All content on this blog from Tim McGhee has moved to the Tim McGhee Substack, and soon, Lord willing, will be found only on that Substack.

Thursday, August 1, 2019

Best- and worst-case scenarios for Boeing

This being a new month, I checked in on transportation policy news, among other things, and noticed that the groundings of the 737 MAX remain front-and-center.

I looked up some of the recent news that made headlines in July, and Tweeted several quotes from those articles. I had a couple questions none of the articles answered for me. The first: How long would a full-scale major design classification certification process have taken?

In looking that up, I instead found a 105-page guide to the process. It does not include an approval process timeline estimate. It also doesn't say anything at all about software which is striking considering (a) how much planes are controlled by software today, and (b) the two fatal 737 MAX crashes were caused by software bugs.

Given that Boeing didn't provide and the FAA didn't have anything on the Maneuvering Characteristics Augmentation System (MCAS) flight control system, one wonders how much software evaluation is part of the approval process prior to the 737 MAX crashes. What makes this so staggering is that pilots have no physically direct control of anything on jet airplanes today; software sits between every control and every part of the mechanics it controls.

Perhaps this is why the FAA so vigorously defends “designees” from manufacturers to have expertise on how planes work. An amazingly necessary question we must ask is, How much about how planes mechanically work today does the FAA itself know and understand? Shifting from designees having expertise and FAA-sanctioned authority to not having FAA-authority and only providing it to the actual independent authority is in order.

A worst-case scenario has two parts:

1. The Boeing 737 MAX entire approval status is sent back to square one and considered a major design classification certification. That could easily send this grounding into next year, and for however much time would remain in a similar certification process.

2. The FAA determines its aircraft approval process is woefully inadequate without a far more thorough evaluation process for software-controlled aircraft, and must rebuild it's approval process from the ground up before any further approval process can move forward. This could add years to an already-slowed approval process which could easily mean the end of 737 MAX production.

Fast Company once featured an article about the code and coding process that went into the Space Shuttle. It noted, “you can’t have people freelancing their way through software code that flies a spaceship, and then, with peoples lives depending on it, try to patch it once its in orbit.” The principles for writing shuttle code included:

1. The product is only as good as the plan for the product.
2. The best teamwork is a healthy rivalry.
3. The database is the software base.
4. Don’t just fix the mistakes — fix whatever permitted the mistake in the first place.

As software has proliferated from the space age to commercial aviation, perhaps it's time to apply more rigorous standards to aviation software development. Our current circumstances with the 737 MAX grounding make point #4 especially relevant which could explain the delay in return to flight.

A best-case scenario would be to provide and approve pilots having a way to fully override all autopilot systems at any time. The two 737 MAX crashes were caused by an autopilot system, so if the pilots had this option, deadly crashes like these caused by software bugs could be avoided. If Boeing could provide this, the FAA could possibly approve the 737 MAX flying again before the fourth quarter.

Establishing this kind of precedent would have other uses, too. Just as sophisticated transportation-related coding has gone from space to air, so coding is increasingly coming to cars and vehicles as well. As we add autopilot, of sorts, and autonomy to land-based vehicles, so too, having single-step access to an override mode to take over control over the vehicle is going to be something drivers or passengers want to have. I called for this years ago, although the regulatory trend has gone in the other direction and not been in favor of driver or passenger control.

No comments:

Blog Archive