cdennyb, the first question Perlman answers is illustrative of the violinist dilemma:  to achieve accuracy (presumably intonation) slow practice and lots of repetition of atomic elements.  Here is the first exercise from Sevcik Op 1 book 1:

Open A, B, C, B, Repeat

Lets say I do this now as follows:  The first stopped note B is out of tune.  Its wrong to correct by moving the finger and then play the next note.  Its wrong to simply ignore it and play the C too.  So you have to try again and if you land the finger correctly in tune, you have to repeat it several times (like about 5 at least) to statistically reinforce learning. I bet 98% of learners don't even recognize this obvious fact.

THIS IS EXTREMELY LABORIOUS. 

Sure, its easy for Perlman to say 5 hours is the limit and there is no benefit beyond that.  Thats because his solution to the precise measurement problem is hard labor:  repetition hits of fingers being dropped from the air.  How can you precisely measure that way in the first place?  You would think practicing 1 and 2 finger( double stop) glissando scales would at least allow the beginner to accurately measure the changing distances between semitones without guessing which is what fingering at the start amounts to.  I couldn't imagine pounding away at the fingerboard with the exhaustive Sevcik more than 5 hours a day due to physical fatigue.  Perlmans world view is from the perspective of learning the violin at 6 years of age who has all the time in the world.

Survive by doing Doublestop glissando scales with rhythmic variations...and tone work.  The finger is always down on the fingerboard, no dropping allowed to start off with while you learn intervals.  This is my idea of an advanced player:  pure intonation, good rhythm, and pleasant tone.  Nobody can do:

B (wrong), B (correct), B, B, B,B,B (rep hits) - C (wrong), C (wrong again), C (correct), C, C, C, C, C...etc etc etc ...then you put your left hand down for some reason, and then start again with a slightly different thumb position, hand angle....thus inducing new variables and finding out you can't hit B or C accurately.

this kind of repetition drudgery for any longer than 5 hours a day.