Books are possibly valid references but:
- They are quite possibly not primary sources. We like verifiable primary sources;
- They need to be cited appropriately;
Simply pointing to a book and saying it supports an answer is not enough for this site.
The answer you cite is a good example of why we want properly cited books and verifiable references. It contradicts basic scientific results and it contains numerous fallacious arguments.
In the answer you are referring to — which I will delete as it has been challenged and it has not been fixed in one year — there are five paragraphs.
The first three simply point to the book (but not to a specific point in the book) and dismiss 70 years of nutritional science with
Mr. Taubes points out some fundamental flaws in the scientific basis of the study, as well as how the conclusions became dogma. He also criticizes many studies for implying causation, when the evidence only supports correlation.
So the first part is a faulty argument ad verecundiam, which points to a possibly flawed argument by Mr. Tabues: even if many studies are wrong, that fact alone doesn't prove his point is correct.
The fourth paragraph is an anecdote, which we explicitly discourage. Furthermore it gives potentially dangerous nutritional advice:
I personally started low carb (and paleo for a brief time) because of this book. It's been two years, and I have lost...
There are literally hundreds of stories of people doing low-carb/paleo and curing...
Therefore it's a flawed hasty generalisation followed by an argument ad populum.
The last paragraph
Once you realize that there really was no scientific basis for the current "saturated fat causes heart disease" hypothesis, it's easier to be open to the contradicting studies that show how we really should be eating.
is the "science is close-minded" argument, which is an ad hominem.