On Dec 19, 2009, at 10:59 AM, David Roe wrote:

> Currently we don't require documentation for __cinit__, __dealloc__  
> and __new__.  Are there any other functions we want to add to that  
> list?  I could see an argument for _add_ and other arithmetic  
> functions too.

A TESTS block is certainly a good thing to have for the arithmetic  
operators.

> David
>
> On Sat, Dec 19, 2009 at 1:41 PM, David Roe <r...@math.harvard.edu>  
> wrote:
>
> > I think that requiring them for every single function is excessive,
> > for example many functions don't take any parameters but self, or
> > don't return output.
>
> OK, good point.  How about requiring INPUT if there are any inputs
> beyond self, and requiring OUTPUT if there are any outputs?
>
> This is what the patch at 7716 currently does (which I've updated to  
> reflect some comments; it still needs a reviewer).  Based on  
> Robert's comments, I would also suggest changing the coverage  
> scripts to not require input or output descriptions for functions  
> that begin and end with underscores.  I'll post a patch on 7716 to  
> that effect shortly.
>
> > Also, what about arithmetic functions like
> > __neg__ and _mul_. (Specifically, I yesterday I was adding a bunch  
> of
> > doctests to monsky-washnitzer, and the thought of adding (in my mind
> > superfluous) INPUT and OUTPUT blocks to these was not encouraging.
> > (I'm not just thinking about file size, developer time and vertical
> > screen real estate are valuable as well.)
>
> Robert, are there any other kinds of functions that you can think of  
> where we don't need descriptions of the input/output?

Not that I can think of at the moment.

- Robert


-- 
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to 
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Reply via email to