Monday, August 17, 2009

Smart Digital Books Metadata Notes #6

ePub eBooks are a dead-end for all publishers.


1) Standalone
2) Flat
3) Static
4) Unconnected
5) A pirate's paradise

Smart Digital Books are what print publishing need to continue existing:

Smart Digital Books:

1) Nodal
2) Multi-dimensional
3) Dynamic
4) Connected
5) A pirate's worst nightmare

In the above graphic of a smart digital book, I'm briefly illustrating the value of metadata connectedness to readers and especially to publishers. Axiom 3:
3 - Connections between books add value to all books

So a smart digital book from publisher A leads to two other books from the same publisher (A), as well as a book from two other publishers (B and C). (Note de Bono's rule here, however: "Patterns are asymmetric. The route from A to B is not the same as the route from B to A." So it might not be likely that the books on the right would lead back to the book on the left. This, however, is irrelevant in the overall network.)

The work of Dave Winer, father of RSS, is most likely important in all of this. Start paying attention to him, especially his rssCloud formulations.

@doctorlaura will probably again wonder about metadata versus marginalia in this example, but I'm blunting her objection here. I am not parsing out the underlying metadata that would make this connection. I'm saying this is the type of connection metadata will make possible.
The following statement made by a noted criminologist illustrates the point: "When men first come into contact with crime, they abhor it. If they remain in contact with crime for a time, they become accustomed to it, and endure it. If they remain in contact with it long enough, they finally embrace it, and become influenced by it."

-- Think and Grow Rich by Napoleon Hill, published 1937

But you don’t think about these trade-offs anymore. You’ve already made this decision many times in the past, so you now assume that this is the way you want to spend your money. You’ve herded yourself -- lining up behind your initial experience at Starbucks -- and now you’re part of the crowd.

-- Predictably Irrational by Dan Ariely, published 2008

I grant 99.9% of people will find this connection to be too obscure. That's because the Predictably Irrational story is too long for a Fair Use quote, so I've cut to the end of the tale. Also, people looking at this are seeing it from the outside, unlike someone -- um, me -- who has recently read both books and can see how they fit together. (While reading PI, I immediately thought of that passage from T&GR.)

The point is metadata will allow conceptual linking. Anyone interested in concept "A" will be able to extract that concept from within other books.

This is a knowledge cascade effect that adds value to the original book and opens the way for sales of other books for everyone.

To bring up Outliers once again. It has a chapter about the 10,000 hours required to achieve distinctive mastery. There is an entire book about that subject alone. How many people reading Outliers as a print book or as an eBook know that or would bother to find out? Such a connection would be possible with smart digital books, eliminating the friction current print and eBooks create for such discovery.

It's this connecting of smart digital books that will save publishing. It will lead to on-the-spot discovery for readers, it will require human expertise, the metadata and connections become a new capital asset, and all books increase in value to readers because new connections are being added all the time.

That race to the bottom pricing with flat, static ePub eBooks comes to a halt with smart digital books.

Moving on ...

A good overview of HTML5.

XML is being used for voice applications: VoiceXML at 10: Fueling Growth in Voice Apps and Hosting.

Previously here:

Smart Digital Books Metadata Notes #5
Metadata Is Money
Smart Digital Books Metadata Notes #4
Smart Digital Books Metadata Notes #3
Smart Digital Books Metadata Notes #2
Smart Digital Books Metadata Notes #1
Dumb eBooks Must Die, Smart eBooks Must Live


laura said...

Technical standards already exist to allow making the kind of conceptual links between books mentioned here and in installment #3.

Expressing such links is possible using metadata models such as RDF, whose triple structure allows not just creating links between resources, but specifying why those links are made through the use of predicates.

So the example here is NOT marginalia, but becomes a way of expressing that:

BookA hasConcept habituation


BookB hasConcept habituation

Thus, it becomes easy to see that these books both mention the same concept. In this way, readers can easily find other books that treat the same concepts.

The example in installment #3 was similar, only the concept being expressed was different.

The Outliers example could be quite easily expressed by triples of the type

BookC hasTitle Outliers
BookC hasConcept mastery
BookD hasSubject mastery

Where BookD is whatever book is dedicated to the mentioned subject.

The main question I have is related to the usefulness of these kinds of links. It might be less useful for me to know that Books A&B have mentioned the same concept than to know, for example, that it was Mike Cane who made the connection.

I think the conceptual links between books are necessary, but their real usefulness comes from knowing who made the links and to what else they apply.

Definitely this means that human expertise and opinion is essential, but I would argue that ontologies that allow expressing the relationships between concepts are also critical, whether human or machine generated.

Today, humans are much more reliable at identifying these concepts in natural language text, while describing them in unambiguous machine-understandable language (with formal semantics) is not always easy.

To do what you describe, we need something that is equally conducive to both machine and human reading and understanding. At least to my mind.

Mike Cane said...

I'd be more specific than "mastery" and I'd refine it down to the paragraph level (the T&GR cite is a single paragraph!).

Machines could do brute-force work, as they do today building a book index. But a human always has to go in there and make it reasonable and functional.

Provenance of a link is interesting!

laura said...

I agree.

I will also say that this series of metadata notes has be very useful and helped me clarify my thinking on this topic.