Your premise is that the same programmers code the AP and 8.1. I don't find that plausible.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Your premise is that the same programmers code the AP and 8.1. I don't find that plausible.
disagree
hos epi to polu
AndrewTX
Member
This is the kind of meaningless statement that 'kinda, sorta' sounds like a promise which gets Musk into trouble. Why spout off on Twitter like this?
AP and FSD likely have separate development groups.
But all of the on board software likely goes through a common group that integrates the software apps, creates the official release build, performs internal and external testing, and then distributes the software updates.
What Tesla has been doing with the 8.0/AP 2 updates, while also trying to bring 8.1 out shortly afterward, usually creates challenges for any software organization.
It makes sense to focus on getting 8.0/AP 2 functionality stable - and then integrate that into the 8.1 release and perform another round of testing (hopefully short) on the combined 8.1/AP 2 software, before distributing it.
It's just taken Tesla longer to achieve their 8.0/AP 2 goals than they had expected (which isn't unusual for aggressive software projects).
Seems like we are getting close...
But all of the on board software likely goes through a common group that integrates the software apps, creates the official release build, performs internal and external testing, and then distributes the software updates.
What Tesla has been doing with the 8.0/AP 2 updates, while also trying to bring 8.1 out shortly afterward, usually creates challenges for any software organization.
It makes sense to focus on getting 8.0/AP 2 functionality stable - and then integrate that into the 8.1 release and perform another round of testing (hopefully short) on the combined 8.1/AP 2 software, before distributing it.
It's just taken Tesla longer to achieve their 8.0/AP 2 goals than they had expected (which isn't unusual for aggressive software projects).
Seems like we are getting close...
aesculus
Still Trying to Figure This All Out
I would think it's more like they want to insure the look and feel on both platforms are in sync so they don't suffer the wrath of AP2 vs AP1 comparisons.
Chill out friend. My post wasn't directed at you, I just used your post to make my point. If you have read the entire thread, then you are aware of some of the hilarious rationalizations of Tesla's behavior that have been made here.
Not everything is about you
No problem. I am chill.
I agree that there are histrionics on both sides (the "Rationalizers", and the "I'll sue!" crowd).
I have no interest in being lumped in with either group, and simply prefer that you use a Rationalizer's post to make your point. As you've said, they fill up this thread, so you have a lot of usable content.
Actually there is another way.
Let Musk be the visionary. Let him dazzle us with his dreams that may one day come true.
Train him to speak slightly differently as he dazzles us
At the same time, develop an effective Communications Team that can take Musk's dreams, run them by their internal experts (software, hardware, manufacturers, operational people, legal, etc) and come up with an implementation strategy that is feasible. Then communicate that to the public.
For instance:
Current Tesla:
(Tweet from Musk): We expect a Tesla will drive from SF to NYC without human input by December 2017!"
(no other communication, so this solidifies from "dream" to "expectation")
My proposed Tesla
(Tweet from Musk): I'm hoping that we can have a Tesla drive from SF to NYC without human input in the very near future! (slightly different wording, but still conveys Musk's vision)
Tesla Communications Team/PR/Speaker: "In our quest to fulfill our vision, Tesla's current time frame is to have AP2 cars reach parity with AP1 cars by XXX date. After this, we will roll out features in this order. A. B. C. This is our current progress on initiative X,Y, and Z. Here are our current hurdles, and how we anticipate overcoming them. For more information, please see the official Press Release at Teslamotors.com"
There isn't necessarily a causal relation. I think all of this "Linux Kernel means a better browser" is based on varying interpretations of an Elon tweet that mentioned the two together, and has just gotten repeated over and over again in various echo chambers until everybody believes it's true.
Having built Linux-based embedded systems for a living...there could reasonably be a couple of reasons: 1) a newer kernel could enable some capability that's required by a newer browser implementation. 2) a newer kernel could be more efficient. So far we haven't seen any hard evidence of either one, and short of actually talking to someone in Tesla engineering who works on this stuff (or a few of the right leaks), we probably won't know for sure. Really we as users shouldn't care about the version of the Linux kernel, distribution, whatever.
(BTW, you may have "used" Linux without being aware of it. Linux and *BSD get used in a lot of embedded software nowadays.)
Bruce.
boonedocks
MS LR Blk/Blk 19”
Ulmo
Active Member
I agree with everything quoted. Linux did browsing very fast in 1996-2003. They don't seem necessarily related (even though there is always the possibility). The application program itself along with how it hogs the present resources seems to be a large portion of the bottlenecks that slow things down these days. Bloat and lack of parallelization (even if due to a few poorly implemented and/or placed lockstep serial flows) are still efficiency killers in today's software, even more so than ever. I find it much more likely the kernel upgrade has around no intrinsic impact (however I can think of at least one extrinsic change that could cause it).
That's not what he tweeted.
Your underlying assumption that nothing has changed in browser technology in the last 5 years is way off the mark.
(As someone who does this stuff as well…)
Agreed. The only circumstances I can imagine where there'd be a strong relationship between updating the Linux kernel and a dramatically better browser experience is if the current older kernel keeps them from being able to update their userland stack beyond a certain version of Webkit (presumably the engine they are transitioning to, given their use of QT) or enabling certain functionality.
And the ones that come to mind would be JIT and hardware accelerated scrolling/rendering, with the latter being most likely.
We can only speculate, at least until we see what comes out and have someone with root access compare with how it's different.
Why do you think EAP and FSD have different dev groups? Doesn't seem more logical that it's one group? FSD will be a feature set extension from EAP, no?
Because he is advising the most prolific twitter spouter Trump and it's rubbing off on Elon
I'm no programmer, but it seems to me that full automation is not necessarily an extension of simple TACC and AutoSteer. The sort of programming necessary for FSD would be on an exponentially more complicated scale, so much so that it seems completely plausible you wouldn't need the same skill sets for one as the other.
Am I the only one worrying about the security angle? Lots of folks fork Linux and Webkit and then let it rot. Every maker of routers, and most phone makers for instance. We've seen the results - which included botnets big enough to cause mass domain name resolution outages. Even Webkit updated weekly is arguably not sufficiently secure for automotive use... Ancient Webkit+2.6 kernel are probably insecurable, barring some harrowing use of isolation techniques, most of which didn't exist in the 2.6 era. How many known vulnerabilities were in git over the weekend in 2007 when Tesla made their forks?
Google is updating it's Pixel phones monthly. How can Tesla not be doing the same and also be secure? My fear is we are driving a Samsung Galaxy S3 on wheels. They should yank out that stupid browser - at least then they could have a hope.
Last edited:
