Maybe I missed you discussing this somewhere (I haven't read all the comments), but it seems to me that Computer Science identified itself as involving discrete mathematics, combinatorics, etc., rather than continuous maths, numerical analysis, etc. I'm thinking of Knuth's "Art of Computer Programming" and of course his "Concrete Mathematics" as quintessential computer science maths books.
I do think there are lots of interesting stories to tell about how new disciplines get founded. We saw this a decade or so ago with Data Science, and I have heard rumblings that the LLM and GenAI crowd are saying they deserve their own field also. (If money is what spawns disciplines, then I guess they have earned it).
The exclusion of continuous math happened much later. In 1961, there were still many numerical analysis and operations research folks at the core of the computing efforts. It was the middle of the Cold War, after all.
Given that the most killer app of GenAI has been cheating at scale, it would be deliciously ironic to build an academic department around it. The chair would have to be a chatbot.
Yes, that's true. The Computer Science department at Sydney University (where I did my undergrad) still had Numerical Analysis courses in the late 1970s.
It's also interesting to think about failed attempts to create new disciplines. You mentioned Operations Research; I remember the enthusiasm around simulations in the 1970s and 1980s. The ability to simulate and the ability to optimise are incredibly powerful, as we read about in your blog most days. But is there a disciplinary home for this? Maybe Economics?
The nature of universities has changed in the last 50 years also. While there always were technical universities, I think many more universities are not trying to cover all the disciplines. This is showing up mainly in Arts and Humanities, but I think it's a broader trend.
I think the GenAI people are still mainly concentrating on building their own conferences. The evolving relationship between AI, Computational Linguistics and GenAI is interesting to watch, because as far as I can tell, essentially the same kinds of GenAI papers get published in all 3 venues.
The Synnoetics story is interesting because there was a similar push for a new discipline to handle the explosive growth of molecular biology data, from DNA and protein sequences to gene expression. It was called bioinformatics. The resistance was reasonable, as there were no such disciplines in other sciences, such as physics and chemistry, where computational techniques were devised by physicists and chemists adept at math and using computers. Biology was different, with most biologists unable to do much with either. Interestingly, bioinformatics is a practice that, AFAIK, is integrated into biology departments, although it has its dedicated journal[s].
In my view, computation should be like writing, a skill that should be attained at some basic level. Just as most people cannot write elegant prose at novel length, they probably will not be able to write large software programs, but they should be able to solve computational problems with computer languages and libraries. These days, languages and documentation are readily available on the internet with few barriers to learning. Unlike writing, coding allows libraries to be glued together to solve problems with less effort. GenAIs are even doing that work with prompts.
Maybe I missed you discussing this somewhere (I haven't read all the comments), but it seems to me that Computer Science identified itself as involving discrete mathematics, combinatorics, etc., rather than continuous maths, numerical analysis, etc. I'm thinking of Knuth's "Art of Computer Programming" and of course his "Concrete Mathematics" as quintessential computer science maths books.
I do think there are lots of interesting stories to tell about how new disciplines get founded. We saw this a decade or so ago with Data Science, and I have heard rumblings that the LLM and GenAI crowd are saying they deserve their own field also. (If money is what spawns disciplines, then I guess they have earned it).
The exclusion of continuous math happened much later. In 1961, there were still many numerical analysis and operations research folks at the core of the computing efforts. It was the middle of the Cold War, after all.
Given that the most killer app of GenAI has been cheating at scale, it would be deliciously ironic to build an academic department around it. The chair would have to be a chatbot.
Yes, that's true. The Computer Science department at Sydney University (where I did my undergrad) still had Numerical Analysis courses in the late 1970s.
It's also interesting to think about failed attempts to create new disciplines. You mentioned Operations Research; I remember the enthusiasm around simulations in the 1970s and 1980s. The ability to simulate and the ability to optimise are incredibly powerful, as we read about in your blog most days. But is there a disciplinary home for this? Maybe Economics?
The nature of universities has changed in the last 50 years also. While there always were technical universities, I think many more universities are not trying to cover all the disciplines. This is showing up mainly in Arts and Humanities, but I think it's a broader trend.
I think the GenAI people are still mainly concentrating on building their own conferences. The evolving relationship between AI, Computational Linguistics and GenAI is interesting to watch, because as far as I can tell, essentially the same kinds of GenAI papers get published in all 3 venues.
The Synnoetics story is interesting because there was a similar push for a new discipline to handle the explosive growth of molecular biology data, from DNA and protein sequences to gene expression. It was called bioinformatics. The resistance was reasonable, as there were no such disciplines in other sciences, such as physics and chemistry, where computational techniques were devised by physicists and chemists adept at math and using computers. Biology was different, with most biologists unable to do much with either. Interestingly, bioinformatics is a practice that, AFAIK, is integrated into biology departments, although it has its dedicated journal[s].
In my view, computation should be like writing, a skill that should be attained at some basic level. Just as most people cannot write elegant prose at novel length, they probably will not be able to write large software programs, but they should be able to solve computational problems with computer languages and libraries. These days, languages and documentation are readily available on the internet with few barriers to learning. Unlike writing, coding allows libraries to be glued together to solve problems with less effort. GenAIs are even doing that work with prompts.