BCRARL

Is this progressing? My version of ForwardDiff does not implement these macros and Icoudl really use it :-)

Yeah, it requires Optim for the optimization. Kamada Kawai solves a non-linear least squares problem. $$ \min_{x_i} \sum_{i,j} (\sqrt{\lVert x_i-x_j\rVert^2} - r_{ij})^2, $$ where $r_{ij}$ is the shortest path along...

You can just use [my fork](https://github.com/BCRARL/GraphPlot.jl/tree/KamadaKawai), @atthom . I haven't kept up to date with the master branch, so some minor manual merging might be necessary. If you need me...

I just merged all upstream changes, but I have not run any tests to see whether anything has broken in the meantime. If you run into an issue open it...

I incorporated the comments as seemed appropriate. Optim is part of JuliaOpt. Putting in an independent non-linear optimization implementation is not a good use of time and resources. I haven't...

@simonschoelly I've addressed most of your concerns, I belive. Is there anything in particular holding back the inclusion?

Try ``` m = BSON.load("Processmodel2.bson",@__MODULE__)[:m] ```

The Symbol error is easy to fix. In AST.jl there are a few references to deprecated symbol instead of Symbol. But after that, you will find that SymbolNode is equally...

The culprit seems to be a call to `next(::Max{typeof{first}}, ::PrivateState, input::PrivateState)` as generated by Zip. input needs to be unwrapped for `first` to apply, although I'm not sure whether `first(unwrap(rf,...