Firmware 6.0 (beta version discussion)

[brianman]

Okay, if you take one sentence out of the two paragraphs, then fine. Disagree with the sentence. But no, using my logic doesn't mean you can have no objective evaluation by a 3rd party of any software project ever undertaken. I didn't say that. That's taking it to the extreme. That kind of extreme statement is great for scoring points in a debate in front of a college crowd, but not so great for an actual discussion about issues.

I'm disappointed, too, like many people here. But not in the software we've seen so far.

There was no extremeness in what I said. You made a statement among other statements, and it seemed like it was stated pretty crisply to stand on its own.

As I stated before, I think the releases should have reasonable cadence and the volume of features in them should be adjusted accordingly. How long did we wait for the suspension "change back" (kinda)? How long do we have to wait for 6.0?

"The big mega release that you have to wait forever for and includes a dozen different things" is a bad model IMO. Tesla seems, at times, to have adopted this model and I find it frustrating and counterproductive for both Tesla and owners.

They have an update mechanism that allows them to push out fixes quickly. They should be able to put out features regularly without bundling them into a giant deliverable. There is no "model year" for the hardware -- the adjustments just roll into the product line. Why is the software delivery more rigid and blocky than the hardware delivery?

 

[bonnie]

There was no extremeness in what I said. You made a statement among other statements, and it seemed like it was stated pretty crisply to stand on its own.

As I stated before, I think the releases should have reasonable cadence and the volume of features in them should be adjusted accordingly. How long did we wait for the suspension "change back" (kinda)? How long do we have to wait for 6.0?

"The big mega release that you have to wait forever for and includes a dozen different things" is a bad model IMO. Tesla seems, at times, to have adopted this model and I find it frustrating and counterproductive for both Tesla and owners.

And those are reasonable points. My frustration is with the picking apart of what does appear to be in the upcoming release (though again, we don't know that's the full list). There are a number of posts that make it sound like [pick your favorite request] should have been easy to implement for even the simplest of engineers, and other posts that put down what they did get.

There are probably some constraints regarding testing/release that are dictated by the fact this is a car with a higher risk profile, rather than an off-the-shelf consumer product. That may negatively impact. But that's within Tesla's control in setting expectations.

I'm not arguing or defending promises made. Only trying to point out that there are a lot of people here that wound up their own expectations to the point that they couldn't possibly be pleased. And if you missed that in my posts, then I need to work on clarity.

 

[brianman]

... There are a number of posts that make it sound like [pick your favorite request] should have been easy to implement for even the simplest of engineers, and other posts that put down what they did get.

Generally I've given up on trying to convince and/or correct such folks of the blunder there. "It should be easy to do X" is a great way for me to tune out of the paragraph of discussion that follows it.

 

[Hodginator]

I'm really looking forward to the updated maps. I live in northern VA and the person with the best traffic information wins (well at least they get to eat dinner with their family). Now that the car will provide alternate routes even based on traffic delays between work and home is great!

I also think being able to use my cell phone instead of my key fob is a big security improvement. I hope I never have to take my fob with me again. Less in my pocket and more secure. What more can you ask for?

I had hoped for better media control through BT, but when you think about it, it's pretty damn amazing that we have cars that will continue to have more functionality. I now for a fact that the features will just get better and better.

I can't wait to see this update pop up on my screen!

 

[Sacrament055]

I like the features but I really hoped to see the open api in this release. Would be nice to get 3rd pay apps going and Elon had mentioned it would be out this year.

 

[TexasEV]

I like the features but I really hoped to see the open api in this release. Would be nice to get 3rd pay apps going and Elon had mentioned it would be out this year.

People have different wants and needs (mostly wants). I'm happy with the improvements that will be helpful for me, and I'm not going to complain about the others as I know they'll be useful to someone else. As for what's not on the screenshot of this beta, everyone should chill until we see what's in the general release and not jump to conclusions.

 

[Haggy]

Not on Audi's or Mercedes it's not. You'll get a warning, and will not be able to restart the car, but it doesn't just die.

I consider that a weakness. Acura and Infiniti won't let you drive away very far without the FOB. There are legitimate reasons why you might not want to kill the engine when you get out of the car, but I can't think of a legitimate reason why somebody should be able to step into my car and drive away with it simply because I'm nearby. If there is some good reason that I'm missing, I'd be fine with Tesla adding an option to enable or disable driving away without a fob. I might want to get out to open a garage door, or have a mechanic drive onto a rack, but fobs are supposed to be more convenient versions of keys, and the good ones stay in the pocket. A car won't drive away without a key.

If somebody wants to tell me why there's an advantage to driving away without a key, I'd like to hear it. That's not because I want to debate it, but out of simple curiosity.

 

[ItsNotAboutTheMoney]

There was no extremeness in what I said. You made a statement among other statements, and it seemed like it was stated pretty crisply to stand on its own.

