Re: matlab rotx, roty, rotz, function and equivalent in octave 4.0 or packages

[JuanPi]

On Mon, Apr 10, 2017 at 4:50 PM, Nicholas Jankowski <address@hidden> wrote:
> On Mon, Apr 10, 2017 at 9:28 AM, JuanPi <address@hidden> wrote:
>>
>> On Mon, Apr 10, 2017 at 12:51 PM, Nicholas Jankowski
>> <address@hidden> wrote:
>> > On Apr 9, 2017 8:32 PM, "Nicholas Jankowski" <address@hidden>
>> > wrote:
>> >
>> > On Fri, Apr 7, 2017 at 9:42 AM, Olaf Till <address@hidden> wrote:
>> >>
>> >> On Fri, Apr 07, 2017 at 07:17:24AM -0400, RT wrote:
>> >> > Greetings,  does Octave 4.0 or one of it's packages have an
>> >> > equivalent
>> >> > to
>> >> > Matlabs rotx, roty, rotz, function
>> >> > https://www.mathworks.com/help/phased/ref/rotx.html
>> >>
>> >> The geometry package has similar functions. CCing the maintainer.
>> >>
>> >>
>> >
>> > with a quick peek geometry seems to have some similar functionality but
>> > not
>> > matlab compatible versions of those functions.
>> >
>> > createRotation would recreate rotz if it took degrees by default.
>> >
>> > the 3d rotation functions createRotationOx, createRotationOy, and
>> > createRotationOz would work if they weren't 4x4 matrices.  (why are they
>> > 4x4? what's the 4th value, length which stays invariant?
>> >
>> > aliases could easily be created for those to implement the rot
>> > functions.
>> > That said, the actual implementation is so simple it's almost a shame to
>> > add
>> > the overhead of the extra function call to create them.
>> >
>> > in Matlab, rotx, roty, and rotz appear to be part of the Phased Array
>> > System
>> > Toolbox.
>> >
>> >>> which rotx
>> > C:\Program Files\MATLAB\R2016b\toolbox\phased\phased\rotx.m
>> >
>> >
>> >
>> >
>> > FYI to close the loop on this I did submit a set of functions to the bug
>> > tracker for this.
>> >
>> > http://savannah.gnu.org/bugs/?50772
>>
>> Hi,
>>
>> if your functions get into core, I will definitely make the geometry
>> functions just wrappers to them.
>> The functions return 3x3 (2d) and 4x4(3d) matrices because they
>> implement affine transformations using homogeneous coordinates[1]. In
>> this way, e.g., it is easy to define a rotation around an arbitrary
>> point in space.
>>
>> [1]: https://en.wikipedia.org/wiki/Homogeneous_coordinates
>>
>> --
>
>
> And now Homogenous Coord's checks the box for 'learn one new thing each
> day'.  :)
>
> regarding the functions, just so I'm not misunderstanding: I put the patch
> together as an add to the geometry toolbox, as it seemed to be a similar set
> of functions & purposes. Did you think these functions should be core Octave
> functions instead?
>

When I was using matlab rot* where on core, I would say these
functions go to geometry package of core. Check in the mailing list,
we can implement them optimized and put them there. I would prefer
they are in core. Would you ask on the mialinglist?


-- 
JuanPi Carbajal
https://goo.gl/ayiJzi
Public GnuPG key: 9C5B72BF
-----
"Why is thought, being a secretion of the brain, more wonderful than
gravity, a property of matter?"
- C. Darwin