Paul Graham - Essays

And yet some applications will still demand speed. Some of the problems we want to solve with computers are created by computers; for example, the rate at which you have to process video images depends on the rate at which another computer can generate them. And there is another class of problems which inherently have an unlimited capacity to soak up cycles: image rendering, cryptography, simulations.

If some applications can be increasingly inefficient while others continue to demand all the speed the hardware can deliver, faster computers will mean that languages have to cover an ever wider range of efficiencies. We've seen this happening already. Current implementations of some popular new languages are shockingly wasteful by the standards of previous decades.

This isn't just something that happens with programming languages. It's a general historical trend. As technologies improve, each generation can do things that the previous generation would have considered wasteful. People thirty years ago would be astonished at how casually we make long distance phone calls. People a hundred years ago would be even more astonished that a package would one day travel from Boston to New York via Memphis.

I can already tell you what's going to happen to all those extra cycles that faster hardware is going to give us in the next hundred years. They're nearly all going to be wasted.

I learned to program when computer power was scarce. I can remember taking all the spaces out of my Basic programs so they would fit into the memory of a 4K TRS-80. The thought of all this stupendously inefficient software burning up cycles doing the same thing over and over seems kind of gross to me. But I think my intuitions here are wrong. I'm like someone who grew up poor, and can't bear to spend money even for something important, like going to the doctor.

Some kinds of waste really are disgusting. SUVs, for example, would arguably be gross even if they ran on a fuel which would never run out and generated no pollution. SUVs are gross because they're the solution to a gross problem. (How to make minivans look more masculine.) But not all waste is bad. Now that we have the infrastructure to support it, counting the minutes of your long-distance calls starts to seem niggling. If you have the resources, it's more elegant to think of all phone calls as one kind of thing, no matter where the other person is.

There's good waste, and bad waste. I'm interested in good waste-- the kind where, by spending more, we can get simpler designs. How will we take advantage of the opportunities to waste cycles that we'll get from new, faster hardware?

The desire for speed is so deeply engrained in us, with our puny computers, that it will take a conscious effort to overcome it. In language design, we should be consciously seeking out situations where we can trade efficiency for even the smallest increase in convenience.

Most data structures exist because of speed. For example, many languages today have both strings and lists. Semantically, strings are more or less a subset of lists in which the elements are characters. So why do you need a separate data type? You don't, really. Strings only exist for efficiency. But it's lame to clutter up the semantics of the language with hacks to make programs run faster. Having strings in a language seems to be a case of premature optimization.

If we think of the core of a language as a set of axioms, surely it's gross to have additional axioms that add no expressive power, simply for the sake of efficiency. Efficiency is important, but I don't think that's the right way to get it.

The right way to solve that problem, I think, is to separate the meaning of a program from the implementation details. Instead of having both lists and strings, have just lists, with some way to give the compiler optimization advice that will allow it to lay out strings as contiguous bytes if necessary.

Since speed doesn't matter in most of a program, you won't ordinarily need to bother with this sort of micromanagement. This will be more and more true as computers get faster.

Saying less about implementation should also make programs more flexible. Specifications change while a program is being written, and this is not only inevitable, but desirable.

The word "essay" comes from the French verb "essayer", which means "to try". An essay, in the original sense, is something you write to try to figure something out. This happens in software too. I think some of the best programs were essays, in the sense that the authors didn't know when they started exactly what they were trying to write.

Lisp hackers already know about the value of being flexible with data structures. We tend to write the first version of a program so that it does everything with lists. These initial versions can be so shockingly inefficient that it takes a conscious effort not to think about what they're doing, just as, for me at least, eating a steak requires a conscious effort not to think where it came from.

What programmers in a hundred years will be looking for, most of all, is a language where you can throw together an unbelievably inefficient version 1 of a program with the least possible effort. At least, that's how we'd describe it in present-day terms. What they'll say is that they want a language that's easy to program in.

Inefficient software isn't gross. What's gross is a language that makes programmers do needless work. Wasting programmer time is the true inefficiency, not wasting machine time. This will become ever more clear as computers get faster.

I think getting rid of strings is already something we could bear to think about. We did it in Arc, and it seems to be a win; some operations that would be awkward to describe as regular expressions can be described easily as recursive functions.

How far will this flattening of data structures go? I can think of possibilities that shock even me, with my conscientiously broadened mind. Will we get rid of arrays, for example? After all, they're just a subset of hash tables where the keys are vectors of integers. Will we replace hash tables themselves with lists?

There are more shocking prospects even than that. The Lisp that McCarthy described in 1960, for example, didn't have numbers. Logically, you don't need to have a separate notion of numbers, because you can represent them as lists: the integer n could be represented as a list of n elements. You can do math this way. It's just unbearably inefficient.

No one actually proposed implementing numbers as lists in practice. In fact, McCarthy's 1960 paper was not, at the time, intended to be implemented at all. It was a theoretical exercise, an attempt to create a more elegant alternative to the Turing Machine. When someone did, unexpectedly, take this paper and translate it into a working Lisp interpreter, numbers certainly weren't represented as lists; they were represented in binary, as in every other language.

Could a programming language go so far as to get rid of numbers as a fundamental data type? I ask this not so much as a serious question as as a way to play chicken with the future. It's like the hypothetical case of an irresistible force meeting an immovable object-- here, an unimaginably inefficient implementation meeting unimaginably great resources. I don't see why not. The future is pretty long. If there's something we can do to decrease the number of axioms in the core language, that would seem to be the side to bet on as t approaches infinity. If the idea still seems unbearable in a hundred years, maybe it won't in a thousand.

Just to be clear about this, I'm not proposing that all numerical calculations would actually be carried out using lists. I'm proposing that the core language, prior to any additional notations about implementation, be defined this way. In practice any program that wanted to do any amount of math would probably represent numbers in binary, but this would be an optimization, not part of the core language semantics.