Content Transformation: Easy, No. Expensive, Yes
July 27, 2009
Content transformation is an expensive proposition. I finished a project and learned that an in house information technology staff did not think about content transformation as a cost sink. The manager assumed that software handled most of the “heavy lifting.” The staff delegated the script fixes to deal with exceptions to contractors. The contractors billed their hours as “programming”. No one looked at the bills. A quick review by my researchers revealed that about one third of the information technology budget was consumed by content transformation; that is, taking information object A in one format and converting it into information object A1 with changes needed to make it indexable, searchable, findable, and updateable. In short, the technical budget was struggling with costs that no one took the time to look at from a bird’s eye view.
You can see why content transformation is expensive. There is a very useful write up from Sys-con outfit called “Processing XML with C# and .NET” with the subtitle “A solution that’s simpler than you might expect.” Yeah, right.
The guts of the article is a rundown of how to convert or transform XML with Microsoft’s tools. The most useful part of the write up is the included code snippets. Very nice. The omission, however, is commenting on how long it takes to adapt this code to handle exceptions from tera-scale information processing flows. There was a question in my mind about the performance of the methods based on the code samples. I also wanted to know, “How can one manage the many different variants of the scripts required to handle the rag tag collection of XML objects that flow through some search and content processing systems?”
To be fair, the write up was aimed at a certified Microsoft professional with a competence in Microsoft Dot Net and C#. The problem is, however, that the cost of this method is a big deal for the chief financial officer who has to budget for code fiddles.
The cost of transformation is a very big deal, and I think economical methods need emphasis. Methods that sell more expensive consulting are likely to run into some rough water.
Stephen Arnold, July 27, 2009