Computer Magazine Banner

Industry Perspective Column

« Back to Viewpoint: The Future of Computing Practice and Education

Reading the classics: Broad first, then deep

Combination View Flat View Tree View
Threads [ Previous | Next ]
I just finished a stint at MIT/CSAIL as a Visiting Scientist. The atmosphere was stimulating. There were many pleasant surprises. For example, it was great to observe what appeared to be a large increase in the percentage of women students.

But one of the things that I was surprised and disappointed about was in discussions with post-docs and grad students about some of the seminal papers in the field. They haven't read some of the early important papers! MIT has a great library. IEEE and ACM have great digital libraries. I downloaded and reviewed the papers in the 25th Anniversary issue of the Communications of the ACM.

My advice? Every grad student should read all of the papers in that issue, and faculty should put together a reading list of what they consider the important papers covering not only their narrow research interests, but the broad body of knowledge this field has become. I think this should be done before graduate students are allowed to specialize. My advice? Students need to go broad first before they are allowed to go deep into their chosen narrower area of specialization.
RE: Reading the classics: Broad first, then deep
2/27/10 3:14 AM as a reply to Isaac Nassi.
I like the idea of broad first, then deep - as well as the idea of collecting the list of seminal papers.

I would include one of my all time favorites:
Computer science as empirical inquiry: Symbols and search
A Newell, HA Simon - Communications of the ACM, 1976

Students would of course need the right progression, since foundations are required to understand the seminal papers.

Also, broad and deep is often called T-shaped --- here is post that agrees with you...

"The kind of people who will be best able to seize these opportunities are those I call 'T-shaped' as opposed to 'I-shaped.' ...they have their specialties—areas of deep expertise—but on top of that they boast a solid breadth, an umbrella if you will, of wide-ranging knowledge and interests. It is the ability to work in an interdisciplinary fashion and to see how different ideas, sectors, people, and markets connect. Natural-born T's are perhaps rare, but I believe people can be trained to be T-shaped. One problem is that our educational system is still intent on training more I's. We need to change that."

From today's post on "Innovation that Matters", By Nicholas M. Donofrio, Senior Fellow, Ewing Marion Kauffman Foundation...