eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
In his keynote speech this month at the Internet World Exhibition in New York, Compaq Chairman and CEO Michael Capellas proposed a dangerously self-serving view of the mission of IT vendors.
Predicting greater enterprise emphasis on what he called “rich media,” Capellas said, “Were used to looking at data that has nice rows and columns … [but] unstructured data means new user interfaces.” Oh, please, no.
I understand Compaqs temptation to define progress in terms of ability to deliver more—and more and more. The easy way to improve IT hardwares price/performance ratio is to define performance in terms of what gets cheaper, that is, in terms of pumping bits.
A voice mail uses more bits than an e-mail of equal content, and video mail is even more extravagant. When an inventory report is a video walk-through of the warehouse, instead of a table of quantities and part numbers, Compaq sells more network and storage hardware; when someone has to skim quickly through that video or turn it into something that can be searched or compared with other “unstructured” reports, Compaq sells more CPU power.
But the true performance measure of IT must include a human element as well, and human performance rises when fewer bits occupy our attention. Thats the value-add of software, especially when supported by professional services.
We think better, as well as communicate better, in systems that enable abstraction. Consider the Roman alphabet: It shapes our way of thinking into one based on encoding and classification, including the simple but powerful notion of sorting lists in alphabetic order. The ordering of Chinese pictographic dictionaries can follow any of several principles—none of them manifest in the symbols themselves. Is this the model Capellas has in mind?
Interfaces should clarify; data should be (in the words of Polonius) “rich, not gaudy.” Hardware vendors may long for the demise of rows and columns, but developers should be finding order and distilling meaning in new data as in old.
Whats wrong with rows and columns? Tell me at peter_coffee@ziffdavis.com.