I enjoyed reading David Diamond’s new book Metadata for Content Management. He’s covering everything a DAM system or CMS customer needs to know about metadata fields, contents and policy and how they affect search, workflows and interoperability. The book is written for practical use, in plain and fun language and with excellent examples for every single point he makes. (Disclosure: I’m listed on the “Thanks” page of the book so I’m slightly biased.) Check out the table of contents, an excerpt and Ralph Windsor’s review.

Good metadata schema design rarely gets the attention it deserves in DAM projects. I’d ask our project managers to send a copy of David’s book to each new DAM system customer so they can prepare better for the requirements workshops we’re doing… if only Adaptive Metadata (as implemented by David’s employer, Picturepark) weren’t so prominently featured in the book. That’s not to criticize the book – Adaptive Metadata is a great concept, and it makes sense to use it throughout the book since it helps keep the examples clear and simple. Our DAM product has somewhat similar functionality (form fields can dynamically be shown or hidden based on arbitrary metadata), but it’s lacking a few features and a good UI. So I guess our customers would, after studying the book, expect our product to do things it cannot yet do. (Which is fine with me – to quote my last blog post, I “love the product we want to build, not the one we already have” – but a problem for our project manager who needs to get things done with what’s available right now.)

My only criticism of Adaptive Metadata is that in some cases, it might lead you to assigning metadata directly to an asset when it should really be on a separate object. David’s examples include modeling workflow metadata as (adaptive) asset metadata. This can be problematic in systems where multiple workflows might be going on at the same time. Example: When two publications are using the same image, one might already have it approved by the editor but to the Photoshop guy is still working on it, while at the same time, the other publication’s editor isn’t yet through with it but their image editing is done already. The best way to model this is to have separate “workflow records” (each with their own set of metadata), not metadata fields directly on the asset. (Though it’ll be hard to have separate objects work as nicely in the UI, so you’ll be better off using Adaptive Metadata if you don’t need concurrent workflows.)

Which leads me to my (short but) favorite section in David’s book: “Everything becomes a Linked Object”. It’s about how metadata attributes, like colors or countries, are actually objects with attributes of their own, and “everything is linked to everything else”. I’ve been writing about this in It’s content, not just DAM metadata and DAM and the Semantic Web, and I love Topic Maps because they’re the perfect data model for it. To quote from the book: “Most content systems still don’t offer functionality that supports this concept of universal linking. Expect that to change.” That’s one of the things I’m really looking forward to…

Thanks for a great read, David!