While that article does explain what set-mining is, don't trust most of what it says. There are many errors and omissions.
Set mining is primarily a concept for NL, since the implied odds are greater. There are some circumstances in which it is right in limit, e.g., if 5 or more players will see the flop, the implied odds usually let you play a low pair for set value. However, this is not worth a lot in limit. Small pairs are still marginal at best. A set has a lower chance to hold up in limit because people have the odds to chase unlikely draws, and you can't get paid off as much in NL.
Position still matters. Position makes it more likely that you can get paid off. When your opponent has a strong hand, he is more likely to commit his stack before finding out that you have a strong hand. When your opponent has a mediocre hand, he is more likely to put chips in to act strong before seeing that you are interested in the pot, or to believe you are just trying to steal if he shows weakness.
The article overestimates the chance of getting paid off. While some people are willing to fold an overpair, this isn't necessary to deprive you of the implied odds needed to make set mining profitable. You lose a lot of value when someone with QQ folds on a K-high board, or when a player with AK unimproved folds.
You can still lose after you flop a set. You lose about 20% of the time against a higher pair. This comes from the possibility that your opponent flopped a higher set ~8%, plus the possibility of spiking 2 outs twice ~ 8%, plus the possibility of hitting a one card straight or flush ~4%. If you get your stack in as an 80% favorite, you only average a gain of 60% of your stack (80% gain of 1 stack - 20% loss of 1 stack), not 100%. If you are sure you are up against an overpair (usually a bad assumption), then you need 60% of your stack to be 1/8.5 times your preflop call. You would break even if it is about 1/14 of your call.
Set mining can be fun. However, for to to be profitable, you either need to find some value from the times that you miss, or you need a lot more implied odds from the times that you hit than just 10:1.