Rob Roi: Open Source and Cost of Technology

Article excerpt

Rob Roy was the fourth in Walter Scott's Waverly novels series. Maybe you remember the 1995 film starring Liam Neeson rather than the novel. If you prefer the book, you can access Rob Roy as an etext on various websites identified by Edinburgh University Library's Walter Scott Digital Archive (www.walterscott.lib.ed.ac.uk/etexts/novels.html#robroy), or, hopefully, you can borrow it from a local library.

I recall this statement from Chapter 24: "A coward calls himself cautious, a miser thrifty." It seems an appropriate commentary for today's views on open source. Open source software seems to be a way to bring "thrift" to the enterprise. "Roi" today means return on investment, not king--although some chief financial officers would beg to differ.

ECONOMIES OF OPEN SOURCE

Is open source software more economical than proprietary software? The answer to this question, like so many in today's fractious world, is, "It depends." A misstep in information technology produces a financial challenge. Stakeholders are often quick to point fingers at managers who rob ROI from the organization's balance sheet.

Most information professionals know that business intelligence is one of the buzzwords competing for their attention. The idea is that software can process structured and unstructured information and generate answers needed to make decisions.

Few software systems are as complex as those that perform business intelligence. The difficulty of handling database or structured content and the torrents of unstructured information in the form of email, Word documents, and Adobe PDF files is high. Content transformation in some organizations consumes as much as 30% of their information technology budgets.

OPEN SOURCE BUSINESS INTELLIGENCE SOFTWARE

In the past year, open source business intelligence software gained momentum in a number of government agencies and commercial enterprises. One of the higher profile vendors is Pentaho Corp. (www.pentaho.com), a privately held company based on open source technology. The company offers a business intelligence suite without a punishing annual licensing fee. The Pentaho system operates on-site or from the cloud. Pentaho puts reporting, analysis, data integration, dashboards, and data mining within reach. The company asserts, "if you want to speed your BI development, deploy on-premise or in the cloud or cut BI licensing costs by up to 90 percent," Pentaho is a solution for many organizations.

If you have looked at products from such proprietary business intelligence vendors as Cognos SPSS (IBM) or Business Objects (SAP), you know that these systems are complex. The same can be said for enterprise search, accounts payable, or enterprise resource planning systems. The next generation of business intelligence products and services offer an enticing plum--to wit, you no longer need specialized knowledge of statistics, mathematics, or data integrity to get the benefits of business intelligence systems. The idea is mouthwatering, but it flies in the face of such information technology realities as "nothing worthwhile comes easy."

UNDERSTANDING THE MECHANICS

What still distinguishes enterprise business intelligence from an application such as word processing is the need to understand the underlying plumbing, such as specific characteristics of a data set or the particular statistical method's implications. Outputs without understanding the upside and downside of an approach can lead to off-base or out-of-kilter decisions. When faced with a choice among statistical methods and complex numerical recipes for performing regression analysis to understand a market trend, most professionals understand that there may be piranhas and snakes in the data pond. A college-level Introduction to Statistics course is essentially useless when business intelligence systems become available. Expertise, therefore, adds to the cost of even an open source business intelligence system as capable as Pentaho's. …