Andrei Berceanu

Perhaps `docker` is the way to go, I see it is also used in `VScode` LaTeX [extension](https://github.com/James-Yu/LaTeX-Workshop/issues/302)

One can also consult eg. [arXiv](https://arxiv.org/help/submit_tex) and [PRL](https://journals.aps.org/prl/authors) guides for LaTeX submission.

So are you suggesting to implement a stand-along dense output routine, which could then be called by the current `oderkf` and the future `DOPRI5`? Or will it be part of...

@acroy I guess that by "something that can be used with `oderkf`" you mean changing `oderkf` itself, since the dense output would need partial results from the runge kutta integration...

I wrote some notes on implementing dense output here: http://nbviewer.ipython.org/github/berceanu/notebooks/blob/master/julia/dense_output.ipynb Unfortunately as there is no support for MathJax in Github (at least as far as I know), I could not...

For the API details on dense output, see #44

`dopri5` is nothing but our beloved `oderkf` using the Dormand-Prince set of coefficients, and with dense output integrated. i see no convincing reason to port it at this point. about...

Ok, so what I had in mind was an API like matlab's, see in particular the `refine` keywork: http://stackoverflow.com/questions/20898040/using-refine-option-in-matlabs-ode45 That would mean always calculating the interpolation, with an extra cost...

I gave your suggestion a bit more thought. Let's say we want the solution for any time `tau` in the interval `tspan`. We define ``` type MyDense t::FloatingPoint h::FloatingPoint x::typeof(x0)...

Sorry for the overlap between our last 2 comments, I posted mine before I saw yours. I guess they converge somehow?