[Haskell-cafe] Open-source projects for beginning Haskell students?

[Brent Yorgey]

On Mon, Mar 11, 2013 at 11:50:38AM -0700, Ben wrote:
> 
> On Mar 11, 2013, at 11:26 AM, Jason Dagit wrote:
> 
> > Myself and several of my friends would find it useful to have a plotting library that we can use from ghci to quickly/easily visualize data. Especially if that data is part of a simulation we are toying with. Therefore, this proposal is for: A gnuplot-, matlab- or plotinum-like plotting API (that uses diagrams as the backend?). The things to emphasize:
> >   * Easy to install: No gtk2hs requirement. Preferably just pure haskell code and similar for any dependencies. Must be cross platform.
> >   * Frontend: graphs should be easy to construct; customizability is not as important
> >   * Backend: options for generating static images are nice, but for the use case we have in mind also being able to render in a window from ghci is very valuable. (this could imply something as purely rendering to JuicyPixels and I could write the rendering code)
> >  
> > * What I would hope from you is a willingness to exchange email and/or
> >   chat with the student(s) over the course of the project, to give
> >   them a bit of guidance/mentoring.  I am certainly willing to help on
> >   that front, but of course I probably don't know much about your
> >   particular project.
> > 
> > I am willing/able to take on the mentoring aspect :)
> 
> i second this, but with a different emphasis.  i would like a ggplot2-type DSL for generating graphs, for data analysis and exploration.  i agree with :
> 
> * it would be great to have no gtk2hs / cairo requirement.  (i guess this means text rendering in the diagrams-svg backend needs to be solved.)  i guess in the near-term, this is less important to me -- having a proper plotting DSL at all is an important start.
> 
> * frontend : graphs should be easy to construct, but having some flexibility is important.  the application here is being able to explore statistical data, with slicing, grouping, highlighting, faceting, etc.
> 
> * backend : static images are enough for me, interactive is a plus.  most importantly : it should be fast enough to work pleasantly with large datasets.  ggplot2 is pretty awesome but kills my machine, routinely.

Not to throw cold water on these ideas (which sound fantastic!), but
the scope of this sounds more like a GSoC project than something a
beginner could accomplish in 10-15 hours in the space of a few weeks.
I'm looking not for project ideas but for small, concrete
contributions they could make to existing open source projects.

-Brent