Yeah, they also allude to "better ways to do it" both in the main text and the last paragraph:
> The next post will discuss better ways to implement higher precision in numerical calculations.
There are lots of automatic row / column swap techniques, particularly for sparse matrices that ensure you avoid catastrophic cancellation. You wouldn't do this automatically, because most matrices are reasonably conditioned (and figuring out the optimal swaps is NP complete and so on), but there are lots of algorithms to do well enough when you need to.
