Is a Self-Modifying Function Bad?

Question

If I want a function to do something different the first time it's run, I could check each time it's run to determine whether it's the first time, or I could change the function; something like this:

foo=function(a,b){
    ...do something1...
    foo=_foo;
}
_foo=function(a,b){
    ...do something2...
}

Is this bad; if so, why? Is there a better way to do this? I'm specifically looking to implement this in javascript, though other language points will be considered.

It really depends on how you want to modify the function. Often, this is used for caching, so it's just an performance-enhancement that the user of the function shouldn't notice. If that's the case, then it's fine. If not, you might need to elaborate on your specific case. — kba, Jan 06 '14 at 19:47
I usually use a static variable (property) as a flag to handle "first time run" scenarios — CrayonViolent, Jan 06 '14 at 19:47
Almost always bad. Back in the day we'd write self-modifying code for resource-constrained systems, of which your browser is not one. Even memoizing isn't really a great excuse, IMO, but that's not a technical debate. — Dave Newton, Jan 06 '14 at 19:48
functions are objects. you'd be better off having one function, with a simple `if (!foo.hasrun) { foo.hasrun = true; first time code; } else { next run code }` type thing going. — Marc B, Jan 06 '14 at 19:48

score 1 · Accepted Answer · answered Jan 06 '14 at 20:08

This isn't a "self-modifying" function in the strict sense - you don't modify any code on the fly (and that would be bad), just assign another value to a function name. This is fine in most cases and makes polymorphic code quite simple and elegant. To make the intent cleaner, you can factor out the "first time" code into a separate function:

function firstTimeFoo() {
    .....
    foo = normalFoo
}

function normalFoo() {
    ...
}

foo = firstTimeFoo

Is a Self-Modifying Function Bad?

1 Answers1