As I stated before, I think the releases should have reasonable cadence and the volume of features in them should be adjusted accordingly. How long did we wait for the suspension "change back" (kinda)? How long do we have to wait for 6.0?

"The big mega release that you have to wait forever for and includes a dozen different things" is a bad model IMO. Tesla seems, at times, to have adopted this model and I find it frustrating and counterproductive for both Tesla and owners.

They have an update mechanism that allows them to push out fixes quickly. They should be able to put out features regularly without bundling them into a giant deliverable. There is no "model year" for the hardware -- the adjustments just roll into the product line. Why is the software delivery more rigid and blocky than the hardware delivery?

Dependencies and testing. It's why most major software projects do not have a rapid release cycle. Major changes hit lots of code and then you end up choosing between increased testing load or bundling changes.

 

[jerry33]

If somebody wants to tell me why there's an advantage to driving away without a key, I'd like to hear it. That's not because I want to debate it, but out of simple curiosity.

The advantage isn't being able to drive away without the key--there should be adequate "The key fob isn't connecting" warnings. The advantage is in not having the car shut completely down and stop when it loses connection with the key. The car can't tell when it's safe to stop, so regardless of how far the "not very long" distance is, it might be too short in a specific case. For example, you're driving across a bridge, though a tunnel, or just far away from anywhere and the fob loses connection to the car for whatever reason. If the distance that it lets you drive is shorter than the bridge is long, the car stops. I'd hate to drive a car like that through Death Valley.

 

[islandbayy]

I consider that a weakness. Acura and Infiniti won't let you drive away very far without the FOB. There are legitimate reasons why you might not want to kill the engine when you get out of the car, but I can't think of a legitimate reason why somebody should be able to step into my car and drive away with it simply because I'm nearby. If there is some good reason that I'm missing, I'd be fine with Tesla adding an option to enable or disable driving away without a fob. I might want to get out to open a garage door, or have a mechanic drive onto a rack, but fobs are supposed to be more convenient versions of keys, and the good ones stay in the pocket. A car won't drive away without a key.

If somebody wants to tell me why there's an advantage to driving away without a key, I'd like to hear it. That's not because I want to debate it, but out of simple curiosity.

I think for me this would be convenient. For one, I do a lot of long distance driving. I'm not going to carry two fobs with me. If I loose both, I'm royally screwed. Likewise, If i'm out of town and loose one, I'm currently screwed for at least 24 hrs, as I would need someone to go to my home, get my spare, then overnight it to me. If I could do without even needing the fob, and start my car via my phone, or have a family member load the app (in the case phone is gone too) and start my car for me, I'm all for that. or better yet, instead of having the phone start the car, you could just use the phone to open the doors, then have a LONG LONG access code that would need to be typed into center console to start the car (say 20+ characters with Capital Letters and numbers required). I would like that better.

 

[Haggy]

These new features don't deserve more than a year of beta testing... The calendar feature is unnecessarily complex!
Where is the new Chrome browser?

This is not the full list of features for sure!

I won't know until I see it, but for now it doesn't seem as if the calendar has enough features. My simple $150 android phone looks at my calendar for me, reminds me of upcoming appointments, offers me directions, and tells me what time I need to leave to get to my appointment on time. It does all that on its own. I don't want to step into a car and tell it all over again. With the phone, I don't even need an app, as long as it's on my calendar. And I can use Google on my computer or phone to put something on the calendar, or have it sync on its own from Lightning or Outlook, etc.

If I step into the car and bring up my calendar, I also see any calendars I'm authorized to see. So if my wife has a meeting scheduled somewhere, I can pull up my calendar in Google, and I would want to click on an appointment and have the car take me there.

Google+ has the concept of circles. I have a circle with my immediate family, one with my extended family, and so on. Google has location tracking. So I can see the location of members of my family at a city level or of my immediate family at a specific location. I can set it any way I want. I don't even need to use the app, since the widget shows me all that, but if I want to pick up my daughter somewhere, I simply look at my phone, swipe until I get to her, and touch navigate. Then I can drive there, following the directions on my phone. This isn't special technology. It's something that anybody with a Google account has by default and something that anybody with an Android phone already has. An advanced vehicle should go beyond what I can already do for free with a cheap handheld device. If I have something I already use, and I end up searching for the perfect suction cup mount for my phone so I can use it instead of a giant built-in screen, then something isn't right.

 

[gg_got_a_tesla]

I wonder if Tesla can rope in Techie owners - many of whom are already contributing to open source projects and such beyond their day jobs - at some point in some sort of a closed source development effort to turn things around faster for less risky, UI-centric firmware enhancements. Things like media player, navigation, trip planning and so on.

