Well, the winning iPod health warning was really good (referencing the iPod advertising silhouette graphics, right?). The code snippet looks monstrous, but I'm too stupid to get the whole implication.

Reasons why it's a bad piece of code:
1) The function is repeated wherever it's needed instead of only being in one place.
2) It doesn't work properly ("12345678P~?" will still validate as a valid SSN)
3) Code for testing each section is repeated. (It should be in a loop)

Even without the *evil* regexp, you should be able to shrink this down to 10 or so lines of code)

in VDF...(My language)
function isValidSSN string sSSN returns boolean
    boolean isValid
    string sChar 
    
    move (true) to isValid
    if (length(sSSN)) ne 8 move (false) to isValid

    for iIndex from 1 to 8
        move (mid(sSSN,1,iIndex)) to sChar
        if ((sChar"9")) move (false) to isValid
    loop
    function_return isValid
end_function

Off topic. It appears that between preview and post, metachat strips out &nbsp's and replaces them with spaces.

Plus it doesn't play nice with gt and lt characters. My Code got mangled.

For a test try posting something containing the phrase  

Damn you code stripper box....That should read...
For a test try posting something containing the phrase &nbsp

Damn it. I've added that bug to the bug list.

I don't really know Java and regexes are still fairly new to me but I guess this would be fairly close to what the coder intended:
ssn.matches("[0-9]{9}");
Still, as they say in that thread, this still wouldn't be the correct way to validate SSNs but it does go to show just how bloated the code is. Nightmarish.

"matches" hah. You young 'ens have got it easy.
I remember when you only had four variables, and you could only compare strings one byte at a time. We 'ad to program uphill both ways back then.

You were lucky!

confirming that my bugfix has worked:
<testme>
 insert text here
</testme>