While the article claims there are no strength benefits to dressing a knot, it is almost certainly still a good idea if only for the fact that it is easier to recognize a mistake in a knot if it is dressed than if it is not.
Where "better" means less likely to fail - in many cases, it could mean "quick" or "easy to teach" for example, which splices are not. Given the tremendous strength of much modern synthetic cordage (especially Spectra / Dyneema), quick-to-tie knots that weaken the line, even by a significant percentage, may be preferable.
For anyone interested in knots and knot tying, the Ashley Book of Knots[1] is a wonderful book to have. I've spent many happy hours pouring over the knots and their descriptions in this book.
It is _the_ book on knots, although it does miss knots invented later in the 20th century, like the Zeppelin bend. I was amazed to learn that it was used as a reference book by mathematical topologists. They will refer to a bend in space as Ashley#1024 for example, since all the knots in the book are numbered in a referential fashion.