This is not quite opening up the SDK to the entire world (which will hopefully happen in 2-3 years) but, more on the lines of a tightly-controlled developer program with very strict NDAs in place

Tesla engineers can still be the gatekeepers for what makes it into the car's firmware updates eventually but, they can crowdsource much of the mundane, sometimes-grunt-work this way.

I know they are probably extremely busy with developing features specifically for the X as well - the "Unfold Falcon Wings" button?! - and I also think there are too few of these engineers to scale these releases up better.

I'd gladly do it for free to improve things in my Model S at a higher cadence than 7-8 months

 

[Haggy]

In the end, almost everyone here wants different things, and I can't see how Tesla could ever begin to meet all those different requests. Although I'm pretty sure "Shuffle" can't be all that hard. :biggrin:

I agree about the shuffle part. It's not that I would ever use it, but it's on so many simple devices and implementation would be easy. There might be a long list of things that people want, but what I see as important is that people shouldn't have to give up what they already have. I want a car to get me from point A to point B. If it's capable of knowing where I'm going, I want it to tell me how to get there. If I want to call somebody's phone number while I'm driving, it should allow me to do that legally and it's a basic feature that even inexpensive cars have. The era of pulling off the road to make a phone call is long gone. I'd put all those things before shuffle, or even before having a radio, since getting there and being able to communicate are far more basic issues.

The problem with car navigators, and the reason I'd choose an external one over the one I had in any car before (they were never optional so I got them anyway) is the navigation software itself is static. The car's design will become final at least a year before it comes to market, and an average car owner will keep the car long enough that on average, the navigator will be about four years out of date. And at the end of a typical ownership, it will be way past that. Tesla has the ability to address the issue by upgrading, since a good computer and a good display could handle completely different software if needed. When a navigator won't show dynamic traffic or show speed limits, and can be upstaged by a handheld one, I have trouble accepting that. Again, I'm not asking for anything fancy. I simply want what I can get at no cost on my phone, or what Garmin built into cheap handheld navigators many years ago but left out of the Tesla.

 

[bonnie]

I wonder if Tesla can rope in Techie owners - many of whom are already contributing to open source projects and such beyond their day jobs - at some point in some sort of a closed source development effort to turn things around faster for less risky, UI-centric firmware enhancements. Things like media player, navigation, trip planning and so on.

This is not quite opening up the SDK to the entire world (which will hopefully happen in 2-3 years) but, more on the lines of a tightly-controlled developer program with very strict NDAs in place

Tesla engineers can still be the gatekeepers for what makes it into the car's firmware updates eventually but, they can crowdsource much of the mundane, sometimes-grunt-work this way.

I know they are probably extremely busy with developing features specifically for the X as well - the "Unfold Falcon Wings" button?! - and I also think there are too few of these engineers to scale these releases up better.

I'd gladly do it for free to improve things in my Model S at a higher cadence than 7-8 months

But then you hit the Mythical Man Month issues, where bringing people up to speed (and subsequent meetings required to handle larger teams) potentially slows things down. Tar pits, here we come!

I love the idea in theory, but can't imagine I'd ever allow it at a company where I worked - you'd have to deal with IP issues, training, getting people up to speed on risk issues, etc. And there would be legal issues because there are always legal issues, especially if you're not paying people but directing their work. And honestly, if Tesla wants to get the work done, they should pay people to do it and keep the technology knowledge in-house.

 

[Haggy]

I wonder if Tesla can rope in Techie owners - many of whom are already contributing to open source projects and such beyond their day jobs - at some point in some sort of a closed source development effort to turn things around faster for less risky, UI-centric firmware enhancements. Things like media player, navigation, trip planning and so on.

This is not quite opening up the SDK to the entire world (which will hopefully happen in 2-3 years) but, more on the lines of a tightly-controlled developer program with very strict NDAs in place

I'd like to see the firmware made hackable. I don't mean that in the criminal sense, but in the traditional sense of the word, so that owners can make mods to tailor thing to their own needs. I'd want it to be able to run something in a sandbox, and simply give me a piece of the screen and access to a data connection. Beyond that, I'd want some macro capability but only to the extent that it could do the equivalent of something I could do by pressing buttons on my own. For example, if somebody wants the car to change radio stations when he drives by his kid's school at 3pm, it should be doable. It's not something I'd ever want to do, and it's not something I'd want Tesla to waste time on, but if a user wanted it on his own, then why not?

 

[galangg]

I'm sure it's PIN or password protected. It would be foolish for it to not be protected in some way, and Tesla are no fools. Presumably it's an emergency option, and would be very handy in the event of a lost key or dead battery. I was hoping for more Waze functionality (essentially Waze rendered for the center screen), and the name thing seems lame. It if was going to be used as a key word to starting listening for an instruction (i.e. Jeeves, play me the best of One Direction (obviously, in this particular scenario a long silence would ensue) then it would make some sense.

Definitely would be great to have WAZE integrated into the center console.

