Weekly Edition
Return to the Development page
| |||||||||||||||||||
Matplotlib 1.2.0 released
After months of hard work by a veritable army of contributors, I'm pleased to announce the release of matplotlib 1.2.0. This is the first time we've released without the assistance of John Hunter, who is sorely missed. I hope this is at least a small way to say thanks for all of his great work. Release tarballs and binaries are available on github. (They are no longer being made available on SourceForge). https://github.com/matplotlib/matplotlib/downloads This is the first release to support Python 3.x (and as a result drops support for Pythons earlier than 2.6). There is new support for outputting PGF/TikZ files. New plot types include 3D trisurface plots, and streamplots. Tripcolor, boxplot, colorbars and contour plots have all grown new features. And under the hood, numerous improvements in stability, flexibility and robustness. For a complete list, see the "what's new" page: http://matplotlib.org/users/whats_new.html For an even more detailed list of 698 issues (!) resolved since the last release, see the github statistics page: http://matplotlib.org/users/github_stats.html Enjoy! As always, there are number of good ways to get help with matplotlib listed on the homepage at http://matplotlib.org/ and I thank everyone for their continued support of this project. Mike Droettboom -- http://mail.python.org/mailman/listinfo/python-announce-list Support the Python Software Foundation: http://www.python.org/psf/donations/ (Log in to post comments)
Matplotlib 1.2.0 released Posted Nov 16, 2012 13:41 UTC (Fri) by debacle (subscriber, #7114) [Link]
Very nice!
I wonder, whether there are news about:
- threadsafety (AFAIK, matplotlib used to have problems when used from multiple threads)
- HTML5 backend (there was some work, but I'm not sure whether it ever reached "production quality")
Matplotlib 1.2.0 released Posted Nov 23, 2012 13:20 UTC (Fri) by njs (guest, #40338) [Link]
My understanding is that the original HTML5 backend turns out not to be maintainable. (It was basically based on porting matplotlib's internals to javascript, a classic boil-the-ocean sort of project.) The new idea is to do all the rendering on the server, and send bitmaps to the browser as needed for interactivity, like VNC. This seems much more plausible to me: http://mdboom.github.com/blog/2012/10/11/matplotlib-in-th...
Matplotlib 1.2.0 released Posted Nov 23, 2012 16:34 UTC (Fri) by debacle (subscriber, #7114) [Link]
Interesting approach. Let's see where it gets.
Michael is probably right, that moving huge amounts of data to the browsers JavaScript interpreter can be a problem. OTOH, letting the browser do the work scales nicely with the number of users, as the server does only have to provide more data, but not do more rendering work. Also, some interactive actions, like switching the visibility or colour of plot lines is sth. browsers can do nicely without any interaction with the server.
That's why we think about re-implementing parts of our matplotlib rendering in JavaScript, e.g. using D3 (https://github.com/mbostock/d3/wiki/Gallery), except for printing.
|
|||||||||||||||||||