Re: [office] Re: [office-comment] Re: sigmoid curve fitting in trend lines

[André Rebentisch <andre.rebentisch@arsaperta.com>]

Am 06.11.2012 16:12, schrieb robert_weir@us.ibm.com:
> 1) Does it make sense to specify requirements for numeric calculations
> that are not guaranteed to yield identical results?  I'd say yes, and we
> already do this.  OpenFormula does not mandate a particular floating
> point processing model, or precision.  So even basic arithmetic
> operations are not guaranteed to be identical, although in practice most
> implementations will. 

A general shortcoming of openformula is that earlier incarnations of the
spec included examples. In many cases removing the examples then
increased spec ambiguity because it is actually very difficult to
explain by words what is explained by a simple example.

Test suites won't harm.

Best,
A