Follow Slashdot blog updates by subscribing to our blog RSS feed


Forgot your password?

Comment Re:More difficult with people? (Score 3, Insightful) 160

beta testing his shitty half-assed autopilot junk on the buying public

Questions Answer: Yes
(1) Have you ever used Autopilot before? 99 %
(2) Are you familiar with the car warnings that Tesla provides about how Autopilot is to be properly used? 98 %
(3) Are you aware that when you first enable the Autopilot, you have to do so through the Drivers Assistance section of Settings on
the center screen? 93 %
(4) Are you aware / Do you know that after enabling Autopilot, you had to agree to an acknowledgment box which stated that
Autopilot “is an assist feature that requires you to keep your hands on the steering wheel at all times” and that “similar to the
autopilot function in airplanes, you need to maintain control and responsibility for your vehicle” while using Autopilot? 99 %
(5) Do you know that each time you activate Autopilot, a message appeares on the screen behind the steering wheel stating:
“Please Keep Your Hands On The Wheel; Be Prepared To Take Over At Any Time“? 96 %
(6) Based on these communications, have you understood that when using Autopilot, the driver is expected to maintain control of the
vehicle at all times? 98 %
(7) Has the name “Autopilot” caused you to believe that the car is fully autonomous, meaning that it does not require the driver to be
supervising the car? 7 % (No : 93 %)

There was an interesting study done (unrelated to the German owners survey above) which showed that the minor autopilot failures (occasional lane drift, unexpected speed changes) are ironically improving consumer safety. Users were well aware of its ability to make mistakes specifically because they're common enough, and this keeps the vast majority of users from treating the vehicle like a tool you don't have to pay attention to it; instead they tend to treat it more like cruise control. As automation improves, the danger may counterintuitively increase as users get used to never having to do anything when the vehicle is driving and thus stop paying attention.

At the same time, despite the frequency of errors, the overwhelming majority of users felt that its failures presented either no risk, or little risk, as they tend to be things that any reasonable driver could react to (in the same way that we don't fear cruise control because if it's looking like it's going to drive us into the rear of the car ahead of us, we slow down). E.g. autopilot never just suddenly jerks the wheel to hard right in the middle of a road or whatnot. They also get quite used to what situations you use it in and what you don't use it in (just like people do with cruise control); the fact that the system won't let you use it when it perceives its ability to follow the road to be too poor doesn't even need to factor into the equation.

Comment Re: Screw it (Score 5, Insightful) 160

That's the thing I don't get. SpaceX is saving the US government huge amounts of money. Yet so many Slashdotters have this weird conception that they're a giant leach sucking government budgets dry. Their conception is precisely the opposite of reality. ULA has been getting an unbelievable sweetheart deal for government launches, getting paid even when they don't launch anything, and charging massive fees when they do, while also getting government subsidy to develop new craft. SpaceX paid back its COTS funding in spades versus what was being doled out to ULA.

Comment Re:I'm shocked! (Score 5, Informative) 160

It's like you didn't even read the article or pay attention to what he said. So I guess someone has to repeat it for you.

NASA's regulations for propulsive landing of a Dragon 2 capsule are too difficult to reasonably meet. So they're dropping propulsive landing from Dragon 2. Meaning it can't land on Mars either. At the same time, they've decided that there's a better approach to landing on Mars than Dragon 2's approach of a bottom-mounted heat shield and side-mounted thrusters.

And for the record, that better approach is what they're looking at with ITS - a side lifting body heat shield with base thrusters for landing. The latter spreads the heat out over a much larger area (Dragon 2 had no option for that because it had no giant, partially empty propellant tanks attached) and increases the length of time over which the heating occurs, slowing the rate.

It'll be interesting to see their changes to ITS. I'm glad to see that "smaller" is among them - I like ambition, but ITS was a step too far, IMHO.

Comment It would be useful to label drugs with issues... (Score 1) 316

Most drugs just deteriorate into something useless but not necessarily dangerous, usually in a decaying exponentiial. A few drugs deteriorate into something toxic, or otherwise having different drug-like effects. Another few may react with their breakdown products to decay at a harder to predict non-linear rate.

IMHO it would be useful to identify, publicly, which are which - especially the ones that get toxic after a while. It would also be useful to have an estimate of how rapidly they degrade under various storage conditions.

That way people could avoid things that get dangerous with age, but use (at their own risk) longer-lived drugs, perhaps slightly raiding the dose, for long after the 95% effectiveness "expiration date".

Comment Re:Ask me how I can tell you're a Republican. (Score 1, Troll) 139

Perfectly willing to ignore that the actual hacking that Democrats are concerned about was related to breaking into email accounts for Democrats, not elections computers ...

Nobody on the R side is ignoring that. But what the Ds are trying to distract everybody from is that the alleged "Russian hacking" consists of exposing what the major Ds were actually saying to each other (about how they cheated Sanders out of a legitimate primary run, what contempt they had for the voters and how they lied to them, and so on).

If the Russians were really behind those leaks, and they did swing the election, it seems to me that the people involved deserve both a close encounter of the law-enforcement kind AND a Pulitzer Prize.

It is absurd to ignore how the GOP has been found by numerous courts to have deliberately targeted selected mechanisms of ID because the opposing voting base uses them, ...

