On Sat, Aug 25, 2012 at 7:40 PM, Michael D Godfrey
<address@hidden <mailto:address@hidden>> wrote:
On 08/25/2012 09:55 PM, Ed Meyer wrote:
Just because it has kept the order of x(2) and x(3) does not
necessarily mean it is a stable sort;
an unstable sort can do either but a stable sort is guaranteed to
keep the original order.
--
Ed Meyer
Yes. But, this thread started from an error on my part. I had thought
that Ben's fix for interp1 was based on the Octave sort not being
stable.
The real problem was with the direction of the discontinuity mechanism
in interp1.
I am pretty sure that Octave sort is stable, but this could be checked
more, of course.
Michael
Comments in the code indicate it is a stable search; the function is
called binarysort but I've never heard of a binary sort so I'm not sure
what it's doing :-(
I think I'll run a bunch of tests to see if it ever does an unstable sort.