[yt-dev] Solicitation for feedback on software

[Matthew Turk]

Hi John,

On Tue, Aug 5, 2014 at 12:21 AM, John ZuHone <jzuhone at gmail.com> wrote:
> Hi all,
>
> Congrats to everyone on a successful release of yt 3.0!
>
> I wanted to drop the developers a line and get some opinions and feedback on a software package I'd like to release in the next week or so. It's called "jt".

I like that the name could mean not just "Julia" but also "John".

>
> jt is a wrapper for yt for the Julia programming language. Julia is a new(-ish) high-level, high-performance dynamic programming language for technical computing. If you want to learn more about Julia, visit http://julialang.org. One of the most important things to note about Julia is that it has a "just-in-time" compiler, so it can be very fast (within a factor of few of C speed in some cases). There is also a lot of interoperability between Julia and Python, including wrappers for Matplotlib and Sympy, and even an extension of the IPython (soon to be Jupyter) notebook!
>
> A short summary of what jt exposes from yt in Julia:
>
> * Datasets
> * Some data objects (e.g., spheres, rectangular regions, slices, projections, profiles, etc.)
> * YTArrays and YTQuantities
> * Simple visualization tools (e.g., SlicePlot, ProjectionPlot, FixedResolutionBuffer)
>
> Here's why I'm mailing the dev list about this. I want jt to be an expansion of yt's capabilities into another programming language, particularly a new one like Julia which is proving popular for scientific computing. I also, however, want to be sensitive to concerns that jt could detract from yt, or become something else that we need to support that we don't really have the time or resources for. So, let me be clear about a couple of things:
>
> 1. No one should be expected to support this software but me, unless any of you want to join in. I do not want questions about it to be handled on the yt-users list, for example. I haven't decided how questions about it should be handled, except that for now they should just be directed to me. I doubt there will ever be enough people using it that there would be a need for it to have its own list.
>
> 2. I do not want jt to be a distraction from yt, since:
>
> a) it requires yt to work
> b) it exposes a small subset of everything yt does
>
> Most of my work is still going to be in Python/yt, but I thought this would be a fun project, useful for myself and others, and a good way to learn a new language. I wanted to strike a balance between making jt a part of yt itself (which would make Julia a dependency of yt, soft or hard, which we do not want), and designing a whole new package in Julia from scratch to do exactly what yt does, which seems like a pointless reinventing of the wheel.
>
> These are probably obvious points, but I wanted to make sure I got them across. I wanted you all to have a chance to have a look at my code and documentation if you wanted (links below), and then give me some feedback. In particular, I'd like to hear from anyone that has any of the concerns I mentioned above (or others).
>

I think this is pretty cool.  And, I think it's going to be a good
opportunity to step back and look at which operations in yt are
primitive operations, which can be codified, what the rough protocol
is, and then perhaps expose them a different way too.  So, nice work!

> I'd like to make a release in about a week, so I'll wait for about that long on feedback.
>
> Thanks all!
>
> John Z
>
> Some helpful links:
>
> jt source: http://github.com/jzuhone/jt
>
> jt Documentation: http://www.jzuhone.com/jt
>
>
>
> _______________________________________________
> yt-dev mailing list
> yt-dev at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org