I learned 7...e5! to refute the Ryder when I was first learning the BDG. This second pawn gambit is only interesting because the trap in it is very tactically instructive. My opinion is: learn the trap, learn one of the refutations, and move on.
Taking the second pawn and refuting in this manner is the best answer for Black, I think. Even so, the game above makes it look much easier than it really is. Nf3 and Bd3 deserve question marks, as I somewhat doubt those are the right squares. Playing Qg4 and then e5 requires cold blood. A less risky, more positional approach exists for the more skittish players.
There've been a few notes posted recently regarding the Ryder Gambit, asking and opining about it soundness or lack thereof. I'd thought I'd offer a little fodder for the discussion.
Diemer's own book, Das moderne Blackmar-Diemer Gambit, first arrived in 1957, though my copy is a later edition in 1983. This book is essentially a book on the Ryder Gambit. It's in German, as is appropriate, but it's easy enough to follow and a cheap German-English dictionary allowed me to understand most of the valuable commentary. The games and game pieces are fun to go through, though I doubt the evaluations and analysis are all accurate - no computers to check things with back then.
Most other texts on the BDG cover the Ryder, if just to convince you that overall it's bad. The most complete text I have found on that particular gambit is Schiller and Crayton's volume that dates from 1995. Their conclusion is pretty much the same as all the rest: if Black makes a mistake, White can often crush Black, but if Black plays the best line(s), White is simply two pawns down or worse.
I will offer the critical line here - at least that claimed critical by Schiller and Crayton.
The embedded game is considered the stem game for Black's defeat of White in the Ryder. Of course, the Devil may still be in the details. Just thought I'd like to offer this up for your enjoyment.
Long live the BDG! (but maybe not the Ryder variation)