Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Does Tesla’s model encourage sloppy programming?

This site may earn commission on affiliate links.

sleepydoc

Well-Known Member
Aug 2, 2020
5,587
9,929
Minneapolis
When Tesla came out with over the air updates it was a revolutionary concept. Other makers are slowly adopting the concept but no one uses it and pushes out updates the way Tesla does. As a new owner, until the V11 interface I would eagerly download the next update wondering what improvements I would get. The total cluster of V11 opened my eyes and made me wonder, does the practice and ability to continually and easily push out an update breed sloppy programming and bug checking?

With traditional automakers a firmware update was a major ordeal, requiring owners to go to the dealer and costing a lot of time and money so they did them infrequently, only when necessary and spent a lot of time testing and verifying to make sure there were no bugs. As an owner, you never got updates but you got fewer bugs, too.

With Tesla, updates come on a regular basis but bugs are a regular occurrence, to the point of being expected. But it’s ok Because the next update will fix the bug. I’m thinking that part of the reason we see so many bugs is the simple fact there’s far less urgency and need to prevent them.
 
...sloppy programming and bug...

The same practice with boasting the next revolutionary step of radar deletion that would free the system of sensor fusion, stationary object collision, and phantom brakes but when it's implemented, it's buggy, and inferior to the radar version, prompting the software recall and NHTSA investigation.

The standard practice should be to test the software in-house team first and once it's proven (alpha, beta...) only then it'll go to customers.

Tesla customers are unpaid bug testers.
 
Last edited:
  • Like
Reactions: _Redshift_
And "In-house" MUST INCLUDE actual drivers (not the devs) in actual cars on actual roads. Not the UI developers with 20/20 vision "testing" it out on their 38" dual monitors on their desk.
And not just roads IN CALIFORNIA-- try a colder climate now and then.
Very true.

And also in the rain too!

 
  • Like
Reactions: HankLloydRight
OTA updates enable continual improvements and A/B testing but also sloppy updates when combined with web business practices instead of safety critical business practices.

Tesla should do more testing before releasing updates to customers. It'd save them work and give customers more confidence.

Beta testers need a bug feedback channel. Meanwhile, few customers are equipped to be beta testers.

A large fraction of drivers won't read forums or the (changing) owners manual, e.g. busy people who aren't car enthusiasts, spouses who rarely drive the car, and car renters.

A bug like breaking the next-FM-station button could've been caught by testing it once, by code review, automated tests, a QA team testing across a sample of car models/years/variations, or beta testers.

Each UI change should get usability testing with a range of drivers in a range of road conditions in a range of locations before building more changes upon it and before rolling it out to beta testers. This is Usability 101.

I suspect few drivers are equipped to be good driving instructors to self-driving cars, esp. without a way to indicate why they disengaged autopilot or why they didn't enable it.
 
The same practice with boasting the next revolutionary step of radar deletion that would free the system of sensor fusion, stationary object collision, and phantom brakes but when it's implemented, it's buggy, and inferior to the radar version, prompting the software recall and NHTSA investigation.

The standard practice should be to test the software in-house team first and once it's proven (alpha, beta...) only then it'll go to customers.

Tesla customers are unpaid bug testers.
But the customers are not bug testers, BECAUSE THERE IS NO FEEDBACK CHANNEL. That at least would help to make this hamster wheel of bugs and janky design decisions make some sense. Also, there is no way in hell the basic car interface should be perpetually in alpha/beta for everybody. We need a 'production' release (like once or twice a year, with some bug fixes only tossed in more often) for people who want boring and well tested, and a 'testing' channel for those who want to try the bleeding edge. And the number of bugs in basic things like the radio controls should be going down with time, not up.
 
But the customers are not bug testers, BECAUSE THERE IS NO FEEDBACK CHANNEL. That at least would help to make this hamster wheel of bugs and janky design decisions make some sense. Also, there is no way in hell the basic car interface should be perpetually in alpha/beta for everybody. We need a 'production' release (like once or twice a year, with some bug fixes only tossed in more often) for people who want boring and well tested, and a 'testing' channel for those who want to try the bleeding edge. And the number of bugs in basic things like the radio controls should be going down with time, not up.

Great proposal!

One idea: People should be able to switch channels at any time, and switching to the production channel should quickly get the current production release.
 
  • Like
Reactions: smogne41
I've heard that Tesla does, in fact, have real beta testers. But they are all basically "friends&family" types and are biased and not actual "real" beta testers. They wear membership in the "beta test" program as a badge of honor. And they rarely supply negative feedback of the latest test versions to Tesla for fear of being kicked out of the beta program and losing their bragging rights.