Interesting article. I've come to find that articles praising one vendor's products over another's wiritten by the vendor are best taken with a pinch of salt. I read Oracle comparisons by Oracle, SQL Server comparisons by Microsoft and DB2 comparisons by IBM skeptically. Not to say that they lie, but they are written in such a way to make the competition look less appealing.
Selecting a database for production useI would imagine is best done on a case by case basis depending on several factors (cost, support, skill set, features, operating system, problem domain etc) which would mean for different scenarios you'd end up with different databases.