'In the Blink of an Eye' - Walter Murch

In this book, Murch talks in depth about the nature of editing, why it works on a psychological and biological level, and how the notion of discontinuity of film was first discovered. He uses a number of different analogies to explain his theories on editing, such as the moving of beehives, the differences in DNA between chimps and humans and the similarities between overactive editors and tour guides. Some of the points, such as his theories on blinking, he reiterates in several other interviews and talks.

Murch goes on to identify his top six reasons where you should cut, which he arranges into a hierarchy. In the second half of the book he describes his experiences working with mechanical and electronic editing equipment, and his predictions for the development of editing (and cinema) in the 'future' - of which most of what he describes has taken place. Below I've summarized the main points he raises in the book:

• Ride of the Valkyries sequence in 'Apocalypse Now' shot as documentary rather than composed shots.
• Cutting a film is not so much putting together but 'finding a path'
• Cuts work even though its a total displacement of time and space for the viewer - and we're used to our day being perceived as a long continuous 'shot'.
• We have difficulty accepting cuts that are neither subtle nor total - such as cutting from a master shot to a slightly tighter shot. There's not enough change in motion or context to avoid it being jarring.
• Films being shot discontinuously was the filmic equivalent of discovering flight - no longer are films earthbound to time and space.
• Film editing is 'cutting out the bad bits' essentially - like making a home movie but much more complex and subjective.
• More audio tracks does not mean a better sound mix. 
• 'Try to do the most with the least' - suggestion is much more effective than exposition. Past a certain point audience becomes spectators and not participants.
• It takes more work to decide when not to cut.
• Overactive editor is like an annoying tour guide - need to let the audience decide what they want to see, and be confident in giving them that freedom.
• Cutting to preserve 3D space is at the bottom of priority list for a good cut.
• Need to be willing to sacrifice other 'rules' to preserve emotion. Satisfying emotion and moving the story along obscures issues lower in the hierarchy.
• The editor is free from the context of the shot - and so has a unique perspective when working with the director.
• Editor has to propose alternate situations to tease out director's vision. Director might not know what they want but they know what they don't want.
• Linear vs Random-access (Non-linear) editing - can discover new ideas whilst having to look through the material in linear process. Too quick to 'NG' a shot when it may be usable later on.
• Re-cut scenes until you can't see yourself or your emotions in it - let the film create itself.
• Test screenings are useful - ask the opinion of audience a couple of days later when they're not skewed by initial emotion.
• Every shot has several cut points - staying on a character or leaving them before they've finished their line has different effects on audience. 
• History of linear/non-linear mechanical/electronic editing formats - and the pro's and con's of each.
• Effects digital editing has had on actual content - are there more cuts now because we can?
• More 'vertical editing' in the future - editing not just after the frame but within the frame, cropping, deleting, masking etc.
• Essence of cinema is 'Lets go out' to watch a movie - dissatisfaction with ordinary surroundings. Can never have true cinematic experience at home no matter how good the equipment is. Will always be watching films in the dark.