Re: Code Submission

[Daniel Carrera]

Cool. I wasn't aware of the alias feature. That's much better.

When can I expect to see the new features in SciRuby?

Daniel.

On Tuesday, July 31, 2012 at 4:47 AM, Carlos Agarie <carlos...@gmail.com> wrote:

Daniel,

In your zeros/zeroes method, you could simply use 

alias :zeroes :zeros

To create an alias (zeroes) for the method zeros. The same for identity/eye. Also, great idea on the in-line constructor! It'd be awesome to show some simple calculations using it on the wiki. :)

And you should send a pull request for the main repo (SciRuby/nmatrix) or someone with commit access could do it.

On Wednesday, July 18, 2012 5:25:15 PM UTC-3, Daniel Carrera wrote:

Attached is a file that contains the function shortcuts with documentation and updated so that only eye() and zeros() have the option to change the storage type.

Cheers,

Daniel.

On Wednesday, July 18, 2012 at 6:17 PM, John Woods <john.o...@gmail.com> wrote:

I agree.

On Wed, Jul 18, 2012 at 8:43 AM, Daniel Carrera <dcar...@hush.com> wrote:

The more I think about it the more I believe that only eye() and zeros() should allow the user to choose the storage format. For eye() and zeros() it makes perfect sense to choose either Yale or List formats, but for one() and seq() I really don't see how it could make any sense.

Cheers,

Daniel.

On Wednesday, July 18, 2012 at 8:52 AM, Daniel Carrera <dcar...@hush.com> wrote:

Certainly the eye() method... and I suppose zeros() as well.

Ok, new attachment. I added a storage type option to most functions, including ones() , seq()  and N[] where it really doesn't make sense. But I hesitated to add that option to the linspace() and indgen() functions which are sort of clones of MATLAB and IDL. I could add them, but does it really make sense to have a linspace() with a :list or :yale storage?

I'm actually tempted to say that only zeros() and eye() should have the storage option. Those are the only functions that could possibly benefit from that, no?

Cheers,

Daniel.

On Tuesday, July 17, 2012 at 10:59 PM, John Woods <john.o...@gmail.com> wrote:

They do, I think, at the very least for the eye method.

Just got my eyes dilated, so I'll have to look later.

Cheers,

John

[John Woods]

Working on it this week. Our priorities are as follows:

- Getting the new C++ templates compiling

- Getting slicing working (mostly done)

- Adding your shortcut functions

[Carlos Agarie]

John, I can commit Daniel's changes and send you a pull request at SciRuby's repo. Ok?

I'd like to write some specs for nmatrix too. :)

2012/7/31 John Woods <john.o...@gmail.com>

--
Carlos Agarie
Control engineering student

Polytechnic School
University of São Paulo

[John Woods]

Hey Carlos,

If you could use my repo (mohawkjohn) or Chris' (chriswailes), that would be better. We're in the middle of a major re-factoring.

John