David Weinberger, over at his Everything is Miscellaneous blog, points out three major problems with Amazon’s Kindle reader for scholarly work:
1) “note-taking and highlighting are jokes”
2) “doesn’t know the original page numbering”
3) “no bibliographical tool”
(From “Kindle is fun but sucks for scholars”)
One of the comments to that post suggests porting to the Kindle a version of Zotero, which could make use of your database portably. Presumably such a database could be loaded via the SD card or, perhaps, it could be shared via Zotero’s new database syncing (currently in preview). I think this is a useful suggestion and could potentially be stunning. However, it will only be stunning if the killer feature of Zotero — i.e. keeping pdfs/docs of items close to their bibliographical entries — is retained. After all, what sets the Kindle apart from any other device is not its organizational structure but its beautiful reading surface. So, combining the killer feature of Zotero with the killer feature of the Kindle sounds to me like a great idea. Of course, one major hurdle is that the Kindle software is proprietary. Unlike converting single books/pdfs into Kindle format, you would want to retain the entire structure and functionality of your Zotero database, which means converting the app. I wonder if a service like Feedbooks would work as a Zotero model: you put a list of hyperlinked out-of-copyright books on your Kindle which download immediately in Kindle format when you click on the link (via 43 Folders). If converted for use with Zotero, you could search your database locally but only call the pdf when needed. To save space the pdf could be set to self-destruct after a certain amount of time, or better, it would somehow ask the user if he/she was ready to destroy the pdf. Of course, your shared database in the cloud would be maintained intact.
To me the iPhone is an ideal device for an implementation of Zotero. The screen on the Kindle is better for reading lots of text, but the iPhone is much the better size for quick searches of your Zotero database while working in the stacks of a research library. Plus, I’m young enough to feel that pulling out my Kindle in public is a lot dorkier than pulling out my iPhone. In terms of content, perhaps a preliminary version wouldn’t necessitate the associated documents, but only the bibliographical entries — maybe adding new entries through photos of the barcode (à la Delicious Library). If it could be implemented, the killer feature would still be the built-in relationships between entries and documents, which is why I use Zotero in the first place. (though pdf support on the iPhone would have to improve) And adding to the database via MobileSafari would be a desideratum as well. Essentially, reproducing all the main features of Zotero on the iPhone are highly desirable to me, especially if my existing database is seamlessly integrated with the iPhone version.
All of this is well and good, but as John Gruber recently noted, developing for the iPhone requires adherence to the terms and conditions of the App Store, which seems to break the GPL. Presumably the Kindle has proprietary agreements for developers as well, and obviously lacks even the semi-open playing field of the App Store. Altogether, this is probably a deal-breaker for any portable Zotero implementation, though I hope I’m wrong. If anyone is working on this and can talk about it, please let me know!
 Feedbooks has an API which has been implemented in an iPhone app called Stanza by the company Lexcycle. I haven’t had a chance to test this app yet but I’ve heard good things about it. This API definitely seems to be a possible starting point for any homebrew iPhone-Zotero implementation, especially if the Zotero development group itself is unable to produce an iPhone app for whatever reason.