FWIW, I have this book:
http://www.amazon.com/John-Standefers-Picture-Perfect-Guitar-Manuscript/dp/0786669365

Bought directly from the author when he came to our church to teach a guitar workshop that I arranged.

Take note of what is on the front cover. I wish it was still in print, because it's the best example of why to use it all, standard notation, TAB, chords, and a great hybrid method that John teaches.

So far I haven't pointed out any 'savant-like' persons.

John could qualify but he also is an incredibly deep student and teacher of all means of playing and transcribing, as it pertains to guitar.

Most of the folks that denigrate TAB think of it as 'lesser' than standard notation. This is where it goes off the rails because, again, for an alternately tuned stringed instrument, TAB becomes much more valuable particularly when one is aiming for a certain 'voice' of the instrument and song-where open strings should be open strings, etc. simply because they have different harmonic content than the same note played on a different string at a certain fret.

This is entirely different than keyboard instruments where for all intents and purposes, you have one place to play a certain note and that is it - no other choices.

Such is not the case on multi-stringed instruments as Bobc points out above. The choices of where to play the right notes of the multiple possibilities means all the world in the nuance and timbre of any stringed instrument, open-stringed harps excepted. Over time, one develops an ear for imagining where to play the right stuff and even if a capo is being used or if an alternate tuning is being used.

When I play electric guitar, I'm typically learning 3-4 new songs a week, where a big portion of the requirement is getting the 'sound' right in addition to playing the right notes. Very little lead work is required because it's for a modern rock worship band - where the popular thing these days is using at least one of the electrics for pads and shimmer; not unlike synth patch programming to be honest. Solos and leads aren't really part of the

John's TAB method is what I use when I have to write notes for myself on creating these vibe-things, when all I am given is a set of lyrics, a key, a BPM and sometimes chords that are accurate, many times that are not. Standard notation falls short for this and many other purposes on stringed instruments.

I hate that Hal Leonard quit publishing the book, because it's so handy - and it encourages knowing both standard notation and taking advantage of TAB if one uses the full potential of what is presented on the pages.

This conversation feels like when our local school started teaching our kids 'Chicago Math' - and it wasn't referring to elections!

One of the concepts there is that different people learn information in different ways, but can all reach the same correct outcome.

The most shining example of this was the re-introduction of use of what is called 'Lattice Multiplication' - which actually has it's heritage centuries before the multiplication method that most of us learned in school. This was taught along-side the method we are most familiar with.

Here's a video from University of Chicago showing an example of 38*57. http://everydaymath.uchicago.edu/teaching-topics/computation/mult-lattice/

It was really interesting to observe my fellow parents' reactions as they taught this. A few were outraged. Most were perplexed. I was transfixed. I saw a beautiful method presented for a guy with pretty crummy handwriting that always got his columns wonky and ended up summing things incorrectly. I knew my multiplication tables, but what I was bad at was handwriting, not the understanding of multiplication.

One the kids have demonstrated that they know the tables, the teachers didn't care what method they used, traditional or lattice.

Similar story with long division. Teach multiple different ways to get to the correct outcome so that students have a greater opportunity for success depending on which method 'clicks' with them. They introduced a method called 'the forgiving method' which encourages getting good at estimation - probably one of the most critical skills any of us deal with when budgeting time, finances, etc.

The UofC website calls this the 'Partial Quotients' method. Again, I saw a beautiful way to do long-division which the way it encourages is patience and recognizing waste. The example shown here is a simple one:

http://everydaymath.uchicago.edu/teaching-topics/computation/div-part-quot/

But note that after doing this partial quotient method for awhile, one would begin to recognize that instead of estimating 10 the first time around, 20 would have been a better starting point for the estimation. Estimation upholding multiplication fact learning.

TAB supporting notation, or vice versa. Try thinking of it that way. You can also think of it as notation. There is no harm there. Probably all of the keyboard first learners poo-pooing TAB had Schaum or Thompson, or Fingerpower, or other primers which indicated which fingers (indicated by number by jove!) were best under the standard notation, particularly for multi-octave runs.

TAB is no different, for 'multi-axis' instruments that is not 'single axis' or purely linear like a piano or other keyboard. Unlike keyboard technique, the benefits of deep knowledge of fingering guides continues to help technique rather than be relegated to primer level activity.

Just ask John Standefer. 2002 winner of Winfield Fingerstyle Competition, and instructor at the Chet Atkins Appreciation Society.