Scholarly publishers want to produce quality longterm-durable books. I get that. But some of us young digital guerilla turks want to combine 3D and augmented technologies so 3D models can be shared and experimented (played) with.
And 3D models should be able to change over time, to link to different scholarly resources and models and links and linked resources should be able to be maintained and modified.

So how do we have stable print materials and changing, dynamic 3D models?

Easy peasy.
Consider the magic book. You put on these special see through glasses. Open a book. The camera recognises the augmented reality tracker (marker) on an open page and on your see-through glasses is projected a virtual 3D object. It can move, it can have animations. But typically the AR 3D object relates to a point on the page. Now I believe the original marker and related augmented 3D shape was stored on the magic book glasses/goggles.
But this won’t do for scholarly preservation.
Say you open a book, You hold your camera phone over a mark (tag, tracker) on the page of the book. On your phone lens appears a 3D playable object. But the phone does not natively hold the 3D virtual object.
Instead, the tracker/marker on the book induces the phone to search through an online library and retrieve the current 3D virtual object that links to that augmented reality tracker.marker/barcode on the book.
And downloads the most recent virtual model.
Why is this useful? The scholarly publisher only has to produce a normal book but with augmented reality markers (as an image).
The library or academic organisation supplies the links (perhaps the 3D virtual object has its own URI). The phone retrieves the most recent 3D AR object from a database (online) maintained by a library or scholarly community.
And as you open/read the book, augmented reality models dance/float/appear on your book (or iPad or Android tablet).
Perhaps the way you move the tablet/eReader changes the appearance or animation of the AR object (for example: lift it up and the model changes forwards in time,lower the tablet and earlier versions of the model appear)..
And theoretically the book will still be useful even 10 years from now if someone maintains the digital assets available at the URI that the book markers point to.

Still don’t understand? Perhaps I need a diagram!

NB below photos are not mine but from http://masters.donntu.org/2012/iii/akchurin/library/article9.htm in “Collaborative Augmented Reality” written by Mark Billinghurst, Hirokazu Kato, Communications of the ACM – How the virtual inspires the real, Volume 45 Issue 7, ACM New York, NY, USA – July 2002, p. 64–70.