E&P over the years has evolved from an exciting and challenging book tightly organized, with no precalculus background, into a tiresome and hugely overweight one with numerous extra chapters on precalculus material, and large numbers of very easy, almost trivial problems. The original book had only challenging problems.
Rearranging of topics also introduced errors such as forgetting to prove the sum law for integrals after moving the topic of antidifferentiation the proof had been based on, or omitting chapter headings in some editions for topics like implicit differentiation.
Other changes include omitting the more significant of the three laws of Kepler which were originally the main application of several variable calculus.
Other changes include simply dumbing down the alngauge used, smaller words, shorter sentences, as well as the annoyingly distrcting increased use of color for the video game generation.
The changes in the problem sets make it harder to assign problems. One used to simply say do the odd problems 1-20, say, but now that will result in doinjg 10 identical and trivial problems, all about parabolas.
For years the proof of the fundamental theorem of calculus had a gap in it, from ignoring the behavior of the function at the interval endpoints, but that was finally filled in later editions.
The proof of the chain rule in early editions also gave the false impression introduced unfortunately by Hardy in his excellent book, that the traditional proof of the chain rule is lacunary, whereas the lacuna is easily filled, as was done in many good books before Hardy's.
E&P 1st edition began with a list of useful interesting applications of calculus, then plunged right into the subject and began deriving those applications. Problems were challenging and deep applications were included at the appropriate places. Later editions delayed the introduction of calculus, introduced huge numbers of more elementary topics and elementary problems, and deleted the more significant applications, and introduced pointless color illustrations and computer related projects and software, as well as web based problem solutions.
In the early editions one had to draw the pictures for the problems oneself, while in the later ones the pictures are already drawn, rendering the volume problems essentially trivial.
There are also errors in some problems, of the sort found also in the differential forms notes followed recently on this forum, in which the Riemann integrability of some unbounded functions like x^(-1/2) on [0,1], is falsely assumed because the antiderivative is continuous everywhere.
I.e. integrability is confused with antidifferentiability.
Still with all these flaws, E&P is still better than most other commonly used books, just not as good as it was itself before the changes.
The same unfortunate phenomenon befell Thomas, Cooke, and Finney in the 10th edition, and Stewart in the third edition.
