Historical Computer Science

Java Logo

Like any well-trained Ph.D. student, I have come to see my own discipline as the master discipline, upon which all other forms of knowledge are based. For instance, I have repeatedly pestered my fianceé, who works in math education, with the idea of teaching mathematics historically. What better way (I enthuse) to teach, say, imaginary numbers than to understand why they were invented in the first place; the historical context that led to their emergence.Still convinced (perhaps quite foolishly) that this is a brilliant idea, I have begun to think recently about how the same concept might apply in computing--how, that is, the history of computing might be used to teach computer science.As an opening first stab at such an approach, I propose that the best way to really understand a programming language is to understand its history. Java provides a case in point: Java was invented by Sun engineers in the early 1990s.  Its original intended purpose was as a language for interactive television, but it was soon evident that interactive TV was not taking off as hoped, while a new phenomenon known as the World Wide Web certainly was.  So Sun re-positioned Java as the language of the Web. This historical context, I propose, can explain many of the features of Java:

  • Why does it have C/C++-like syntax?  Because its creators were steeped in the Unix world of Sun, which made was based primarily on C.  They also hoped to attract others like them with a familiar syntax.
  • Why does it run on a virtual machine? Because it was intended to operate in an on-line environment, where programs would be downloaded over the network onto a variety of different machines.  The virtual machine provided both cross-platform code compatibility and the potential for a secure 'sandbox' where untrusted code from the network could operate.  
  • Why is it object oriented? In part because object-oriented code fit well with a client-server model of networked computing: only the class files needed on the client-side.

The advantage of this approach would be to connect otherwise abstruse concepts like object-oriented programming with human motivations, which might make them easier to learn and understand.  It would also give a better sense of the openness and potential of computing.  Rather than simply seeing Java as a given (this is what we have to learn, this is what programming is) students might get the sense that there are many, many ways to design a programming language. What do people think?  Is there anything to be said for using the history of technology to help teach technical disciplines?


Funny you should mention that thought. Bill Aspray had a similar inspiration during his time at the Computing Research Association and got an NSF grant to run some workshops in 2001 and publish a book on the topic. The workshop presenters were primarily historians and those with an interest in history who had tried to teach technical students. CS educators were supposed to come and learn from us. It was an enjoyable and interesting process, but I was left mostly with a sense of the challenges involved in actually doing this. Seven years teaching in a library & info science school has only reinforced that. The main problem is that in CS and professional disciplines teachers tend to think of history as factual material like dates and capsule biographies, of which students should perhaps know a smattering. Whereas we think of history as a set of core skills to do with interpreting sources, understanding connections, exposing embedded assumptions, etc. So if they do try to incorporate history they teach the former, not the latter. Also that most tech students have chosen to take tech courses to learn specific tech skills, rather than to engage in a liberal arts style education. So they don't see the point. So my summing up is that we all tried valiantly and came up with a good case for why CS should be taught from this perspective, but not really for how it could be. However don't let my jadedness put you off!

The book is at http://www.cra.org/uploads/documents/resources/workforce_history_reports/using.history_.pdf and includes an article I wrote to try to tell computer scientists how the history of computing really worked as a scholarly field and why they should care, as well as an early version of what became the SIGCIS resource guide, and contributions from SIGCIS members such as Atushi Akera (Aspray's coeditor),  Jen Light,  Paul Ceruzzi, Nathan Ensmenger, etc. Bill sent me some spare paper copies so if anyone wants an original bound version I'd be happy to mail it out.

Thanks, Tom for the very useful reference, and for the dose of reality :-)