Generate range graphs

[steveno]

No description provided.

[steveno]

I'm thinking this will probably involve Clutter or Cairo. I'm not 100% sure yet at this point how to implement these.

[steveno]

As part of Issue #12 I'm implementing a vapi for Gnuplot

http://www.gnuplot.info/

Which as of this posting is breaking CI.

[steveno]

I'm also considering using flotr2 though. It's Javascript based. It would appear that Vala easily integrates with webkit and this would require nothing fancy.

Flotr2
http://humblesoftware.com/flotr2/index

Blog post about Flotr2
http://swizec.com/blog/flotr2-my-favorite-javascript-graph-library/swizec/4558

Webkit + Vala demo
https://wiki.gnome.org/Projects/Vala/WebKitSample

[dylanbeaudette]

Anyone interested in composing graphics in something like R? It might also be a fine interface for running simulations, linking to external databases, etc.. Happy to help.

[steveno]

I would be happy to accept contributions with graphs generated in R.

[dylanbeaudette]

Great! There are a couple of options:

• R interface to the libbalistica library
• R interface to a command-line version of balistica
• loose coupling via intermediate files (CSV, JSON, etc.)

The best approach is a self-contained R package that includes the sources for libbalistica. This is pretty common for libraries written in C or C++. I have no idea if it is even possible with vala.

Also, an interface to R would make it possible to plot simulated trajectories in 2D or 3D via all of the GIS functionality.

[steveno]

If you're waiting on me to tell you which of the three I'd pick or prefer I'm sorry. I'm not familiar enough with R to make an informed decision there.

I will tell you however that I've considered dropping vala and rewriting everything in a more popular language like D or rust. I've also considered going crazy and using zig.

Your thoughts are welcome.

[dylanbeaudette]

I'd vote for rust because there is a relatively simple framework for interfacing with R. I'd be happy to put in the time required to get the R package developed, documented, and published on CRAN.

[steveno]

I haven't committed to either yet, but D doesn't look that difficult either:

https://dlang.org/blog/2020/01/27/d-for-data-science-calling-r-from-d/

[dylanbeaudette]

That might be a fine way to get R-generatd figures / simulations into a GUI-driven application, but would not work for an R package (the inverse). Interesting read, I had never heard of D.