Re: [boost] [GSoC18] Astronomy Project Proposal

Subject: Re: [boost] [GSoC18] Astronomy Project Proposal
From: Pranam Lashkari (plashkari628_at_[hidden])
Date: 2018-03-19 18:43:43

>
> Pranam, I've read your proposal and I think it's lacking a motivation
> section.
>
I'll add motivation section and re-post it to the mailing list.

  Let's suppose I'm comfortable with C++. From the topics you've chosen it
> seems that your library is an astrometry library rather than an astronomy
> one. This is _very_ specific, I would even say too specific for Boost, but
> suppose for now that I am interested in these topics.

I am proposing new library here so It is not possible to implement all the
functionality(e.g: spectroscopy) in the period GSoC, in the proposal I have
written only things which I am gonna implement in GSoC. I am planning to
complete this library by this November and reach to review ready state.

You plan to plot curves. My question is then how are you going to implement
> that? There is currently no GUI library in Boost (there was a discussion
> not too long ago, but AFAIK it resulted in "nothing"). Are you going to
> depend on external libraries? I'm afraid a pretty big part of Boost
> community will be against that.
>
I am planning to implement graphs with SVG files.  SVGs can store simple
elements like boxes and ellipses and also a little more elaborate ones like
Bézier curves, filled segments and some more. It should be enough for all
graphs and for images Boost.GIL can be used.

Moving on to FITS... there are a few libraries out there which provide the
> functionality you describe, so what are the advantages of using your
> implementation?
>
In astronomy, all the things are dependent on the observation and
observation data is mostly stored in the FITS. As you said earlier external
dependency is not a good idea. So if boost is planning to feature astronomy
library in its collection than there are no other options then implementing
boost's own FITS handling module. In addition, existing FITS library lacks
good documentation also where boost can be helpful to the science community
as always.

And finally, coordinate systems. Do you really need a new library for that?
> I mean we have Boost.Geometry, why can't celestial coordinate system just
> be added to it?
>
I think Boost.Geometry is more of a generic library and implementing all
these specific coordinate systems in it won't be a good idea. But I was
planning to implement these coordinate systems by extending Boost.Geometry
in this library.

Don't get me wrong though, I'm not trying to discourage you. I'm just a
> concerned physicist :)
>
Feedbacks are always welcome they make us better. If you have any other
ideas what else can be implemented in this, I'll be glad to give it a try.
I hope I have given satisfying answers to all your questions.

Thank you,
Pranam Lashkari