7 ms·

This reminds me of programming languages designed with an English-like syntax in the hope that people would be able to develop software without learning to code

by Zak 2y ago

This reminds me of programming languages designed with an English-like syntax in the hope that people would be able to develop software without learning to code. It also reminds me a bit of the Ercoupe, an airplane designed in 1937 with the ailerons and rudders mechanically coupled so that pilots would have fewer controls to operate and reduced ability to put the airplane into uncoordinated flight. Neither approach has remained popular into the modern era. Of course it also reminds me of Airbus, which has a similarly simplified fly-by-wire control system that abstracts away the details of controlling the aircraft under normal conditions. Airbus promised that system, introduced on the A320 would make planes uncrashable, which pilots promptly disproved on that aircraft's first passenger flight[0]. Helping people get pilot's licenses without learning to fly coordinated turns does not strike me as particularly useful, but there is no doubt room for improved avionics and planning software. [0] https://en.wikipedia.org/wiki/Air_France_Flight_296Q https://en.wikipedia.org/wiki/Air_France_Flight_296Q

jhugo 2y ago
The crash you reference was most likely caused by a pilot choosing to fly too low (they performed the flyover at 30ft instead of 100ft) and striking trees which had not been marked on charts provided to them. The pilot did try to blame the fly-by-wire system, but it wasn't very convincing.
- Zak 2y ago
  I don't think the pilots were correct to blame the system either, however I do think some of their decisions to put the aircraft in a dangerous situation were influenced by a belief that the computers would keep them out of trouble. Airbus trying to market the plane as uncrashable probably contributed.
nl 2y ago
> This reminds me of programming languages designed with an English-like syntax in the hope that people would be able to develop software without learning to code... Neither approach has remained popular into the modern era. Au Contraire! MS Office macros continues to use VBA (Visual Basic for Applications) which traces its roots to BASIC - one of those "English like syntax" languages. MS Office macros remains one of the most popular languages in the world.
- Zak 2y ago
  I wouldn't put BASIC in that family; it's probably closer to simplified FORTRAN. SQL is probably the best counterexample.
  - nl 2y ago
    BASIC is literally "Beginners' All-purpose Symbolic Instruction Code" To quote Wikipedia:"They wanted to enable students in non-scientific fields to use computers."
calmbonsai 2y ago
Oh wow. The Ercoupe. Sadly, the original design was found to be an unsafe plane due to very poor spin recovery, but mostly due to being under-powered for its airframe. It actually ended up having a higher-than-mean accident rate for the time which defeated its raison d'etre. There are still more than a few (~200) actively FAA registered today though I'm certain with engine power and control surface upgrades. It turns out that pilots DO need uncoupled yaw, pitch, and roll controls in order to maintain authority and/or recover from certain behind-power-curve conditions, but they didn't know all the edge case stuff back then. Many of these early designs, including other notables such as the Dyke Delta, sacrificed general airworthiness for a single notable flight envelope performance characteristic. In the case of the Dyke Delta, very efficient high-speed cruise, but it's far too sensitive to CG, has little practical cargo capacity, poor stall approach/recovery (think snap-stalls), very poor visibility, and longer take-off/landing runway requirements. I have zero doubt that modern fly-by-wire controls can enable a very car-like piloting experience, but I also have zero doubt that there's no way such an experience can be brought to the GA market that's remotely price-competitive with current non-fly-by-wire offerings. Additionally, the regulatory environment (at least in the U.S.) still mandates pilot proficiency without any dependency on these flight aides. To use a (very) rough analogy, though we have auto-driving features on modern automobiles, they can't be enabled when obtaining a driver's license. Like with many other human activities, you need to train and certify for the statistical tails though you operate in the fat part of the curve. The automobile market is also absolutely HUGE with (exponentially) lower upfront regulatory compliance costs that can potentially offer an return on investment to absorb the R&D of fielding a few "radical" early "luxury" models to enable the mass-manufacturing of a "base model" and achieve economies of scale. The general aviation market is TINY and its large upfront regulatory compliance costs only effectively allow for a new manufacturer to field ONE model for many, many years so that ONE model must be designed VERY conservatively and, essentially, only iterated throughout its lifecycle--never fully redesigned.
- Zak 2y ago
  > The general aviation market is TINY Airhart seems to be planning to expand that market by selling a $500K version of an airplane that normally costs about $250K. I wish them good luck in finding a viable pivot when than plan goes as well as can be expected.

