One of the problems with a lot of technical analysis is that there are conflicting definitions in different books and there is often no objective way of resolving which is "right". Who is an authority? How can we be sure that even the most respected authority did not make a mistake?
There is a third version I am aware of - Steve Nison, who says that it opens above the high of the white candle, but does not have to penetrate as much as 50%, just penetrate "significantly". However, there is much more to Nison than this, as we will see below.
I wrestled with this when I wrote the subject notes for E171 Specialised Techniques in Technical Analysis. What I settled with was basically Nison's position, because it is logical to me:
Dark Cloud Cover is a bearish reversal pattern. Thus, it only occurs after an uptrend. It requires two candle lines: a long white candle, followed by a black candle that opens near its high and above the upper shadow of the white candle and closes near its low and significantly into the range of the white candle's real body.
The more the black candle's real body covers the range of the white candle's real body, the stronger the pattern. If it more than covered the white candle's real body, it would be an engulfing pattern. So, the closer it is to an engulfing pattern the better.
Also, the closer the black candle is to a black marubozu, the stronger the pattern. It follows too, that the stronger the white candle, the more significant the pattern. So, Steve Nison suggests that two opposing marubozu would be a very strong reversal signal.
I am very much against teaching "rules". Rules are for the less intelligent to rely on, but for the more intelligent to use as a starting point for thinking and understanding. You should be able to see the way I like to think around the issues in the above quotation. It is also evident in Nison's book, which Is why I tend to rely on it most.