Published in IT World
The same data in the same place
By Sean Mc Grath
Leopold Bloom has an excellent definition of the word 'nation' . A nation, according to Bloom, is 'the same people living in the same place'.
I like that. Bloom's conceptual model of a nation either beautifully sidesteps the question or gives a deeply insightful answer to it. Take your pick.
It never ceases to amaze me how critically important conceptual models are in determining the structure of enterprise applications. Take business-level reporting for example. The goal of a good report is to consolidate and summarize key information into a single place. Typically, this is a sheet of paper or its electronic equivalent such as a Web Browser.
Now, imagine for a second that you are a busy executive with no IT knowledge. You have two separate, computer-generated reports on your desk. One report gives you excellent consolidated and summarized information on business activity A. The second report does an equally stellar job on business activity B.
You find yourself punching numbers into a spreadsheet in order to create a consolidated view of the information in report A and report B. Eventually, you tire of this and call in IT.
Executive: I want a consolidated report covering A and B.
IT: But A and B are generated by two different computer systems.
Executive: Then go build one computer system that combines all the features of systems A and B so that I can get my consolidated report.
Executive: Just do it.
Behind this little exchange lies an assumption on the part of the executive that, in most cases, is not correct. An expensive assumption that can lead to unwarranted system redevelopment. The underlying assumption is that in order to get consolidated reporting, the data underlying the reports must all be made 'the same' by putting it all in 'the same place'. In other words, merging the two separate computer systems.
The reality is that it is often possible to achieve executive-level 'sameness' of data for reporting purposes without re-engineering existing applications. All that is required is for the relevant reporting data to be pulled from the separate systems to create a throw-away, read-only silo of data for consolidated reporting. No significant re-engineering required.
These days an excellent weapon is available to those who need to communicate this idea. Namely, web search engines. Think of a search engine as a reporting tool. The search engine has the effect of consolidating information into one place for reporting purposes. How does it do it? Each individual system simply creates replicas of key information in the form of Web pages. The search engine then indexes these web pages. No significant re-engineering required.
The same applies to other forms of consolidated reporting. More often than not, there is no need to physically consolidate computer systems in order to consolidate them at the reporting level.