There's a quotation by Alan J. Perlis that says:

There are two ways to write error-free programs; only the third one works.

I recently heard this quote from my friend, and was unable to understand the deeper meaning behind it.

What is Perlis talking about here?

  • 1
    you do realize the fallacy in this parody as well, as it is possible to write non-trivial programs that are error free, it just takes discipline. – user7519 Dec 5 '11 at 19:22
  • 1
    This type of question is now being discussed on our meta-discussion site. – user8 Dec 5 '11 at 20:20
  • 1
    recommended reading: Discuss this ${blog} – gnat Feb 20 '14 at 13:05

11 Answers 11


It means there are really no error-free programs. A profound quote about ways to avoid errors with an error itself is parody.

| improve this answer | |
  • 3
    Alan Perlis certainly had a way with words. – Frank Shearar Oct 5 '10 at 15:42
  • 2
    It's the "parody" that is important in this quotes meaning. – Adam Harte Oct 5 '10 at 19:18

There is no third way.

There is no way to write error-free programs

| improve this answer | |

I'll answer with another quote...

A strange game. The only winning move is not to play.


| improve this answer | |
  • 5
    +1 for war games reference! – Jason Oct 5 '10 at 19:14
  • Obligatory xkcd link following: xkcd.com/601 – Baelnorn Oct 6 '10 at 19:59

As many other answers have already pointed out, there is no way to write an error-free program.

But what I'd like to point out is the quote's potential meta nature. It's essentially an out of bounds error. In the first statement, he defines the universe or "list" having only two possibilities or elements. Yet in the second statement, he makes reference to a third. Which is absurd! Illegal even! A third element given a two element boundary is itself an error.

Truly profound in that the quote is able to demonstrate the very essence to which it is referring.

| improve this answer | |
  • There is a way to prove that a program behaves as specified. That is used e.g. for nuclear facilities... – user1249 Nov 21 '10 at 17:00
  • 1
    @Thorbjørn Ravn Andersen, as specified does not mean it's error-free. – CaffGeek Dec 5 '11 at 20:08

This means that all non-trivial programs will have bugs. It's just a funny way of saying there's no way to write an error-free program.

| improve this answer | |

It s possible to write error-free programs, even non-trivial ones and even prove them correct. Consider for example, languages like Coq, Epigram or Agda where this is done.

The halting problem states that it's not possible to do this for the general program.

| improve this answer | |
  • Go back farther, to Don Good's team at UT Austin, and their work in the 1970s and early 1980s with the Gypsy Verification Environment. They demonstrated that error-free code was possible, by delivering a proven error-free Message Flow Modulator to the Navy. The acceptance test suite was developed by a completely different group. When the MFM saw the acceptance test suite, for the first time, at the Acceptance Test, it passed, with NO deviations, waivers, or "yeah but"s. – John R. Strohm Oct 6 '10 at 14:43

This reminds me of a nerd shirt I saw: There are 10 types of people in the world. Those who know binary and those who don't.

It could also be a play on the fact that sometimes lists are 0 indexed. $var = array('First','Second','Third'); And you can access this list as such: $var[0] = 'First' $var[1] = 'Second' $var[2] = 'Third'

So the literal array index 2 points to the "Third" index.

| improve this answer | |
  • ... and those who start indexing at zero – user1249 Nov 21 '10 at 17:00

This is already explained in other words, but not as clearly as I think it should be. It simply means you will try both ways, they will have errors, and finally you will fix your bugs and have an error-free program. Compare with another quote:

The only way for errors to occur in a program is by being put there by the author. No other mechanisms are known. Programs can't acquire bugs by sitting around with other buggy programs. --Harlan Mills

(Alternatively, you could read this as Pierre said (which I think is a stretch). (The third way, which does not exist in the domain, works.) Like I said, it's a a stretch, but true.

| improve this answer | |

This is the same quote my dad use to tell me when I make excuses. The saying tends to go like : "There are 3 sides to a story. Their side, Your side, and the right/true/correct side".

Putting this into context with development (and being a software tester by prof. ), I would say since there are so many ways to code something it would make sense to go with "There's 3 sides to coding. Your code, Their code, and the Refactored code."

I think this is because programmers/developers tend to refactor once the product is getting stable which is mostly too late, but most of the time the refactor is done to improve something that you and buddy did not do so well in the first place.

Hope this helps.

| improve this answer | |

I think, technically speaking, that you could write a error free non-trivial program, but due to the Halting Problem it’s impossible prove that it’s error free. So, one must work under the assumption that all programs have bugs since it’s impossible to prove otherwise.


Update: You can prove a particular algorithm will return the right answers, but that’s not the same thing as proving it’s totally correct. http://en.wikipedia.org/wiki/Correctness_(computer_science)

However, my point was that the quote is referring to the fact that one must assume a program always has bugs and trying to explain why that is the case. http://en.wikipedia.org/wiki/Software_bug#Bug_management

| improve this answer | |
  • 2
    as Tony Morris said, it is possible to prove that a particular program is correct. It is not possible to write a program that can in general prove that any program that is correct, is correct. – Max Strini Oct 5 '10 at 21:43

As additional insight, the "two ways" might be a reference to this quote by Tony Hoare:

There are two ways of constructing a software design: One way is to make it so simple that there are obviously no deficiencies, and the other way is to make it so complicated that there are no obvious deficiencies. The first method is far more difficult. It demands the same skill, devotion, insight, and even inspiration as the discovery of the simple physical laws which underlie the complex phenomena of nature.

Meditate on that a little and you'll see he's saying the same thing: if your piece of software is non-trivial, it has bugs (but complicate it enough and they won't be obvious bugs).

| improve this answer | |
  • this does not answer the question asked – gnat Feb 21 '14 at 5:49
  • @gnat I don't see how it doesn't - it's right there in the second parragraph. Perhaps the wording wasn't clear, but when I said "saying the same thing", I meant "saying the same thing as Alan Perlis". That is, Perlis's quote is likely a humorous parody of Hoare's. – Doval Feb 21 '14 at 15:34

Not the answer you're looking for? Browse other questions tagged or ask your own question.