Sage GSOC 2016

240 views
Skip to first unread message

Samuel Lelièvre

unread,
Jan 18, 2016, 7:21:49 AM1/18/16
to Sage-devel
Hi all,

Sage will be applying to Google Summer of Code (GSOC) again this year.

Information for applicants is on this wiki page.

http://wiki.sagemath.org/GSoC/2016

Anyone with ideas for projects, please add them there.

See last year's Sage GSOC wiki page

http://wiki.sagemath.org/GSoC/2016

for inspiration and formatting (copy-paste and adapt from there).

Samuel

Samuel Lelièvre

unread,
Jan 18, 2016, 7:38:46 AM1/18/16
to Sage-devel
Mentors should also consider subscribing to sage-gsoc

https://groups.google.com/forum/#!forum/sage-gsoc

where applicants will introduce themselves.

Jeroen Demeyer

unread,
Jan 18, 2016, 8:25:05 AM1/18/16
to sage-...@googlegroups.com
For GSoC, we should really insist that the coders try to get their code
on trac from the start. In particular, they should follow the "one
ticket, one issue" rule and not just dump everything on one ticket.

I have seen some GSoC projects where one person codes for the summer,
puts the code on Trac and then disappears. The code will never get
merged because it has issues or because it is simply too big to be
reviewed in a reasonable way.

As an example of how to do this right, I recommend:
http://trac.sagemath.org/ticket/17601

Jeroen.

kcrisman

unread,
Jan 18, 2016, 9:07:04 AM1/18/16
to sage-devel


See last year's Sage GSOC wiki page

    http://wiki.sagemath.org/GSoC/2016


For the record (especially newbies), that's http://wiki.sagemath.org/GSoC/2015
 
for inspiration and formatting (copy-paste and adapt from there).


But not *too* much copy-paste ;-)

Samuel Lelievre

unread,
Jan 18, 2016, 10:00:54 AM1/18/16
to sage-devel


Mon, 2016-01-18 06:07:04 -0800 (PST), kcrisman:

Actually archives of several previous editions are linked to from

    http://wiki.sagemath.org/GSoC/

The copy-paste advice is mostly for formatting style for project proposals,
and for renewing project proposals which are still of interest but were not
funded in previous years.

Johan S. R. Nielsen

unread,
Jan 19, 2016, 10:15:43 AM1/19/16
to sage-...@googlegroups.com
How does it work with Sage's application to GSoC and what are the
relevant deadlines? Also, in what period should the project be conducted?

I have an idea for a project I would like to mentor, but I have to be
sure that I'm available in the period.

Best,
Johan
--

Harald Schilly

unread,
Jan 20, 2016, 11:49:30 AM1/20/16
to sage-devel


On Tuesday, January 19, 2016 at 4:15:43 PM UTC+1, Johan S. R. Nielsen wrote:
How does it work with Sage's application to GSoC and what are the
relevant deadlines?

Hi, the program is explained on the GSoC website, and the timeline is here:

The start/midterm/end times are given by that timetable.

If you have an idea, please add it to SageMath's wiki page, for the year 2016. 

-- harald
 

Karan Desai

unread,
Feb 1, 2016, 12:12:16 PM2/1/16
to sage-devel
How about adding support for machine learning algorithms in sage ?
Hello developers,
I am just pitching in a very rough and amateur thought for the idea of 
upcoming GSoC, and we can refine it further due to your inputs.

Sage can have a support for a variety of Machine Learning algorithms.
We can wrap around python's scikit-learn library, it has been great
since a past few months.

Plus a whole new domain would be a huge addition to the things which
Sage can provide.

A good amount of Object Oriented Programming in Python is required
for that. I am not sure much yet, as I am a newbie, but I say if the veterans
of Sage in this mailing list have a look at this idea, I can take it up myself
for this year's GSoC.

Nicolas M. Thiery

unread,
Feb 17, 2016, 2:01:32 AM2/17/16
to sage-...@googlegroups.com
On Mon, Feb 01, 2016 at 09:12:15AM -0800, Karan Desai wrote:
> How about adding support for machine learning algorithms in sage ?
> Hello developers,
> I am just pitching in a very rough and amateur thought for the idea of
> upcoming GSoC, and we can refine it further due to your inputs.
> Sage can have a support for a variety of Machine Learning algorithms.
> We can wrap around python's scikit-learn library, it has been great
> since a past few months.

scikit-learn is indeed great, and it is trivial to install (with pip
IIRC) and use from Sage. What specific development would you have in
mind? Did you spot compatibility issues or additional glue that
would make the combination more useful than it is now?

Cheers,
Nicolas
--
Nicolas M. Thiéry "Isil" <nth...@users.sf.net>
http://Nicolas.Thiery.name/
Reply all
Reply to author
Forward
0 new messages