However, many clients do have data that is woth protecting. Take a look at the link I posted above.
That is not protecting "the code" that is protecting "the feature table".
A feature table can, of course, be programmed into the actual code of the chip and would then depend on the lock bits, this, in my opinion is the wrong approach.
A feature table should be encrypted and quadruple checksummed and have a "chip match code" to avoid copying a full features table into another system.
This would be the right approach for a features table, not lock bit dependency or such.
The article seems to blame technology when the problem is that the approach is wrong.

Erik