on 2 October 2010
This is a hugely important and influential work on the subject of designing a data warehouse solution.
Inmon has completely different opinions to Kimball in many key areas. For example Kimball advocates a dimensional schema for the warehouse, whereas Inmon insists that a relational structure must be employed.
Kimball is very clear in his books that he believes that data-marts must be aligned to business processes whereas Inmon maintains that data-marts should be aligned to business departments.
Disagreeing with Kimball on the subject of data warehousing is a bit like disagreeing with Newton on the subject of gravity. But Inmon has such experience and authority within this field that he can do exactly that and still be credible.
It's really refreshing to read such a radically different perspective, even if at times I may not agree with everything Inmon writes. To those people who will only entertain the writings of Kimball and dismiss all others I would quote the old adage about every problem looking like a nail when the only tool you own is a hammer.
If you want to get a decent understanding of data warehousing then read this book, Kimball's "Data Warehouse Lifecycle Toolkit" and Kimball's "Complete Guide to Dimensional Modelling".
Really think about the differences in the approaches between the two authors without being judgemental, and you will probably understand Data Warehousing more than 99% of the so called experts in the industry today.
on 28 December 2007
If you work for a large corporation which has millions of $ to spend on DW projects, maybe you should look at this book and even consider some of the ideas that it contains.
But if you need to develop a data warehouse using limited resources and within a certain timeframe, your time will better used reading other books, because following the Inmon approach will lead you to an unnecessary complicated and expensive design.
I found that the arguments used by Inmon to demonstrate the limits of the dimensional approach are not convincing at all. For example, at page 142 he says "Because there is a different data structure for each data mart, making any data mart into a data warehouse doesn't make sense."
Having personally implemented several data warehouses using the "conformed dimensions" approach, I can guarantee that it worked and produced a very elegant and clean data model.
on 21 April 2011
I needed this book to get a clear perspective of what the Inmon approach is. And I think it is good for that purpose only. It is very haphazard - especially if to one who has gone through the Kimball works.
The material repeats itself simply because there is no clear approach. Also, I agree with some other viewers that practical world doesn't follow this approach. No business is willing to pay a fortune for a solution that does not deliver a high performance reporting solution within a couple of quarters at most. A 3NF model does not deliver that and the duration is hardly enough to take the solution through phases outlined in the book.
The approach defined here is great in theory and is a lip smacking opportunity for any person who finds a sponsor patient and visionary enough to implement it. The book does not help for any sort of implementation as it lacks content and clear approach.
It is also a bit out of line in outright rejection of star schema given that it is being adopted (successfully) by designs of market leading products.