n(xxx) - bits or digits of precision?!

[Dima Pasechnik]

sage: pi.n?

Docstring:

   Return a numerical approximation this symbolic expression as either

   a real or complex number with at least the requested number of bits

   or digits of precision.

this is clearly ambiguous; can it be clarified?

(it bites e.g. if one wants to create meaningful doctests, such as on #18888)

[Volker Braun]

Whats ambiguous?

The call signature is missing, thats a bug. Ideally it would be documented in an INPUT block.

[Dima Pasechnik]

On Saturday, March 12, 2016 at 7:47:46 PM UTC, Volker Braun wrote:

Whats ambiguous?

 (decimal) digits or bits?

a part of my confusion comes from a puzzling (OSX-specific) error on #18888.

A Mathematica object has .n() and .N(), and they are not synonyms.

(n() is Sage's method, N() is Mathematica's method).

However it seems that on OSX N() acts as n(), I don't understand why. Some case-insensitive stuff, I guess.

[Volker Braun]

Default is bits, the digits= keyword argument its for decimal digits instead. 

Signature: pi.n(self, prec=None, digits=None, algorithm=None):

Docstring:

   Return a numerical approximation this symbolic expression as either

   a real or complex number with at least the requested number of bits

   or digits of precision.

   EXAMPLES:

      sage: sin(x).subs(x=5).n()

      -0.958924274663138

      sage: sin(x).subs(x=5).n(100)

      -0.95892427466313846889315440616

      sage: sin(x).subs(x=5).n(digits=50)

      -0.95892427466313846889315440615599397335246154396460