Profile

Cover photo
Kannappan Sampath
Attends Indian Statistical Institute
Lives in Bangalore
25,996 views
AboutPostsPhotosVideos

Stream

Kannappan Sampath

Shared publicly  - 
 
Hm, I wonder if we even have a primitive trash collection system here in India. I am embarrassed to say that we care little about what happens to us by way of littering all over our country.
 
Slightly over-engineered trash-bin collection system in #zurich , #switzerland
4
Add a comment...

Kannappan Sampath

Shared publicly  - 
 
 
Development in Sage - My process
================================

Prelude:
--------

In the git development process, I was having frequent (re)compilation of the Sage library. This is not pleasant since the desktop I work on is not that powerful and it takes ages to do any quick development process. I started a thread in sage-devel (see https://groups.google.com/d/topic/sage-devel/WpDnmLZJKDE/discussion ) in order to look for ways to reduce (re)compilation. Thanks to +Nathann Cohen  I have come across some ways to achieve potentially minimum recompilations.

Initial Steps:
--------------

1. Set up Sage by downloading a development version, by having an account on trac.sagemath.org, and by having ssh-keys set up. I won't go into the details here, but I will assume that after you are set up, you are mildly familiar with git, and you are inside SAGE_ROOT. The following command should show "origin" as a remote branch:
$ git branch -r | grep origin
  origin/HEAD -> origin/master
  origin/build_system
  origin/develop
  origin/master
  origin/old_master
  origin/upstream

1. Set up a develop branch that tracks the develop branch of origin
$ git checkout --track origin/develop -b develop

2. Enable "SAGE_INSTALL_CCACHE": In your ~/.bashrc, set the following
export SAGE_INSTALL_CCACHE="yes"

3. Enable simultaneous compilations: Before compiling or building the documentation, set the following MAKE variable in your terminal (assuming you are running Linux)
$ export MAKE="make -j$(grep -c bogomips /proc/cpuinfo)"

4. Upgrade Sage to the latest develop version
$ git checkout develop
$ ./sage -upgrade develop

5. Install git-trac-command. I will assume that ~/bin is in our PATH
$ cd .. # Assuming we were in SAGE_ROOT, I go one directory above SAGE_ROOT
$ git clone https://github.com/sagemath/git-trac-command.git
$ ln -s "$PWD"/git-trac-command/bin/git-trac ~/bin
$ cd -  # We go back to inside SAGE_ROOT
$ git remote add trac ssh://git@trac.sagemath.org/sage.git
$ git trac config --user=Myself --pass=s3kr1t


Development with up-to-date develop branch - existing ticket:
-------------------------------------------------------------

Suppose you want to test an existing branch in a trac ticket #12345. Now, your develop is up-to-date, but the branch in the trac ticket may not be based on the latest develop version of Sage. So, it is better to "pull" the branch in order to check out the changes. Hopefully, the patch does not depend on changes in packages/spkg.

1. Pull the remote trac ticket
$ git checkout develop -b ticket/12345
$ git trac pull 12345
$ git diff develop  # check changes against the latest develop branch

2. Recompile sage and do your tests, etc. It won't take long to recompile.
$ ./sage -b

3. If you are the author of the branch, you can push new changes easily:
$ # edit files and make git commits
$ git trac push 12345

4. (Todo:) How to push changes if you are not the original author


Development with non up-to-date develop branch:
-----------------------------------------------

This is not a recommended method: use at your own discretion

Suppose your develop branch is not up to date, and the ticket you want to check out is on a newer development branch of Sage. In this case, you can still check out the ticket and test the changes it has made. It might be "good enough" for making a basic review of the ticket. However, this requires that you know well enough that the changes in the ticket do not depend on Sage features that have been introduced after your current develop branch. Moreover, a full review of the ticket will probably require you to upgrade to the newest develop code and test that it hasn't broken the existing code.

1. Fetch the branch in the ticket. Suppose the branch is u/username/patch . Note down the SHA hashes of the commits in that branch.
$ git fetch trac u/username/patch

2. Cherry-pick the commits you want. Look at the manual of git cherry-pick to see how to provide ranges of commits.
$ git cherry-pick <hash>

3. Recompile sage, and do tests, etc.
$ ./sage -b

In many cases this method will not work because the changes in the ticket may be extensive and may involve a long trail of patches, some of which are merges of the develop branch on to the ticket. So, you have to be extra careful when doing this. However, if this procedure is successful, then it will save a lot of time that would be otherwise spent in recompiling Sage or upgrading Sage (especially if the changes in the ticket are trivial).


Development with new ticket:
----------------------------

Suppose there is a new ticket; new in the sense that there are no existing patches in the ticket. In this case, we can simply do a checkout of the ticket and let the "git trac" command handle creation of a new branch.

$ git trac checkout 12345   # will create a branch based off our develop
$ # edit edit, and commit
$ git trac push             # push our changes to the ticket


Feedback is welcome on other ways to reduce compilation during development!
1
Add a comment...

Kannappan Sampath

Shared publicly  - 
 
The combinatorics docs are getting a make over...
 
Documentation updated to 6.2. Got word, to especially check out the Combinatorics part: http://sagemath.org/doc/reference/combinat/index.html
1
Add a comment...

