John Deakin is certainly an experienced pilot, and he's paid well for his writings, much of which is right on target.
Not this one though.
He does create some insight to the difference between certification flights by test pilots versus everyday flying by others, ...but his suggested techniques aren't really very useful at all for several reasons.
We all rely upon the certification data for predicting our ability to clear the obstacle or get airborne within the takeoff run distance available. To toss in imaginary differences in everyday operations using ad hoc techniques ...is equivalent to tossing out all the valuable test-data obtained by the certification process. I.E., the only way to rely upon the certification data is to duplicate the recommended techniques.
This is supported by his story about the Sacramento Ice Cream Parlor crash.
First, he denounces the FAA comment: "Remember that an attempt to pull the airplane off the ground prematurely, or to climb too steeply, may cause the airplane to settle back to the runway or into the obstacles."
...Then he states "Not exactly. This is true only if the performance is marginal, as at high elevations and density altitudes, or with very underpowered airplanes, which is pretty much the same thing. "
...then he supports his denouncement
by using examples which exactly prove the FAA comment. He uses a 172 in Colorado that cannot continue to climb due to density altitude, and a Sabre jet which tries to fly by rotating prematurely thereby creating excess drag which in-turn prevents acceleration to flying speed...
...exactly what the FAA comment implies.
He denounces takeoff configurations which begin the roll with unnecessary drag...then recommends full flaps application for the shortest possible takeoff rolls.
In another contradiction, he seems to endorse the idea of "popping" flaps immediately prior to the end of the takeoff roll and/or obstacle. How does that compare with the full flaps takeoff he recommended earlier? And what about the excess speed (which equates to unobtained/lost altitude) using that technique?
In my opinion, the reason many pilots like the flap-popping technique is because it reduces the amount of time exposed to the emotional concern of "will this work?" (The relief of excess speed being available to hop-up over the obstacle is more reassuring than the Vx struggle to do the same....completely overlooking the fact that arriving at the obstacle with excess speed means the obstacle is cleared with less altitude to spare.)
He criticizes FAA for recommending a "safe altitude"...questioning whatever that definition might be...then states unequivocally " ..a "safe" altitude is anything from about ten feet above the runway on up. ..." Huh? I thought we were discussing obstacle-clearance.
He questions the technique of a taildragger using a "tail low" technique for shortest take-off run... then shifts to using only examples of tri-cycle gear airplanes for proving his points.
He demeans the FAA recommendation of accelerating to a safe "flaps up" speed, then retracting the gearr... by scolding the "dog"... and then he recommends virtually the same technique several paragraphs later : "...at least until some speed is gained, the flaps can be eased up to a "high lift, low drag" position, and the gear retracted."
This article is good for inspiration to THINK about the subject. But it offers nothing new, or particularly good, for the everyday operations by line pilots. IMO.
In my B-model,
Soft field/Short field:
1-Keep rolling. (make a properly-leaned, power-on U-turn if it helps) 2- Use takeoff flaps and full power as recommended by the OM and documented by the AFM. 3- Let the tailwheel come up naturally (i.e. use nuetral elevator until it lifts the little wheel clear of the mud/grass, then keep it there) until the airplane lifts up/out of the soft-field. 4-Accelerate to Vx or Vy (as obstacles require) to a "safe altitude", retract takeoff flaps while accelerating to your preferred climb speed, but only after close-in obstacles are cleared.