Re: inconsistency with Inequality testing and Floor
- To: mathgroup at smc.vnet.net
- Subject: [mg60037] Re: inconsistency with Inequality testing and Floor
- From: "Brett Patterson" <muckle.moose at gmail.com>
- Date: Tue, 30 Aug 2005 04:42:57 -0400 (EDT)
- References: <dek8a1$aao$1@smc.vnet.net><dep7p6$f1a$1@smc.vnet.net>
- Sender: owner-wri-mathgroup at wolfram.com
Hi Jean-Marc, Thanks for your response...but I don't really understand it! My problem is that the logical test (<) and the numerical operation Floor seem to give different and inconsistent results on the same input. Regardless of the reasons for that, due to machine representations, etc, I feel that these operations should be designed to be consistent. My input is the value of a variable x. The logical test tells me that x >= 1, but the result of Floor[x] is 0. This is surely unintuitive and inconsistent. In your example, setting $MaxExtraPrecision to a larger number did not change the value or Floor[x], and it does not change the result of x >=1. So I don't understand what you were trying to demonstrate by setting $MaxExtraPrecision to 1000. Cheers, Brett