Whoa, reading the last few replies, I already get the general concept of what merged mining is... and I'd never heard of it before reply # 285 where it all made sense:
hashing -> found result goes to (user-run) bitcoin pool server -> credit in your bitcoin mining account
Merged...
hashing -> found result -> bitcoin pool -> credit in your bitcoin account
^---> other secondary network pool -> credit in your other pool account
Since the two networks don't communicate or have any relation to one other (aside from protocol), the hashes from one network can be used in the other as well!
... And the results could be economically disasterous. O_O
If hash rate of the Other network (namecoin, in this case, which I don't care much for) increases as more people do this "merged mining" thing with hardware used for the high difficulty of Bitcoin, the difficulty of Other Coin's network will increase to adapt to the rate as well.
The other side of the matter is that people are essentially "double spending" their production - spending the same amount of resources hashing for one network, then "double spending" that redemption into other currencies, devaluing both of them in the process (if they spend it).
Will it work? Yes. Will it work really well? Yes. Easy to implement. Yes. Will people get more "money" out of it? Yes yes yes yes yes... is it something I'd do? Sure. Is it a "pure idea" with no "oh god, how long can we keep dumping nuclear waste into this small pond" behind it? Eh... nah. In other words, on the scale of "white and pure" to "dark and evil", it's somewhere light-grey.