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

Firmware 5.8.6/7/8/9/10

This site may earn commission on affiliate links.
A lot of the angst over software releases could be removed if Tesla had an "open beta" program, allowing owners to "opt in" or "opt out" of having access to beta software releases.

Though, they also need to make two additional changes to provide owners more control over the software running in their cars. First, they should have the ability for owners to know WHICH VERSION is being installed and the RELEASE NOTES of what will be installed, when the owner is presented with the option to install a new software update. Owners can then decide if and when to install the update.

Second, Tesla should always allow owners to "rollback" to the previous stable software release. This is something that Tesla should be able test easily with each new software update. This doesn't have to be an arbitrary rollback to whatever the owner was running previously (which would be better) - at least a rollback to a single, specific release.

The combination of these changes would provide early access to more owners, provide owners more control over what they are running, and if/when any problems are encountered, owners have an escape clause back to a stable release.

Reducing the risk of installing new software would allow Tesla to get early software releases into the hands of more owners - something that many software companies do today - and with more testers - they can more quickly find problems - and get them fixed - both improving the quality of the released software and shortening the testing cycle.

And, for those owners that are willing to accept some problems in order to get new features they think are really important - such as the auto lowering, improved navigation or improved entertainment software - I suspect there are a number of us willing to accept some bugs, if we have the ability to report problems back to Tesla - and, as a last resort, we could go back to something that worked reasonably well.

Tesla's solved the hard problems in building a great car - resolving the problems with their software release process should be much easier - and is something that other companies have already solved...
 
A lot of the angst over software releases could be removed if Tesla had an "open beta" program, allowing owners to "opt in" or "opt out" of having access to beta software releases.

Though, they also need to make two additional changes to provide owners more control over the software running in their cars. First, they should have the ability for owners to know WHICH VERSION is being installed and the RELEASE NOTES of what will be installed, when the owner is presented with the option to install a new software update. Owners can then decide if and when to install the update.

Second, Tesla should always allow owners to "rollback" to the previous stable software release. This is something that Tesla should be able test easily with each new software update. This doesn't have to be an arbitrary rollback to whatever the owner was running previously (which would be better) - at least a rollback to a single, specific release.

The combination of these changes would provide early access to more owners, provide owners more control over what they are running, and if/when any problems are encountered, owners have an escape clause back to a stable release.

Reducing the risk of installing new software would allow Tesla to get early software releases into the hands of more owners - something that many software companies do today - and with more testers - they can more quickly find problems - and get them fixed - both improving the quality of the released software and shortening the testing cycle.

And, for those owners that are willing to accept some problems in order to get new features they think are really important - such as the auto lowering, improved navigation or improved entertainment software - I suspect there are a number of us willing to accept some bugs, if we have the ability to report problems back to Tesla - and, as a last resort, we could go back to something that worked reasonably well.

Tesla's solved the hard problems in building a great car - resolving the problems with their software release process should be much easier - and is something that other companies have already solved...

Sometimes in software rollback is a difficult task to implement, since during updates internal data structures are migrated to a new format beyond just business logic, making a rollback is as difficult to develop and to test than the time required to develop the update, I could understand why Tesla doesn't want to go in the business of rollbacks since they want to minimize costs and thinking on update for forward improvements and not spending the R&D on rollbacks, however in software there is a very common practice of Betas and there is where Tesla should be focusing the efforts.

So implement Betas and let the user take the risk.

At this point I see every time a user go the Service Center they become the new Beta users without knowing

I work in Software and felt for the firmware update and I'm paying a small price until next mayor update, now I'll read TMC before agree to install new firmware.

Anyway I think they are doing a great work since they improve and fix changes quickly, they need to focus a little more on the collateral effects of a fast pace firmware development to minimize bug randomization and keep the high standards, if they lose customer trust then will be a bigger problem for them
 
-1000 on open betas. Part of the idea of a beta is to help find bugs, Bug + car = very bad,

They do have betas, just not open ones. The last thing you want is to have some unfriendly media downloading a beta and then create headlines about how flakey the firmware is, and how unsafe the car is, and how it's not ready for prime time. And you can bet that's what would happen as there is a lot of unfriendly media. Open betas are fine when the worst that can happen is that your internet search fails but they are not great for critical equipment.
 
Sometimes in software rollback is a difficult task to implement, since during updates internal data structures are migrated to a new format beyond just business logic, making a rollback is as difficult to develop and to test than the time required to develop the update, I could understand why Tesla doesn't want to go in the business of rollbacks since they want to minimize costs and thinking on update for forward improvements and not spending the R&D on rollbacks, however in software there is a very common practice of Betas and there is where Tesla should be focusing the efforts.
...

