Practical RDF [Paperback]

The Resource Description Framework (RDF) is a suite of W3C specifications; consequently there is much information available on the Internet covering both the formal detail and related background reading.

For readers new to RDF, "Practical RDF" stands as a good companion to these online documents (it helped me no end); it introduces the concepts (triples, graphs, commonly encountered schemas, tools, query languages, implementations, etc) and provides a useful base narrative from which to read around the subject.

For a first edition the online errata numbers very few errors and I'd not been inconvenienced or led astray by any of them.

To my mind the letdown for this book is the index - and this is true of many books from this publisher. This is an XML subject and consequently namespaces are everywhere, so when I want to know (for example) about the "type" element, I want to look under "T" - I don't necessarily know or care what namespace prefix it was under when it was indexed, so the fact that "type" is indexed under "R" (for "rdf:type" which is therefore a very long and messy index section) and not also under "T" is a constant niggle.

Consequently as a reference for the more experienced RDF practitioner, "Practical RDF" is not as useful as a good search engine, however, notwithstanding this publication-specific snag, I can confidently recommend this book to the newcomer, who will appreciate it's cohesive coverage of what is (in places) a highly complex subject.

When I read page x of the preface, where the author describes herself as a "practical" person. This was possibly the only place in this book where I felt a connection with author. My expectations from reading this however were soon dashed... Have you ever finished reading a page from a book and then ponder to yourself. "Huh? What did I just read?" (Particular pain points that stand out for me were: P28 last paragraph. P47 on rdf:type Property).

90% of this book was like this for me. I just didn't connect with the author. The old adage "a picture is worth a thousand words" springs to mind. I found the material to be poorly organised, full of bad sentence construction: e.g. "The type element has a range that determines the type of values associated with it". This was some of the author's finer work! What about this? : "To this point, we've looked at recording only individual properties, but RDF needs to record multiply occurring properties." I read this a couple of times then spotted the typo. This is just the tip of the iceberg.
The author flits from topic to topic leaving you feeling like your suffering from information overload and nothing seems to register in a cohesive fashion. What do you think of this for mixed signals?: P77 Note says Working Group removed bagID. Then just below author goes on to describe obsolete method as best way of doing things! How about just sticking to the most current spec at time of writing. No need for excess baggage it muddies the waters.

I think the author really knows her stuff but didn't do a good job of storyboarding the book to make sure the newcomer could get to grips with the subject. The editors at O'Reilley need shooting. With the right pairing of editors, there is the potential for a fantastic book.

I've yet to be blown away by an O'Reilly book. They really could do with following Manning's lead of providing a forum to interact with author and breaking examples down by numbered bullet points helps. P104-105 is a prime candidate for this topic. At tail end of P105, I wasn't sure if I was supposed to be looking at the Element column or the Property column for "Content" element. "Resource Bio" doesn't jive with either!

The text is plastered with errors and outdated links. If you look at the confirmed/unconfirmed O'Reilly errata pages, you'll see what I mean... I reported a load more issues that aren't in the unconfirmed draft even as we speak. So who knows what the complete picture is! I'm extremely disappointed these have not been fed back in to a reprint. Five years have gone by since book was first published. Come on guys, you are not going to impress your customers with this kind of sloppiness. (I have since heard back from O'Reilly and it appears there was a reprint in Nov 07. I have that version and unfortunately loads of errors do still exist!)

I think my biggest gripe stems from the choreography of material covered. Its all 'back to front' and just floors the beginner. My gripe about P28 and all its techno babble glory is a prime example of this "node-edge-node pattern". Finally author gets around to discussing this on page 35 when she discusses Striped Syntax! The problems stem from the fact the author does everything in a 'top down' fashion. When you're building a semantic vocabulary, it helps to give a high level 'top down' overview, but when you get to the nitty-gritty details, describe the foundations first and build upon them in a 'bottom up' fashion.

This book would get zero stars if I could award it that. It gets its one star for harvesting the links which may help me further unravel RDF. At the same time I may get a more up to date perspective on obsolete references like P122: ublin Core dcmes-xml! P149: Jena 1.6.1 (2.5.6 as of now). P151: DAML+OIL. Replaced by OWL (Do a Wiki search). P157: rdfcontent url leads to project with nothing to download. Groovy!
P312-314: Also whole section in Ch15 on Tucana Knowledge Store. From what I can tell this got bought up by Northrup Grumman. Can't find it anywhere. Maybe it got renamed.

On a final note, I found the author to be most unhelpful when I did ask for an example to break down what author described on P28. I got a sense of, "Thanks for your patronage and buying my book. Now get lost"!