On an experimental basis, I'm working with a visiting student to distribute born-digital html-based articles as "stand alone" files that include all assets and are easily downloaded. See:
https://isawnyu.github.io/isaw-papers-xhtml-standalone/
They key is that images are included as base64-encoded data urls. The strategic goal is to explore #ScholarlyHTML as an open and easily parseable equiv to PDF. More work to do, but I'm pleased with her progress.
@sebhth It sounds interesting, but why choose it over EPUB? I'm guessing it's because Web browers are already set up for HTML, but to compete with PDF a format optimally needs to support pagination for citation purposes. I looked at some of the examples via the link and they seemed unpaginated.
One quick answer:
Here are the ePub+Mobi versions of same content:
http://isawnyu.github.io/isaw-papers-ereaders
Longer:
The "standalone" files have RDFa + other metadata. They remain as parseable as any html file. More amenable to various forms of computational analysis.
Longer, longer:
In general, ISAW Papers (the journal) avoid pagination. Each paragraph is directly citable as a more fine grained solution. E.,g.,
http://dlib.nyu.edu/awdl/isaw/isaw-papers/13/#p34
Many thanks for responding/asking!
Yeah, pagination is problematic in formats that reflow in viewers. I agree that we need to move toward citation mechanisms that correspond to semantic structures inherent in the work, rather than arbitrary chunking driven by a particular display medium, and can be encoded across all formats used.