 

[Haggy]

Generally I've given up on trying to convince and/or correct such folks of the blunder there. "It should be easy to do X" is a great way for me to tune out of the paragraph of discussion that follows it.

Exactly. It's easy to do X is a great way in general to say what anybody wants on a wish list, but saying it should have been in this release is pointless. It's fine for people to wish for things in the future all they want. Not everybody will have the same priorities, but just because I won't use something doesn't mean I can't see why others would.

 

[markwj]

Sometimes the reasons for a major (x.0) release are under the hood restructuring with little visible changes.

 

[Haggy]

Actually there are effectively two digital security barriers, you need the username AND the password if you're going to duplicate someone's phone app control. I've read many times that the majority of car thefts are opportunistic and it's hard to see why someone would want to target one specific VIN and then download an app and go to all the trouble of hacking someone's login credentials.

If you sign out of the app the password is needed to log back in (regardless of whether the phone is locked; having it password enabled means that the phone (at least) is one level of security higher than just using your key fob. If the phone is locked also then it's two levels better than the key. Even if you don't lock your phone, keeping the app live is no less secure than carrying your key fob around; just remember not to lose either.

A lot of security related apps, such as ones that control your home door locks or your online banking have yet another level of security: a pin. There's a reason you'd want one in addition to a password. I want my password to be as difficult as possible to guess, since anybody with a web browser can try to get into my account. I want a long string of random numbers, letters and special characters. It doesn't even matter if I can remember them, because I can use a password manager and have to remember a single 20 character password for that alone to get to my account. For an app on my phone, I have a few more levels of security. I have it in my pocket, as opposed to anybody having a web browser. I can have a pattern or fingerprint unlock or a password for the phone itself. At that point, I have the app. If I can give it the user name and password, but have it ask me for a pin whenever I go away from it, then I can get into it easily without having to log into it each time. Some apps of this type need individual devices to be authorized, so the app is downloaded, a web session lets you generate a code, and the app prompts you for the code if it doesn't already have your credentials. But the bottom line is that I can use apps such as these and only had to give them my password once, but I still have a pin for an extra level of security.

All that makes it way more secure than a fob, and in the worst case, if somebody stole my phone, I'd be able to change my password (or deactivate the device if it's implemented that way) before anybody came close to getting to the app.

- - - Updated - - -

Sometimes the reasons for a major (x.0) release are under the hood restructuring with little visible changes.

That's true, and for bigger things such as a new navigator, it could mean significant changes to the code base. Other features that don't require changes at that level could be introduced earlier rather than making people wait. The catch is that it might require developing them in both the new and old code base, since nobody could hand somebody something complete to work on prior to implementing each feature. I've had to do that before. It's more work than releasing everything at once. But giving somebody a change, and having to implement it and test it twice is sometimes a good idea if it keeps people from waiting.

- - - Updated - - -

A solution would be to add another router located between your router and your Tesla and near the edge of the range of your first router.

How far is the distance between your wifi router and your car?

Actually, a wireless access point. I'm not trying to be picky, but I want people to know what to look for. Some routers have WAP features built in. Others don't.

 

[iadbound]

I agree about the shuffle part. It's not that I would ever use it, but it's on so many simple devices and implementation would be easy. There might be a long list of things that people want, but what I see as important is that people shouldn't have to give up what they already have. I want a car to get me from point A to point B. If it's capable of knowing where I'm going, I want it to tell me how to get there. If I want to call somebody's phone number while I'm driving, it should allow me to do that legally and it's a basic feature that even inexpensive cars have. The era of pulling off the road to make a phone call is long gone. I'd put all those things before shuffle, or even before having a radio, since getting there and being able to communicate are far more basic issues.

The problem with car navigators, and the reason I'd choose an external one over the one I had in any car before (they were never optional so I got them anyway) is the navigation software itself is static. The car's design will become final at least a year before it comes to market, and an average car owner will keep the car long enough that on average, the navigator will be about four years out of date. And at the end of a typical ownership, it will be way past that. Tesla has the ability to address the issue by upgrading, since a good computer and a good display could handle completely different software if needed. When a navigator won't show dynamic traffic or show speed limits, and can be upstaged by a handheld one, I have trouble accepting that. Again, I'm not asking for anything fancy. I simply want what I can get at no cost on my phone, or what Garmin built into cheap handheld navigators many years ago but left out of the Tesla.

Agree that the car needs to do the basic things that people expect from the technology they already have. In another thread (or maybe it was earlier in this one), I made a long post suggesting that Tesla will need to bring the functionality that people can already get from their phones or other cars on the market to the Model S and the Model X (at launch) because the target market for the Model X will demand it and reviewers won't give Tesla a pass if the technology hasn't advanced by then. I'll emphasize that this doesn't have to happen in 6.0, but it better happen before too long.