[AstroPy] Introducing Myself, GSoC '14

[Tanuj Rastogi]

Hi,
I am a 3rd year CS student from Thapar University, India.
I find your astropy project very interesting and have gone through the GSoC
'14 ideas list.

One of the ideas that I find interesting is:

*Implementing uncertainties in Quantity*

If possible, I would like to apply with a proposal for the same.
I have forked the astropy repository and am exploring the codebase and the
documentation related to
this idea (mainly *nddata* and *units* modules).

*This is what I have done so far:*
I have gone through the code in *astropy.units *module and
*astropy.nddata *module
and have gained some insight
into its working.
I have specifically understood the* NDUncertainty* and
*StdDevUncertainty*classes along with their usage in *NDData
*class
and the general structure of *NDData* class.
------
>From this preliminary study, my basic understanding of what I would have to
do is that:

1. Implement the uncertainty attribute in the *Quantity* class ( as an
object of *StdDevUncertainty* ).
2. Modify all operations on the *Quantity* class to ensure proper
propagation of uncertainty.
3. Possibly, work on expanding the *Uncertainty* class, for eg., writing
support for uncertainty with correlated variables.
------
Of course, I have not fully explored the *Quantity* class (or the
*astropy.units* module in general) and there might be some things
that I might be missing or getting wrong.
Plus, I am not sure about the importance(or urgency) of this project from
the organization's perspective and am open to ponder and work on other
ideas if/as
suggested by you.

Having said this, I would appreciate if you could guide me on how to
approach this idea further and building a proposal for the same.
I do have some specific questions:
1. What would be your advice on approaching this idea further?
    Should I:
    * Try to fix basic bugs listed in the github isuues page.
    * Spend time on understanding the codebase further to get a clearer
picture of what I have to do.
    * Focus on brushing up and reading some statistics and math related to
uncertainty, to form a more concrete and clear idea of how the uncertainty
      would propagate and how to implement it.
    * Any other general advice you would like to give.
2. What should be the scope of the proposal that I should aim for. For eg.,
should it include expanding or modifying the present Uncertainty framework.
------
I have a good background in Python. In my 3rd semester I did an online
course from Massachusetts Institute of Technology (through edx platform) on
'Introduction to Computer Science' which thought Python along with many CS
concepts. I also earned a certificate for completing it. I have used Python
extensively in a lot of coding contests.
In addition to this, I did a course on 'Introduction to Astronomy' from
Duke University (from Coursera).
I would be happy to answer any questions you might have.

Regards,
Tanuj Rastogi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/astropy/attachments/20140301/9fd08bb0/attachment.html>