by Flemming Funch
Slashdot has a thing on functional programming languages. Which I didn't really know much about. But functional programming might best be compared with how a spreadsheet works, or how the SQL database query language is working. I.e. you specify the structure of the result you want, rather than having to figure out step-by-step how it would come about.
I'm a programmer, and despite useful advances such as Object Oriented Programming, my job would mostly consist of coming up with steps for solving problems. There are big complex problems I'd kind of love to solve, but which I can't, because I haven't figured out the steps that would solve them. So, it is refreshing with different ways of thinking about stuff like that. Where you rather work on specifying the results you'd like to get, and a compiler or interpreter will optimize the best way of getting there. Bugs don't have quite the same meaning when you work that way. You might not get the answer you're looking for in the first shot, but it is not a matter of the whole thing crashing, which is how normal programming is. You adjust your formulas untill you get the type of results you hoped for, rather than spending most of the effort getting your steps to work at all.
Now, how about that for solving big complex problems involving many people. I'm mainly thinking about information sharing in social networks. I hold certain kinds of information and certain skills and I have certain interests. I could use certain things from other people, either generally, or in specific situations, or to solve problems that suddenly come up. Maybe I need an accountant who knows about tax issues and company formation differences between the U.S., France and England. Maybe I need the side window for a 1959 Cadillac Seville. Some things already fit into solutions that people have made. Like eBay would fit for the latter. But what if nobody has spent the resources for piecing together a specific solution. Or if I don't know where it is.
So, the thought is that I might possibly express my needs in a functional way. Well, we're talking about those things that might be put into some kind of reasonably precise expression. I'm looking for this thing, and I'll be willing to pay that much, and I want it by this time, and from somebody within this radius from my house. And that some super parallel distributed interpreter comes along and grabs my query, together with everybody else's, and optimizes them all, so as to use the least possible resources in order to get everybody approximately what they want, if it is available.
Oh, there are lots of problems in that, of course. It is merely an idea. It is not clear how one would express complex requests into an unambiguous query. And I'm not sure anybody would know how to write such an interpreter. But it reframes the problem. One task is to be able to express clearly the types of things people want, and which we expect that computers might facilitate. And the other task is to write the engine that processes all those queries and gives the best possible results. And each of those separate areas would of course be improved as people find better and better ways of doing it.
|
|