Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.soft-sys.math.mathematica > #15166

Re: inconsistent refinement behavior

From Alex Krasnov <akrasnov@cory.eecs.berkeley.edu>
Newsgroups comp.soft-sys.math.mathematica
Subject Re: inconsistent refinement behavior
Date 2013-07-12 06:29 +0000
Message-ID <kro7o8$f19$1@smc.vnet.net> (permalink)
References <krj23l$1jf$1@smc.vnet.net> <20130711060248.8267169C4@smc.vnet.net>
Organization Time-Warner Telecom

Show all headers | View raw


Firstly, x==0 also implicitly assumes that x is in Reals, since 0 is in 
Reals, as the following examples demonstrate:

In:	Assuming[x==0, Refine[Element[x, Reals]]]
Out:	True

In:	Assuming[{Element[x, Reals], x==0}, Refine[Infinity/x]]
Out:	ComplexInfinity

Secondly, ComplexInfinity results from the unknown sign of x at 0, not any 
property of complex numbers.

Alex


On Thu, 11 Jul 2013, Helen Read wrote:

> There is nothing inconsistent about this. When you include the
> assumption x>0 or x>=0, you are implicitly assuming that x is an element
> of the Reals, so with either of these assumptions the result will be
> Infinity rather than ComplexInfinity.
>
> Recall that there is no ordering in the complex numbers.
>
> On 7/10/2013 3:22 AM, Alex Krasnov wrote:
>> The following behavior appears to be inconsistent:
>>
>> In:	Assuming[x==0, Refine[Infinity/x]]
>> Out:	ComplexInfinity
>>
>> In:	Assuming[x>0, Refine[Infinity/x]]
>> Out:	Infinity
>>
>> In:	Assuming[x>=0, Refine[Infinity/x]]
>> Out:	Infinity
>>
>> The third example should return unrefined given the first two examples. Is
>> there an explanation?
>>
>> Alex
>>
>
>
>
>

Back to comp.soft-sys.math.mathematica | Previous | Next — Previous in thread | Find similar | Unroll thread


Thread

inconsistent refinement behavior Alex Krasnov <akrasnov@cory.eecs.berkeley.edu> - 2013-07-10 07:22 +0000
  Re: inconsistent refinement behavior Helen Read <readhpr@gmail.com> - 2013-07-11 05:57 +0000
  Re: inconsistent refinement behavior Alex Krasnov <akrasnov@cory.eecs.berkeley.edu> - 2013-07-12 06:29 +0000

csiph-web