Software is like this too
John Baldessari is a painter whose works I return to frequently. He always makes me think, and not only about painting. After sharing this painting with friends over the years, I realized only recently that this is also a description about my relationship to software development. As I look back at the trends in my work, the core reason for developing any of the companies, products or algorithms that I have created in the past was to embody the most recent understanding of a burning question.
Early in my career, I was deeply intrigued by image processing in the human visual system. How do we see what we see and make sense of it? Cognitive processing of the visible world is quite a miraculous phenomenon. At the time, large scale image database retrieval was the particular burning issue that occupied my every creative moment. At the birth of digital imaging there was the creation of a method to encode an image for storage and display. But that very representational scheme did not account for the time when there would be massive quantities of image files that needed to be found, all more or less named nnnn.jpg.
My burning, primary daily question was: “How do people search for images?” This created a number of sub-questions all hanging off the first. Once a result set is gathered, how do users continue the filtering process until a final selection is made. What representational model, metadata and methods need to be in place for the most flexible downstream retrieval? Is it possible to create image metadata with “bounce”, meaning, can you shift the task point of view (i.e., throw the query ball from a different direction) on a single representation and still get a relevant result set.
These burning issues eventually were only partially resolved in a variety of patents, companies and products over the years: eMotion, PictureQuest, Picture Network International, ClearView Networks, and a host of others. In painting terms, these were the “group exhibitions” I was a part of, since there are very few “one man exhibitions” in software.
When I talk to my tech friends I always inquire about what they are working on, hoping to hear what they are “really” working on. I may now start asking them more bluntly, what are you exhaustively studying in order to build your product … and when will your group exhibition open?