Kannappan Sampath

Shared publicly  - 
 
An interesting title with an interesting description!

/p/s/ One for the IMSc (and ISI's too) library... 
 
With school out for the semester and warmer weather upon many of us, perhaps you’re now pondering where to escape to for a summer getaway. Let us offer a suggestion: how about Asymptopia?

“Asymptopia,” says author Joel Spencer, “is a magical mathworld in which parameters approach, but never reach, infinity. Details melt away, the major term remains in blinding simplicity.”

In Spencer’s brand new book of the same name, he presents ideas of how to approach asymptotic problems that arise in discrete mathematics, analysis of algorithms, and number theory. A broad range of topics is covered, including distribution of prime integers, Erdős Magic, random graphs, Ramsey numbers, and asymptotic geometry. This book is aimed at strong undergraduates, though it is also suitable for particularly good high school students or for graduates wanting to learn some basic techniques.

Asymptopia is indeed a beautiful world. Enjoy the journey! http://bit.ly/1nxXLmc
4
Add a comment...

Kannappan Sampath

Shared publicly  - 
4
Brent Longborough's profile photoDavid Roberts's profile photoKannappan Sampath's profile photo
3 comments
 
I can remember several of these: coffee, mushrooms, octopus, onions... and I kicked myself each time I discovered how delicious they were
Add a comment...

Kannappan Sampath

Shared publicly  - 
 
This year's winners: Simon Donaldson, Maxim Kontsevich, Jacob Lurie, Terrence Tao and Richard Taylor.
"Mathematics is essential for driving human progress and innovation in this century," said Facebook CEO Mark Zuckerberg, one of the funders of the awards.
3
Add a comment...

Kannappan Sampath

Shared publicly  - 
 
 
Eigenvalue. Eigenvector. Eigenface. And now we have eigenmoses and eigenjesus.

Your homework: randomly choose a noun, prepend "eigen", and develop the appropriate linear algebra for your new concept.
1
Add a comment...

Kannappan Sampath

Shared publicly  - 
 
Yay! We now have the requisite number of followers! Let's add and vote on questions! Of course, sharing the proposal far and wide is still wise!
2
2
Theron Hitchman's profile photoDavid Roberts's profile photo
Add a comment...
 
Hello Sage Users: 

I proposed a Sage Stack Exchange site for Sage. Please go and follow it; add example questions and vote up (down) questions that you think should be on-topic (resp. off-topic) for the upcoming site. 

http://area51.stackexchange.com/proposals/70511/sage?referrer=8adE6ec0VyqQCyB1kFgF7g2

Looking forward to your support and cooperation for the success of this adventure on Stack Exchange. 

#sagemath  #stackexchange
5
5
Sajin Koroth's profile photoDavid Roberts's profile photoKannappan Sampath's profile photoKamalakshya Mahatab's profile photo
6 comments
 
+Amritanshu Prasad: No. The problem is that ask.sagemath.com is not hosted by StackExchange already. OTOH, MathOverflow was hosted by Stack Exchange but I am not sure how the deal was struck. 
Add a comment...

Kannappan Sampath

Shared publicly  - 
 
 
Quote from a letter of Minkowski to Hilbert (December 1890) indicating that he would not be returning as usual to Königsberg for the vacation period:

I do not know if I need to console you though, since I think that this time you would have found me thoroughly infected with physics. Perhaps I even would have had to pass through a ten day quarantine period before you and Hurwitz would have admitted me again, mathematically pure and unapplied, to your joint walks. .... In order to have points in common with other mortals, I have surrendered myself to magic---that is to say, physics. I have my laboratory periods at the Physics Institute; at home I study Thompson, Helmholtz, and their consorts. And, starting at the end of next week, I will even work several days a week in a blue smock in an institute for the production of physical instruments, thus a practical training as shameful as you can imagine.
3
1
Aniket Joshi's profile photo
Add a comment...

Kannappan Sampath

Shared publicly  - 
 
 
Daniel Quillen, one of my mathematical grandfathers, apparently was a voracious note-taker, and after he passed away a few years ago, his notebooks have been scanned and made public:

http://www.claymath.org/publications/quillen-notebooks

by the Clay Mathematics Institute. This is a boon to mathematicians everywhere.

(Via +Peter Woit , here: http://www.math.columbia.edu/~woit/wordpress/?p=6861)
2
Add a comment...

Kannappan Sampath

Shared publicly  - 
 
 
Vassiliev (of Vassilev invariants fame) has been sentenced by a (kangaroo, if you ask me) court in Moscow for "resisting police". He is asking for maximum publicity about this shameful event...
1
Chandan Dalawat's profile photo
 
AMS President David Vogan has sent a letter protesting the conviction and fine to the Russian science newsletter Troitsky Variant :

http://www.ams.org/news/vogan-vasilliev-letter.pdf
Add a comment...
People
Education
  • Indian Statistical Institute
    Mathematics, 2011 - present
Basic Information
Gender
Male
Story
Tagline
Interested to learn the intricacies of Mathematics and enjoy its Beauty
Places
Map of the places this user has livedMap of the places this user has livedMap of the places this user has lived
Currently
Bangalore
Previously
Trichy - Chennai
Links