by Zak 2y ago

jhugo 2y ago

The crash you reference was most likely caused by a pilot choosing to fly too low (they performed the flyover at 30ft instead of 100ft) and striking trees which had not been marked on charts provided to them. The pilot did try to blame the fly-by-wire system, but it wasn't very convincing.

Zak 2y ago
I don't think the pilots were correct to blame the system either, however I do think some of their decisions to put the aircraft in a dangerous situation were influenced by a belief that the computers would keep them out of trouble. Airbus trying to market the plane as uncrashable probably contributed.

nl 2y ago

> This reminds me of programming languages designed with an English-like syntax in the hope that people would be able to develop software without learning to code... Neither approach has remained popular into the modern era. Au Contraire! MS Office macros continues to use VBA (Visual Basic for Applications) which traces its roots to BASIC - one of those "English like syntax" languages. MS Office macros remains one of the most popular languages in the world.

Zak 2y ago
I wouldn't put BASIC in that family; it's probably closer to simplified FORTRAN. SQL is probably the best counterexample.
- nl 2y ago
  BASIC is literally "Beginners' All-purpose Symbolic Instruction Code" To quote Wikipedia:"They wanted to enable students in non-scientific fields to use computers."

calmbonsai 2y ago

Oh wow. The Ercoupe. Sadly, the original design was found to be an unsafe plane due to very poor spin recovery, but mostly due to being under-powered for its airframe. It actually ended up having a higher-than-mean accident rate for the time which defeated its raison d'etre. There are still more than a few (~200) actively FAA registered today though I'm certain with engine power and control surface upgrades. It turns out that pilots DO need uncoupled yaw, pitch, and roll controls in order to maintain authority and/or recover from certain behind-power-curve conditions, but they didn't know all the edge case stuff back then. Many of these early designs, including other notables such as the Dyke Delta, sacrificed general airworthiness for a single notable flight envelope performance characteristic. In the case of the Dyke Delta, very efficient high-speed cruise, but it's far too sensitive to CG, has little practical cargo capacity, poor stall approach/recovery (think snap-stalls), very poor visibility, and longer take-off/landing runway requirements. I have zero doubt that modern fly-by-wire controls can enable a very car-like piloting experience, but I also have zero doubt that there's no way such an experience can be brought to the GA market that's remotely price-competitive with current non-fly-by-wire offerings. Additionally, the regulatory environment (at least in the U.S.) still mandates pilot proficiency without any dependency on these flight aides. To use a (very) rough analogy, though we have auto-driving features on modern automobiles, they can't be enabled when obtaining a driver's license. Like with many other human activities, you need to train and certify for the statistical tails though you operate in the fat part of the curve. The automobile market is also absolutely HUGE with (exponentially) lower upfront regulatory compliance costs that can potentially offer an return on investment to absorb the R&D of fielding a few "radical" early "luxury" models to enable the mass-manufacturing of a "base model" and achieve economies of scale. The general aviation market is TINY and its large upfront regulatory compliance costs only effectively allow for a new manufacturer to field ONE model for many, many years so that ONE model must be designed VERY conservatively and, essentially, only iterated throughout its lifecycle--never fully redesigned.

Zak 2y ago
> The general aviation market is TINY Airhart seems to be planning to expand that market by selling a $500K version of an airplane that normally costs about $250K. I wish them good luck in finding a viable pivot when than plan goes as well as can be expected.