I keep seeing suggestions that people don’t need to learn programming because computers will do it better. I am skeptical. In part I am a skeptic because I have been hearing this sot of thing for scores of years and I don’t think we’re getting close to this that quickly. But more importantly I worry that it will take language processing a long time to get good enough at understanding people and language.
For example you may have seen this old joke:
A wife asks her husband, a software engineer "Could you please go shopping for me and buy one carton of milk, and if they have eggs, get 6!" A short time later the husband comes back with 6 cartons of milk. The wife asks him, "Why the hell did you buy 6 cartons of milk?" He replied, "They had eggs."
I like to share that with my students to help them understand ambiguity in language. Today I saw this somewhat macabre version.
I’m still not 100% sure I want to use this one in class but it does highlight the risks of unclear speech.
One of the things I learned developing applications early in my career was that users use language that works for them and for people in their field but which may not be clear to others. It is important to understand their vocabulary, their understanding of the tasks, and how they work. Without a common set of assumptions trouble is an unavoidable result.
In the first joke above there is no clear connection between eggs and milk and bad assumptions are made. I’ve learned to question my wife closely about the shopping lists she gives me. Things that are totally clear in her head may be totally opaque to me.
Will computers be able to know how to ask the right questions? Or will they make assumptions of their own? I’m guessing this will take a while to work out. In the mean time we, people, are going to have to be very careful about how we instruct computers.
No comments:
Post a Comment