Actually, it's hard for a 9-character password to beat a 12-character password even though the latter has a larger alphabet/key-space (unless I've completely blown the analysis below, which was done before coffee :).
The first has a key-space of 36^12 (36 possible characters in each of 12 positions), or about 4.7e18. The second has a key-space of 62^9 (62 upper/lower case letters and digits in each of 9 possible locations), or about 1.4e16.
If, in addition to adding the uppercase letters, you added the possibility of needing to test symbols, such as ~`.,/:;!@#$%^&*-=_+ (another 19 symbols), and changed the latter password to "tlpW#NT2m", then the searchable key-space for all 9-character passwords becomes 81^9, or about 1.5e17.
RE-EDIT: Sorry. I should have read the article first. I'm not sure why the latter would be more secure. Obviously "WENT" would be in a dictionary, so I'd think that "tlpWENT2m" would fall to a combinator attack very quickly, too.
The first has a key-space of 36^12 (36 possible characters in each of 12 positions), or about 4.7e18. The second has a key-space of 62^9 (62 upper/lower case letters and digits in each of 9 possible locations), or about 1.4e16.
If, in addition to adding the uppercase letters, you added the possibility of needing to test symbols, such as ~`.,/:;!@#$%^&*-=_+ (another 19 symbols), and changed the latter password to "tlpW#NT2m", then the searchable key-space for all 9-character passwords becomes 81^9, or about 1.5e17.
RE-EDIT: Sorry. I should have read the article first. I'm not sure why the latter would be more secure. Obviously "WENT" would be in a dictionary, so I'd think that "tlpWENT2m" would fall to a combinator attack very quickly, too.