Clear[f]; > In[7]:= Sum[1, {f[x, 1*1], 1, 3}] > Out[7]= 3 > In[8]:= ?f > > The way that f gets set by the Sum to the final value is a bug is it > not? I would guess that Sum unsets the literal expression f[x, n] or > f[x, 1*1], rather than exactly what it set in the first place. I can > get around this by f[x,n]=. whenever I use it... but I would rather > not have a line with a comment saying "Mathematica bug work-around". > Integrate works fine under similar circumstances. > > Thanks, > Martin > > > >