Re: [boost] [gsoc18] New Astronomy Library

Subject: Re: [boost] [gsoc18] New Astronomy Library
From: Pranam Lashkari (plashkari628_at_[hidden])
Date: 2018-01-25 12:49:18

Dear Jean-Claude;

As you mentioned in your emails, you are having some ideas what can be
implemented in this library as per requirement of professionals in this
field.
Can you start discussing those ideas so that we can get a better idea of
how to structure this library?

On Mon, Jan 22, 2018 at 1:40 AM, JC Passy <jcpassy_at_[hidden]> wrote:

> Dear Pranam, dear Vinicius,
>
> not sure that my initial email made it to the mailing list. If it did,
> sorry for repeating myself.
> I am working as a computer scientist at the MPI for Intelligent Systems in
> Tübingen. Before that, I did my PhD in astrophysics and a postdoc. In
> general, I am also interested in using boost and developing a library that
> could be used by astronomers and astrophysicists.
>
> Looking at your proposal, I would second Vinicius’ opinion in the sense
> that one should try to make it a bit more specific. In particular, you
> should emphasize the *why* (what do you wish to accomplish? why would
> someone need/use it?) and the *who* (who is your target audience?).
> Regarding the latter, I see two options:
>
> 1) *amateur astronomers*: I personally do not know what their needs are,
> but I could try to find out. There are some open source projects dedicated
> to amateur astronomers (for instance https://openphdguiding.org to which
> we contributed), so that could be an option;
>
> 2) *professional astronomers*: I think a lot can be done, in particular
> regarding observational data analysis. They used *REALLY* old tools (IDL,
> IRAF, …) and given the future and upcoming surveys that will produce very
> large amount of data, they might have serious performance issues. This is
> why I believe they might benefit from using cutting-edge C++ libraries.
>
> I would be quite interested in option 2, and it could impact potentially a
> large audience IMHO. Of course, the “selling” part won’t be easy (people,
> especially astrophysicists :) won’t give up their old methods easily even
> though the new method is 10x faster) but we can facilitate this process by
> pin-pointing features that they really need. I will try to talk to some of
> my former colleagues this week and see what they think about that. In this
> case, this is a project that would need a bit of time to be properly
> defined and the deadline for GSoC18 is coming too soon. As far as I am
> concerned, this is no problem as this project does not need to be tied to
> it.
>
> Cheers,
>
> Jean-Claude
>
>
> On 21. Jan 2018, at 15:00, Vinícius dos Santos Oliveira via Boost <
> boost_at_[hidden]> wrote:
>
> 2018-01-21 8:07 GMT-03:00 Pranam Lashkari <plashkari628_at_[hidden]>:
>
> I have made this proposal for this new library(check attachment)
> Suggestions are invited.
>
>
> You need to focus more on "selling your project". Why is it important to
> have such library in C++ specifically? Isn't this one of those projects
> where other languages will do just fine? What market are we missing by not
> having this library? What C++ have to offer to make such library be
> inviting to write in C++?
>
> You mention "gather information from observation data". Does this mean
> you'll include a parser for some format?
>
> Can you compare this library with some non-C++ library and detail a little
> how will scope between the two differ (e.g. your library will focus
> initially only on X given the library is only starting)?
>
> For now, this should be good. When it's time to submit the student form
> through the GSoC platform, you'll also focus on selling yourself. I'll give
> more feedback on this when the time comes. Don't worry.
>
> Should I publish this on the wiki page of Boost GSoC 2018?
>
>
>
> Yes, please. You can put my name as a potential mentor.
>
> --
> Vinícius dos Santos Oliveira
> https://vinipsmaker.github.io/
>
> _______________________________________________
> Unsubscribe & other changes: http://lists.boost.org/
> mailman/listinfo.cgi/boost
>
>
>