2019.40.50.7 installing now

[mswlogo]

ACTUALLY, that's not at all true. the new hotness is to create model-based designs, then feed that into a 'code generator' and that makes provably safe and secure runtime code. its been this way for a while (not on linux, though) and it moves the 'logic' up a level.

google 'autosar' for what I'm talking about. or add 'autosar asil d' for your search.

one search result: AUTOSAR | Vector (although I hate to send people to Vector; their stuff is so expensive and they have such an attitude, as a company. but they are the #1 in that field, fwiw)

So you are saying they could just pour their whole NN database, compiler etc. and out pops perfect code.
Awesome. Problem solved.

Or that they could pour in say even the voice command code and out pops perfect code that would work for every voice !!

I'm sure there are some closed systems that you can prove all bases are covered and the code won't crash but this is was more complex than that.

 

[kavyboy]

So what I am hearing is you are okay with bad updates as long as you can choose which bad update you are on? I am on 50.5 still so I am not sure what the angle is here. I have not been on 50.6/7 yet.

I still think the advanced update argument is irrelevant for what I am talking about. I am talking about poor releases in general. It doesn’t matter if I decide to stand my ground and say no thank you to your future updates...the fact is I am currently on a badly rushed update right now. The whole idea behind these patches is to fix bugs/stability, and that is clearly not happening. I completely understand if it was for core releases, but what are supposed to do when bug fixes make it worse?

the fact that they can push an update and the Tesla community can find issues immediately shows there is a huge lack in QA testing.

If this is a response to me, no, I'm not at all OK with it. As a software developer myself, it rankles me how bad Tesla's processes must be. I've just learned the best way to live with it. That includes never expecting the next update to fix more than it breaks.

 

[Kennymb5384]

If this is a response to me, no, I'm not at all OK with it. As a software developer myself, it rankles me how bad Tesla's processes must be. I've just learned the best way to live with it. That includes never expecting the next update to fix more than it breaks.

this is was not directed at any one person. There were multiple people that mentioned just picking which version you wanted to be on, and staying there. I don’t like the idea of just living with it though. I think we need to be vocal about the things we don’t like, or things that aren’t working. Seems to work for future features haha. I work with software developers everyday, so I know it’s gotta piss you off to see stuff like this. As a dev, what do you prefer? The way they are releasing so many patches in so little time, or the idea of spending more time on a specific patch? That is a serious question too. I am not a developer by any means, so more curious...

 

[kavyboy]

this is was not directed at any one person. There were multiple people that mentioned just picking which version you wanted to be on, and staying there. I don’t like the idea of just living with it though. I think we need to be vocal about the things we don’t like, or things that aren’t working. Seems to work for future features haha. I work with software developers everyday, so I know it’s gotta piss you off to see stuff like this. As a dev, what do you prefer? The way they are releasing so many patches in so little time, or the idea of spending more time on a specific patch? That is a serious question too. I am not a developer by any means, so more curious...

Frequency isn't an issue with me, really. I use Netbeans and Postman, for example. The Netbeans release schedule is set years in advance. Postman updates constantly. Both are fine because they don't keep breaking things and the new features/fixes are actually desirable.

When things like the heating steering wheel stop working out of the blue (and that happened a while back), that makes me really uneasy (and I don't even have a heated steering wheel!). Some of the UI changes have been questionable at best, too.

What I would really love, and would pay for from a third party, is to be able to pick a long-term stable release and just get safety and security updates.

 

[Twiglett]

It would be nice to have Microsoft O365/Windows style of release options where you can choose from just Stable releases up to "Beta and be damned"
Then default the car to the slowest Stable release schedule so that you only get releases that have specific update docs while all the others just get the standard "Bug fixes and improvements" type notice

 

[Kennymb5384]

Yeah, there is no perfect answer. Looks like you just deal with the current process, and hope it improves over time. Pros still outweighs the cons with these cars!!

 

[Twiglett]

Yeah, there is no perfect answer. Looks like you just deal with the current process, and hope it improves over time. Pros still outweighs the cons with these cars!!

exactly my thought as well. There are so many posts on here about something breaking or not charging as fast or capacity dropping etc.
But I know Tesla will fix it and things will return to normal.
For me it was the really uncomfortable update to AP that made it so unsettled in the lane, especially around semi-trailers or large trucks.
This last update fixed it completely - and also made the lane merge issue much better.
There just isn't any point getting bent out of shape over it.

 

[nepenthe]

Still doesn’t seem like you are getting my point. I totally understand that software will have bugs, and no software will be free from bugs, but if you are releasing patches that are meant to fix said bugs, do you not expect it to do just that? If bugs with for example voice commands were reportedly not working in 50.1 and you release 50.5/6/7 and they are not getting better, or worse according to some, is that really not a reason to be worried. I totally understand you cannot make everyone happy, but isn’t this forum a place to express that? Which ever side you are on...

Presumably, there is a QA/Test/Validation process for each release, even on the Advanced chain. A bug introduced and or found seven days ago, probably shouldn't be fixed on a release that follows within days, never-mind a week - if there is in fact this kind of validation process in place. Providing it wasn't a major or critical bug that is.

But I do get your point, some releases feel rushed, happens industry wide, from largest to smallest companies. There's just not another car company that operates like Tesla in the software realm. So really, they are more comparable to a software company and in that light, it does not seem surprising or irregular.

