Metadata Solutions: Using Metamodels, Repositories, XML, and Enterprise Portals to Generate Information on Demand [Paperback]

The author covers the topic thoroughly, starting with an exhaustive discussion of information and its value to business that spans the first six chapters. Some excellent case studies are included to reinforce concepts. Part II consists of five chapters that take information up one level of abstraction and introduce metadata. Topics in this section start with requirements, and move to modeling. It's worth noting here that the OMG Open Information Model notation is heavily used in this section of the book. This should not be a problem because the notation is easy to follow and is widely known and used, reducing any learning curve to zero for experienced architects and modelers.

Part III's three chapters delve into the underlying infrastructure technologies and tools. I especially liked the chapter on repositories, as well as the one that provided a balanced view of standards. Part IV is where the pace picks up with four chapters devoted to factors you need to consider before designing the solution. The case studies that are included in some of these chapters are important reading because they portray the issues and challenges that are associated with the factors. The last chapter in this section leads you through the design selection process.

I especially like Section V, which is a catalog of solutions that thoroughly examine each aspect of the solution. It starts off with a chapter that describes a typical disaster, followed by viable solutions presented in a format that can be easily recast into design patterns. I also like Section VI, which covers post-implementation administration and maintenance. The three chapters in this section cover topics that are often overlooked until a solution is ready to go into production or (in too many cases) after it has been released into production.

Key points about this book: (1) It is unique in that it addresses the data abstraction requirements of portals, and does so comprehensively. (2) Provides an end-to-end view of metadata from concept to final solution using real life examples and sound techniques. This is a refreshing change from many books that address metadata, but are so abstract themselves that it is difficult to transform the concepts into a working solution. (3) Addresses post implementation issues, which is something I have never come across in a book on metadata.

This is an important work and is essential reading for key players in a portal design and implementation project. Five stars and my highest recommendation.

As a practicing Data and Metadata Architect, I have noted with interest Ms Tannenbaum attempt to offer solutions to a very difficult problem namely first of articulating what metadata is and then managing it to make it useful in full understanding of the data/information/knowledge chain within an organization.

The book traces historical roots of the issue by introducing the data/ information systems developments in organization and respective development of complexity. She then introduces the world of metadata. The treatment of identification of what metadata is, various meta-metadata models is indeed in-depth but makes very dense reading (only data bigots will love these sections). Her treatment of solution is also covers a range of possibilities and various case histories at the end of book only go on to illustrate that metadata problems and solution are varied and complex.

The book suffers from two omissions. One, is that the definition of the metadata is leaves out a significant area namely that of the business metadata i.e. metadata that would adds further information/understanding to organized data e.g. reports. Examples of Business metadata are industry reports, internal documents and communication which form a major portion of the current metadata universe but receive scant attention.

The second omission is examples of what constitutes good metadata vs. mediocre (bad) metadata. Writing good definition is no trivial task. "Receipt_Date - Date when the goods are received" is typical of the definitions in any organization dictionary and yet these definitions are less than useless as they lead you believe that the data(s) are defined when in fact they are not. Take for instance an oil tanker disembarking it's cargo which may take over two days - what is the Receipt_Date?

Overall the book tries perhaps to cover too large an arena and fails to offer depth that would make it more useful.