I'm moving the discussion of the paper to the mailing list, as requested by Stefan. Here's a summary of the points so far:

* Josh brought up a Computational Science and Discovery special issue as not only a possible venue for the paper but a general renewed call to action about this paper.

* I vetoed the CSD special issue, despite having no formal power of veto =P, and suggested we opt instead for a fully open access journal, such as F1000Research.

* This and Johannes's email sparked a discussion about possible homes for it. Current suggestions:

   - F1000 Research

   - JMLR MLOSS [sklearn published here]

   - Journal of Open Research Software [mahotas published here]

   - Image Processing On Line (ipol.im)

   - Journal on Image and Video Processing (jivp.eurasipjournals.com)

* Stefan suggested as the author list, "currently active core contributors", but would certainly add more authors that have "made a substantial contribution to the package." I feel the same way and I imagine other current core devs would not object to this either. (?)

* There is a markdown template PR here:

Johannes gave a +1 to merging that, and I give another, so that makes +2, I'll merge following this email. =)

(I have a question about this point: for managing LaTeX papers on git, I have usually stuck to the convention of "1 sentence: 1 line". Do we want this for this paper, or wrap at 72 characters, or something else?)

I think that's everything, though I'm sure the discussion will continue!

Stefan asked me to elaborate on my suggestion of F1000. I must admit I don't know much about the other journals on the list, and would need to look into them. Things that I expect from our eventual home are:

 - open access.

 - CC or similar licensing that allows text mining applications.

Further niceties offered by F1000Res:

 - papers published immediately as preprint.

 - open peer review

 - once two reviewers have signed off on the paper, it is considered "peer reviewed". Reviewers can request modifications, and full paper and revision history is maintained.

 - peer reviewed articles are indexed by PubMed.

Essentially, the review model is quite similar to the GitHub PR process, which sounds great to me. PeerJ offers a similar (identical?) model, but is currently not LaTeX friendly, which pretty much rules it out for this.

Juan.