Kevin Reid at 16:38, 7 March 2009

Kevin Reid — Sat, 07 Mar 2009 16:38:11 GMT

New page

The '''<code>Float64</code>''' type is E's primary floating-point [[number]] type. Its members are IEEE 64-bit binary floating-point numbers. ({{XXX}} is this appropriate definition language?) Float64 is one of the [[E literal types]].

The [[universal scope]] contains <code>float64</code> (no capital), the guard for this type.

? float64
# value: float64

== Protocol ==

===Number protocol===
Float64s implement the [[number protocol]]. ({{XXX}} figure out if there's benefit in templating to include generic number tests here)

{{instance msgdoc|abs|0||[[Float64]]}}

As specified in the [[number protocol]]. Specific tests:

? 0.0.abs()
# value: 0.0

? 0.5.abs()
# value: 0.5

? (-0.0).abs()
# value: -0.0

? (-0.5).abs()
# value: 0.5

{{XXX}} is (-0.0).abs() supposed to be negative?

{{instance msgdoc|sin|0||[[Float64]]}}
Returns the trigonometric sine of this number as an angle in radians.

? 0.0.sin()
# value: 0.0

{{instance msgdoc|cos|0||[[Float64]]}}

Returns the trigonometric cosine of this number as an angle in radians.

? 0.0.cos()
# value: 1.0

{{instance msgdoc|tan|0||[[Float64]]}}

Returns the trigonometric tangent of this number as an angle in radians.

? 0.0.tan()
# value: 0.0

{{XXX}} test nonzero parameters for trig functions

{{XXX}} write tests for all other methods (get from eocl and eojava)

{{XXX}} Design issue: Is the canonical initial letter case for this type name (and int, boolean) upper or lower? [[Category:Unresolved design issues]]

[[Category:ELib specification]]

Float64 - Revision history

Kevin Reid at 16:38, 7 March 2009