(1) Peter identifies Wikipedia as a good example of new models for publication. Open source software, considered as a publication / collaboration process, also suggests some really interesting new ideas. A recent development is distributed version control systems such as git and Mercurial (Linux is now using git, which was developed by Linus Torvalds) which are potentially far more useful for large-scale collaboration than old-fashioned version control systems like those used in Wikipedia.

It’d be <em>very</em> interesting to mashup a wiki like MediaWiki (or something more lightweight) with something like git. Probably only take a few days to get something workable if you picked the right projects to mash.

(2) <a href=”http://scholarpedia.org” rel=”nofollow”>Scholarpedia</a> is an interesting example of a new tool that is getting input from some very eminent scientists, along the lines advocated by Peter in his last paragraph.

Presently you are right, it’s difficult to write papers in this way. I think the main reason for this is that the required software tools are not in place. Before the Wikipedia was in place one could have made the same criticism of highly modularized, hyperlinked articles. But once the Wiki software was in place and tailored for this particular application it became quite straightforward to do this. I think the same ought to apply to the suggestion I made.

Why don’t you try writing a paper in this way? (Or rewriting a paper you already have?) Everything you need is out there. And it would make your suggestions much more concrete.

1) Contrast. Reading papers on, well, paper, is easy on the eyes.

2) Annotate-ability. Writing notes wherever I want on paper is easy. Doing the same on PDFs is hard, especially if I want to write equations.

3) Works on the plane. Paper doesn’t require an internet connection to be readable.

1 and 2 will eventually change as tablet PC technology improves, but for now, they’re big stumbling blocks. The best compromise I can think of is to allow people to “customize” their version of the paper by collapsing explanations they don’t need and expanding details they want to read. Once the paper is electronically customized, it can be printed if further reading is desired. Of course, I would only need this for the (small) subset of papers that I actually read in detail, but those papers are the ones I care about most. I have one or two papers I’ve actually had to re-print because the original copies got so worn and covered with scribbled notes that they became hard to read.