On Oct 25, 2013, at 8:43 PM, boris_stitnicky (Boris Stitnicky) wrote:

> @david_macmahon, mohawkjohn: Colon is busy, how about harassing % ?
>=20
> %s/1 1e6 step 12/ # %s would mean series, returning an enumerator

I don't think that offers any benefit over just creating a factory =
method for creating Ranges (augmented with a step size attribute).  For =
example

def _(first,last,step=3D1); Range.new(first,last,false,step); end

...which would be used like...

_(1,1e6,12)

Of course the "real" definition would probably use *args and be smart =
about 2 vs 3 args to allow step to be given as an optional 2nd argument.

Even still, using the colon to specify ranges is far more compact and =
already used in a variety of other languages for that purpose.  In =
addition to Matlab/Octave and Python, it also is used to specify ranges =
in R, Verilog, and to some degree even in Excel spreadsheets (for ranges =
of cells).

Is your objection to using the colon technical (e.g. impracticality of =
parsing) or philosophical?

Thanks,
Dave