* networklevel: 
** quantitative modularity (Marcio Araujo et al. 2008 Ecology: Dieta1): currently only a windows-exe is available!
** WINE-artiger von Marco
Arajo MS, Guimares PR, Svanbck R, Pinheiro A, Guimares P, Reis SF, Bolnick DI (2008) Network analysis reveals contrasting effects of intraspecific competition on individual versus population diets. Ecology 89:1981-1993
(downloadable from https://webspace.utexas.edu/ma4775/software.html)


* include all quantitative host-parasitoid webs

* modularity (Ecology paper by Oleson?); visualise modularity in plotweb


* in second.extinct, use plot.it="higher"/"lower" to plot only one of the trophic levels (wish of Georg Anderson).

* use dirmult (in the same package) to fit Dirichlet-distribution to network data; use these to draw from random distributions in null models?

* check what package tnet does (it somewhere uses the term "weighted" and "two-mode")


* implement the hub-species-detection procedure of Gonzales et al. (Ecol Complex)


In the .Rbuildignore file the syntax is a bit strange:
\..* means (in Perl): starts with . (has to be escaped, hence \.) and is followed by characters ("."), and as many as you want ("*" is a wildcard).
So \..* refers to any file/directory starting with a ".".
To exclude specifically .hg, I guess it would be \..hg or so.

It also doesn't seem to work ...