[Elfwreck]

Originally Posted by bigtext

As a programmer, why would you write instructions that says "IF sideloaded content THEN don't store metadata into database that syncs to cloud ELSE store metadata into database that syncs to cloud".

No, I suspect it's "IF content is not in Amazon's database THEN ignore." I don't think sideloaded content is viewable online nor synced across other devices on the same account; it's only present on the machine it was loaded to. (Please, someone correct me if I'm mistaken. How do Kindles deal with personal documents, either those mailed to its service or those loaded through USB?)

No, its easier to write a program that just stores everything in the device database regardless of how the book was loaded.

Store in device database: yes.
Upload and sync across other devices: no; Amazon is not paying to upload your personal files across their servers.

If side loaded content displayed in a different library view and reader than items purchased from Amazon then there would be a different story.

I believe they are (or can be) stored in different places--a "documents" folder instead of the "books" folder.

And they don't need to upload the entire book, they just need to grab the meta data parts. This metadata will be much smaller than what it takes for them to serve up a single web page when you go to their site.

Metadata from well-formatted ebooks is easy to grab. Metadata from poorly-made ebooks and personal documents is a mess.

Most purchased books (not interested in personal PDF documents that you made yourself)

If they're tracking sideloaded books, they need to deal with those... or they can just assume they're all legit, but that's an awfully big assumption.

will have a ISBN that looks like this in the OPF: <dc:identifier id="ISBN" opf:scheme="ISBN-HERE">ISBNHERE</dc:identifier>. That's a unique identifier for the book. You also have Title <dc:title> and Author <dc:creator> that in many cases end up being a unique identifier when combined.

But those will not tell you if the book was legitimately acquired as a non-DRM freebie (for example, when Tor.com was setting up, it distributed a number of free ebooks as promotional gimmicks), or a review copy, or was acquired while in a nation in which that book is in the public domain. (Not so likely with SOIAF... but there are a couple of nations that don't have copyright laws.)

And for good measure most books have <dcublisher> also. You talk about filenames, but if you have used Calibre and loaded books into Calibre then you must realize that filenames are irrelevant.

So is metadata, in a personally-created ebook. I can set the title, author, ISBN and publisher to anything I want it to say. While this isn't likely to happen through direct Calibre conversions, it *can* happen with Sigil, where a person grabs one ebook to use as a template for others, and doesn't bother to change the metadata.

I don't believe anyone actively monitors people's accounts. What happens is that all this information is stored in a database. In general, these databases are permanent and information can be sorted, search, manipulated in any way imaginable at any time in the future.

Hm. New Discordian project: distribute a swarm of free ebooks with bestseller ISBNs. Declare them to be parodies recognizable by silicon-based life-forms.

It's very hard for someone who hasn't worked with some of the tools out there to realize how powerful the technology for understanding and manipulating data is.

I have worked with enough to know how hard it is to restrict data to *useful* content, unless starting with strict limits like "only track books purchased at Amazon."

I don't think Amazon is interested in turning people in for DRM violations. I don't know if they track whether the files are encrypted or not. I think its an easy thing to do if they wanted from a technical perspective.

I think they could spot DRMd-vs-not-DRM'd content on a Kindle. I'm not sure they could easily track info (like title, author, etc.) about non-DRM'd, sideloaded content, because the formatting options are too great--they'd have to decide what to do with all the documents they can't get info from. I'm very unsure they'd be willing to open themselves up to the liability issues of attempting to tackle those isues.

And that's not a small, side issue to the technical side of things. "We don't track that info so we can't be held liable for getting it wrong, and can't be forced to hand it over" is a BIG reason to avoid it.

I think there is more of a concern with a company likes Barnes and Noble that is financially unstable and could be bought by anyone and sold off in pieces.

B&N has a *sharp* separation between books bought through it, and sideloaded content. They definitely don't track or sync the sideloaded content.