Most of them probably won't, but there's nothing stopping them from decoding it. Especially since people have been talking about the idea.

Yeah, it's a known method for hiding email addresses, and I've used it. But I think it only takes one or two function calls in PHP to decode it.

Graphics are a better bet, but then the address won't be clickable.

Hmmm, well if you used the url-encode method, did it work? For how long did you not receive spams to the url-encoded email address?

It wasn't my address that was published, it was two others, and they were already getting spam anyway.

If you are looking for a solution to obfuscate email address in source code, there is javascript.

Example:

// mailto:address@domain.tld
var a = new Array('in.tld','lto:','mai','dress@doma','ad');
document.write('<a href="'+a[2]+a[1]+a[4]+a[3]+a[0]+'">'+a[4]+a[3]+a[0]+'</a>')
document.write('<a href="'+a[2]+a[1]+a[4]+a[3]+a[0]+'">contact</a>')

Or an alternative:

var a = new Array('in.tld','lto:','mai','dress@doma','ad');
function mailto(ref) {location=a[2]+a[1]+a[4]+a[3]+a[0]+'?subject='+ref+'; }
<...>
<a href="javascript:mt('something')">email us</a>

Any standard web browser can take any encoded characters and any related Javascript code and convert it do an e-mail address. There's no reason to think that an e-mail harvest program would not do the same. It's not hard to do in a program.

I've used this solution on a heavily visited site, with a specific address reserved for this usage. So far no spam in this mailbox (the address can't be guessed).
Email harvest programs are not standard browsers. And so far there is no need for them to understand js. They must harvest as much email address as possible in a short time, enabling js for every page will slow the process quite a lot.