 

[Kennymb5384]

Presumably, there is a QA/Test/Validation process for each release, even on the Advanced chain. A bug introduced and or found seven days ago, probably shouldn't be fixed on a release that follows within days, never-mind a week - if there is in fact this kind of validation process in place. Providing it wasn't a major or critical bug that is.

But I do get your point, some releases feel rushed, happens industry wide, from largest to smallest companies. There's just not another car company that operates like Tesla in the software realm. So really, they are more comparable to a software company and in that light, it does not seem surprising or irregular.

Yeah, definitely. Will be interesting with the Mach E since Ford will be doing OTA updates as well...

 

[pilotSteve]

ACTUALLY, that's not at all true. the new hotness is to create model-based designs, then feed that into a 'code generator' and that makes provably safe and secure runtime code. its been this way for a while (not on linux, though) and it moves the 'logic' up a level.

google 'autosar' for what I'm talking about. or add 'autosar asil d' for your search.

one search result: AUTOSAR | Vector (although I hate to send people to Vector; their stuff is so expensive and they have such an attitude, as a company. but they are the #1 in that field, fwiw)

Thanks that's interesting info. I was mostly kidding, but model-based or specification-based designs that use code generators should be better than traditional methods.

But its always about the specifications. GIGO right?

 

[NHK X]

I work for a SASS company, so I definitely get how some things can go missed. But how many patches does it take before you fix an issue? I think there is more concern around the idea that this is my car and the safety side. When you hear that someone’s car just shutdown while driving after the latest patch, it gets worrisome. May be a one off, but If they can spend the time releasing multiple fixes in a span of a week, they can take the time to regression test and spend more time on one solid release. Just my opinion

You sure it wasn’t just a mean screen shutdown and reboot? The whole car turned off? I read the release sections and don’t remember reading about this.

 

[linux-works]

Thanks that's interesting info. I was mostly kidding, but model-based or specification-based designs that use code generators should be better than traditional methods.

But its always about the specifications. GIGO right?

not when the GUI enforces policy when you try to enter things that can be checked real-time, then you can do various sanity-checks before you save/export, then there are other phases where you can run simulations and validate. its quite interesting, if you like to understand how aircraft and other mission critical things have been designed (approached) for a while, now.

 

[linux-works]

strange, I drove around for a while, enabled wifi at home, turned software on and off (updates) and could not trigger one. it won't gimme one, today, it seems. [grapes]maybe .5 is the better version, anyway[/grapes]

 

[Soey22]

I can confirm with this update I can do automatic lane changes on city roads. I assume it was on the 50.1. I did about 300 miles on NOA and the MX P Raven never messed up. I was extremely impressed and continued using automatic lane changes on city roads.

 

[bhakan]

I can confirm with this update I can do automatic lane changes on city roads. I assume it was on the 50.1. I did about 300 miles on NOA and the MX P Raven never messed up. I was extremely impressed and continued using automatic lane changes on city roads.

I have been using TACC and AP on local roads for quite some time and the car would switch lanes and turn the turn signal off also once it finishes the lane change. It does not initiate auto lane change like NOA on hwy, with my MADMAX setting, but once I press the turn signal stalk it changes lane and turns off the signal also. I use AP on local roads extensively and it is mostly flawless for sometime. From local roads on AP and NOA it enters the hwy without any hesitation after 50.1 even without lane marking. This is new to me.

 

[Alset Srotom]

I can confirm with this update I can do automatic lane changes on city roads. I assume it was on the 50.1. I did about 300 miles on NOA and the MX P Raven never messed up. I was extremely impressed and continued using automatic lane changes on city roads.

If you are talking about hitting the turn signal lever to trigger a lane change, we have been able to do this for a while now...

 

[follow.johnny]

Just received this update coming from 2019.40.50.5

 

[outdoors]

yeah, I was only talking about the high frequency in the latest patches. Maybe I am just new to the club, and not used to seeing 4 versions/patches released within a month. When they initially released the new voice commands it never worked for me, not even once. As the patches were released, it slowly got better and was working about 60% of the time. Then the latest patch that I am on (50.5) came out and now it only works some times with the steering wheel button. Doesn’t work at all if I click on the microphone button on the screen. This is the only reason I used voice commands as an example.

While you would like you fix for voice commands I understand, but I kind of want my charger to go back to full power.

Story:

For every software bug fix release it could be fixing someone's car that just plain doesn't work, or charges at half rate. I have seen multiple updates in a month. 3 or 4. Couple one day after the previous. If it is a critical fix, then guess what. Everyone gets it, doesn't matter if either update was yesterday, and if you didn't install the first one they pull it from you. While the notes could say. Hey couple cars couldn't charge right, might not be you. That would create nightmares from people seeing each and every issue.

Model S charges at half rate. 3 charges fine at home. Thinking I have to replace the on board AC charger for S. Tesla Virtual service comes in and says it is a firmware issue on 2019.40. They are working on it, and not a hardware issue.

So should Tesla wait on my critical update till they figure out how to tell the glove box to open in different dialects? Or should one just not click advanced? Or if they click advance they can see the update. Go read about it here after a week and go at it. Really Tesla can't win in this case. Some want the latest with the bugs, others want the update that is to their liking less buggy.

Edit: Car had ears ringing. 50.7 coming in. Pressing install.

 

[jrmnet]

I was also on 50.1 (hardware 2.5) and voice recognition went way downhill whereas prior to that it was awesome. Installing 50.7 now.

 

[linux-works]

yay! finally, the tesla gods heard me.

I did have to sacrifice a young goat. only one, this time, though.