matplotlib-users Mailing List for matplotlib

On Sun, Aug 12, 2012 at 01:23:35PM -0700, jonasr wrote:
> Hello, 
> 
> i am working on some 3d stuff with plot_surface() , my problem is that  i
> want to use a stride smaller then 1.

The stride refers to the *array* stride. So a stride of < 1 makes no
sense.

> Since my data is only on an intervall from -1 to 1, in x and in y direction
> i want to plot a 3d grid with at least 20 lines in each direction, is there
> a possibility to do this ?

An rstride of 1 will plot every row. A cstride of 3 will plot every 3rd
column. If your data is in a 2D array of dimensions 100x100, say, then
setting rstride=5 and cstride=5 will plot every 5th row and every 5th
column, giving 20 lines in each direction. The kwargs rstride and
cstride do not care about the domain of your data.

Hope this helps.

-- 
Damon McDougall
http://www.damon-is-a-geek.com
B2.39
Mathematics Institute
University of Warwick
Coventry
West Midlands
CV4 7AL
United Kingdom

Hi Ben,

On Sat, Aug 11, 2012 at 8:30 AM, Benjamin Root <ben...@ou...> wrote:
> I have said this before, and it can't be repeated often enough.  The work
> that you and your team has been doing the past few years with the notebook
> is *already* revolutionizing how we teach python.  10 years from now,
> programmers will point to this as the *killer* feature of python.

well, your kind words are very much appreciated, truly.  It's been a
ton of work, and at this point far more credit goes to the rest of the
team than to me.

One thing I'd like to emphasize is how strong, productive and positive
the collaboration between IPython and matplotlib has been over time:
we have managed to allow both projects to fully retain their identity
(we don't even have a hard dependency on mpl in IPython, and
matplotlib doesn't even import IPython at all), and yet the two
projects complement each other very well, benefiting both of them, and
ultimately all of our users.  A good combination of communication and
collaboration has allowed us to maintain a strong separation of
concerns while providing users a feel of integrated functionality
where it matters.

I have every reason to believe that, as we push into the second decade
of this effort with the vision of challenges and ideas that John and
Michael D. recently laid out (at the SciPy'12 keynote and in Michael's
posts), this is only going to get better.  The web work is going to be
a pretty tough challenge, but at the same time it's a great
opportunity to revisit key parts of matplotlib with a lot of hindsight
we've accumulated.

That kind of hindsight is what let us refactor all of IPython over the
last few years, so that while the user experience at the terminal from
0.10 to 0.11 remained mostly unchanged (we did have some regressions
but they were pretty mild), we had a completely new architecture under
the hood that paved the way for the qt console, the notebook and the
current parallel machinery.  I hope we'll see similar benefits as the
web forces us to rethink matplotlib for a multiprocess model.

Cheers,

f