It's not that they're hard to implement, it's that they're very hard to tack on afterwards to a design that didn't account for such a feature to begin with. That aside, I'm sure that given enough information ahead of an upgrade and the choice to accept or decline, would significantly reduce the need for an actual rollback in most cases. It becomes a merely "nice to have" when you aren't surprised by the content of an update.
 
Other companies with software products have open betas. If problems are found in the beta release - and then publishes them online in the forum - that should be OK - because it's a BETA! Beta releases will have problems - and as long as Tesla corrects the problem before the official release, there shouldn't be any media concerns. And, if by having more beta testers Tesla finds and fixes more of the problems before the official release - some of the negative comments posted in the forums would go away.

And, if Tesla is in this for the long haul - if their current software can't support rollback - then they should fix that problem.

Tesla's done a pretty good job in designing and manufacturing a high quality car in the Model S - and their software implementation and release system should be improved to match the quality of the car's hardware...

They aren't doing anything different in their software that other companies have been doing for years - so no "breakthroughs" are needed - just more discipline in their software process and design.
 
If I had to guess, controlled beta is critical part of Tesla's software approach. They have lots of owner and telemetry data and I think they use it to progressively roll out releases to representative sample to surface issues before broad release. As a result, critical issues, such as when a few IL owners couldn't charge, have been successfully minimized. This is also why we see so many dot releases in rapid succession. Open, self-select beta would defeat this. There might actually be some innovation going on here. Agree something like this is required for a car vs say a browser. I know others will say balderdash but I think many are stuck in a PC software paradigm.
 
Open betas are great when it's easy to roll back to a prior version … or simply use a different URL. Don't currently have that option with the Tesla. Then, imagine the customer support issues when many people download a beta who don't really understand what they're getting into. It's definitely in Telsa's interest to keep this a closed beta program.
 
Open betas are great when it's easy to roll back to a prior version … or simply use a different URL. Don't currently have that option with the Tesla. Then, imagine the customer support issues when many people download a beta who don't really understand what they're getting into. It's definitely in Telsa's interest to keep this a closed beta program.

+1 and in owners' and shareholders' interest as well
 
I think the difference between alpha and beta should be noted here. Alpha may have notable bugs. Beta has been carefully vetted but may still have minor hidden bugs. No, Tesla can't guarantee there's no fatal flaws, but then they can't guarantee that for full releases either... I'm +1 for Open Beta. You just have to sign a release first. That will solve the risk issues.
 
Most of the software changes in each release are not related to the critical operations of the car. You can see that when you reboot the touchscreen - while the car is moving. While the touchscreen is blank and going through the reboot sequence - the car drives normally, and other than losing the sound from turn signals - haven't noticed anything that impacts operation of the car. So the risk of new software releases causing a serious driving failure are probably pretty small - as long as Tesla has properly isolated the mission critical operations (which they appear to have done).

As someone who has experience in software projects ranging from complex, mission critical, real-time, systems to web-based portals - Tesla should be able to do much better on their software releases. It's taking way too long for major functionality to be implemented and deployed - they are considerably behind other auto manufacturers in areas such as navigation and entertainment features. And the more frequent "point" releases seem to have frequent problems, requiring additional point releases.

If Tesla is in this for the long haul - then re-investing in their software architecture and methodology would be much easier to do now - than later, as they start dealing with the more complex problems of supporting cars running in many countries - with many languages - and multiple generations of the car's hardware and features.
 
Most of the software changes in each release are not related to the critical operations of the car. You can see that when you reboot the touchscreen - while the car is moving. While the touchscreen is blank and going through the reboot sequence - the car drives normally, and other than losing the sound from turn signals - haven't noticed anything that impacts operation of the car. So the risk of new software releases causing a serious driving failure are probably pretty small - as long as Tesla has properly isolated the mission critical operations (which they appear to have done).

As someone who has experience in software projects ranging from complex, mission critical, real-time, systems to web-based portals - Tesla should be able to do much better on their software releases. It's taking way too long for major functionality to be implemented and deployed - they are considerably behind other auto manufacturers in areas such as navigation and entertainment features. And the more frequent "point" releases seem to have frequent problems, requiring additional point releases.

If Tesla is in this for the long haul - then re-investing in their software architecture and methodology would be much easier to do now - than later, as they start dealing with the more complex problems of supporting cars running in many countries - with many languages - and multiple generations of the car's hardware and features.


You seem to assume that part of the update doesn't include updates to other pieces of firmware in the car. Just because you can reset the screen does not mean that other things there were updated don't keep running.
 
My Model S wouldn't connect with my iPhone last week prior to 5.8.8. TM had me disconnect the phone from the car, the car from the phone, reboot the car, turn phone off/ on and then reconnect to solve the problem.

FWIW, my Jaguar (prehistoric pre-Model S ICEberg) used to do this every time it got a software update. From my experience, bluetooth = unpredictable. Not sure you had to do the reboot, probably would have sufficed to delete and re-pair the phone.