TITLE: Do We Need Folks With CS Degrees? AUTHOR: Eugene Wallingford DATE: February 19, 2008 5:11 PM DESC: ----- BODY: Are all the open jobs in computing that we keep hearing about going unfilled?
Actually -- they're not. Companies do fill those jobs. They fill them with less expensive workers, without computing degrees, and then train them to program.
Mark Guzdial is concerned that some American CEOs and legislators are unconcerned -- "So? Where's the problem?" -- and wonders how we make the case that degrees in CS matter. I wonder if the US would be better off if we addressed a shortage of medical doctors by starting with less expensive workers, without medical degrees, and then trained them to practice medicine? We currently do face a shortage of medical professionals willing to practice in rural and underprivileged areas. The analogy is not a perfect one, of course. A fair amount of the software we produce in the world is life-critical, but a lot is not. But I'm not sure whether we want to live in a world where our financial, commercial, communication, educational, and entertainment systems depend on software to run, and that software is written by folks with a shallow understanding of software and computing more generally. Maybe an analogy to the law or education is more on-point. For example, would the US would be better off if we addressed a shortage of lawyers or teachers by starting with less expensive workers, without degrees in those areas, and then trained them? A shortage of lawyers -- ha! But there is indeed a critical shortage of teachers in many disciplines looming in the near future, especially in math and science. This might lead to an interesting conversation, because many folks advocate loosening the restrictions on professional training for folks who teach in our K-12 classrooms. I do not mean to say that folks who are trained "on the job" to write software necessarily have a shallow understanding of software or programming. Much valuable learning occurs on the job, and there are many folks who believe strongly in a craftsmanship approach to developing developers. My friend and colleague Ken Auer built his company on the model of software apprenticeship. I think that our university system should adopt more of a project-based and apprenticeship-based approach to educating software developers. But I wonder about the practicality of a system that develops all of its programmers on the job. Maybe my view is colored by self-preservation, but I think there is an important role for university computing education. Speaking of practicality, perhaps the best way to speak to the CEOs and legislators who doubt the value of academic CS degrees is in their language of supply and productivity. First, decentralized apprenticeship programs are probably how people really became programmers, but they operate on a small scale. A university program is able to operate on a slightly larger scale, producing more folks who are ready for apprenticeship in industry sooner than industry can grow them from scratch. Second, the best-prepared folks coming out of university programs are much more productive than the folks being retrained, at least while the brightest trainees catch. That lack of productivity is at best an opportunity cost, and at worst an invitation for another company to eat your lunch. Of course, I also think that in the future more and more programmers will be scientists and engineers who have learned how to program. I'm inclined to think that these folks and the software world will be better off being educated by folks with a deep understanding of computing. Artists, too. And not only for immediately obvious economic reasons. -----