Firefox is developing some interesting offline capabilities for the next version. The idea is that you can go offline and continue working in web-based applications as if you were still connected, which needs all the scripts and data to be temporarily stored locally. According to posts last month from Chris Double and Robert O'Callahan, this involves marking necessary resources with <link rel="offline-resource"> (rather than using rel="enclosure", which is fair enough, but I don't know why they're using 'link' and not 'a') and bundling everything up into a jar (zip) archive.
This is how I'd like to see papers published for on- and offline use from now on, rather than PDF:
- an XHTML document with CSS stylesheets for screen, handheld and print
- all the images, data, 'accessed on' caches of web pages and supplementary data marked as 'offline-resource'
- an XML file containing the bibliographic metadata as MODS or RDF
Firefox would bundle all this up into an archive, which could be opened on any device that supported the archive format, dragged into a word processor or bibliographic manager to cite the paper, printed out, or annotated using javascript - all in a self-contained archive.
See also: Zotero and compound documents.
Comments
All fields are optional, email address will not be shown; no HTML, URLs are automatically hyperlinked.
One point for using 'link' and not 'a' is there is a need to be able to reference all items to be stored in the offline cache when the user visits the first page of the application.
Using 'a' would mean that there would need to be links to these resources on that page somewhere. 'link' on the other hand does not have a visual presence on the page.
The Jar file thing is optional too. You can list each item in individual 'link' elements but it makes it easier to bundle them all in a single file and just reference that. Since the Jar protocol is already supported by Firefox it made sense to use that for the demos.
Chris: That makes sense, though I still think allowing rel="offline-resource" on 'a' tags (as well as 'link', for invisible resources) would be better than having duplicate links in 'head' and 'body'.
Reading your last sentence it sounds like Firefox doesn't actually bundle all the files up into a jar archive - perhaps it just stores all the offline files in a folder somewhere.