It's not absurd if:
  - the opposing voting base uses them because they're easy to fake, AND
  - the not-so-easy-to-fake alternatives are both easy to obtain (by any QUALIFIED voter) and free. ... when the GOP is found by numerous courts to have engaged in ... gerrymandering ...

Gerrymadering is a two-team sport. If you don't believe the Ds also play a good game, take a look at California, just for starters.

Also: "... found by numerous courts ..." cuts no ice when egregiously left-wing and activist judges are one of the grievances of both the Rs and the voters for them.

But thank you for playing straight-man. B-)

Comment BLE or classic Bluetooth (Score 1) 70

It's not completely clear to me from TLA whether this builds on BLE (Bluetooth Low Energy a.k.a. Bluetooth Smart), though the reference to the 4.0 level of the standard, the many instances of the word "Smart" in the article, that it's a mesh, and the nature of the protocols all suggest it's BLE.

If so, it will be interesting to see how they keep it from eating up the devices' batteries. BLE devices get a couple years out of a coin cell by spending about 99.5% of their time "asleep", drawing roughly three orders of magnitude less current (~5 uA rather than ~5 mA) than when awake and with the radio on. Typically only a clock using a watch crystal is running during that time.

(Yes, that 99.5% isn't rhetoric. An advertisement takes about 5 ms, so a configuration of one advert per sec comes out to a duty cycle of about 1/200.)

They get away with that because they have a distinction between centrals (which have line power or (like smartphones) big batteries with frequent recharging) and peripherals (little battery powered devices that must only sip electrons). Peripherals can transmit when they feel like it. But centrals have to spend a lot of time listening, and the receiver (which has a lot to do) is (counterintuitively) slightly more power hungry than the transmitter.

If you try to build a mesh network out of what were formerly peripherals, not only do they have to spend several times as much battery power forwarding other devices' messages than they do handing their own (if everybody is equally chatty), but if the scheduling isn't set up right (or while listening for new players) they may need to leave the receiver on for substantial periods listening to the crickets chirp. That would just KILL battery life.

So I await the opportunity to peruse this addition to the spec. with bated breath.

(But not held breath. Even without this, the BLE v4.1 standard was 2,841 pages of some of the crummiest prose I've had the misfortune to have to try to understand. And I may be the only person to successfully implement a T1/T2/T3 framer from just the Bell standards and a logic analyzer bitstream capture.)

Comment Re:ZigBee & Z-Wave. (Score 1) 70

Zigbee on the other hand performs well and has been growing by leaps and bounds.

And, as I hear it, BLE / "Bluetooth Low Energy" / "Bluetooth Smart" - a different networking protocol from original Bluetooth in the lower layers but sharing some of the upper layers, or at least the upper layer design approaches - was largely created (absorbed into the Bluetooth standard from its inventors, with their gleeful cooperation) in reaction to ZigBee's success.

Comment Re:Dear clients: (Score 3) 243

Where I am, our code doesn't work with newer versions of a dependency library. Two developers have tried to work around the incompatibility, and failed. So until we can scrounge together enough time to redevelop the frontend from scratch, we're stuck installing old versions of the library, and just hoping that no OS changes render older versions of the library inoperable - otherwise we won't be able to upgrade OSes either.

That said, this doesn't fit the topic, because our boss knows all of this. We keep our boss well in the loop. He used to work as a programmer the software, and still does some work on it from time to time. That's IMHO how it should be.

Things work best when workers aren't afraid of their boss. I hate Machiavellian workplaces.

Comment I've said it before, and I'll say it again: (Score 2) 125

Autopilot is the best excuse for a driver getting into an accident that ever was invented. "No officer, it wasn't me! My car did it on its own!"

Thankfully, it's easy for Tesla to avoid legal liability for things like this because the car logs when autopilot is actually in use and what it's doing. Unfortunately, it doesn't help with the PR aspect, as the media just blindly reports that it was Autopilot before taking the time to find out if it actually was.

Comment Maybe it is your case. Here's my notes. (Score 1) 483

I looked at your post again and I think maybe it WAS the same problem. (Mine was a Toshiba - satellite, I think - bought surplus from work that went to black-screen-on-login after it updated to Natty 11.4 and was still broken with Oneiric 11.10)

So here's my notes from the install.

Sat May 19

  - Backed up critical part of home directory to white 4G thumb drive.
  - Used "upgrade" button in update manager to go to new revision.
        - It went to Natty. (11.04?)
        - Broken:
              - New workspace.
              - Freezes once taskbar shows.
              - Could get console login on pseudoterm ctl-alt-F1
              - Console login prompted for another upgrade.
  - Upgraded again (using command interface.) Went to onieric (11.10)
        - Still frozen.
        - Live CD can't see encrypted disk.

Sun May 20.

  - Much debugging:
        - Initially thought it was mouse buttons. But:
              - Mouse worked fine on login screen.
              - Control-alt-delete, enter: Logs out. Screen flashes then
                  back to login.
              - So it looks like a permission issue.
              - Much looking around on net found stuff on it:
                    - Related to Nvidia driver:
            - (compiz?) isn't changing permission somewhere like previous
                workspace did so userspace stuff can't touch screen.
            - User got past it by:
                  - Going to classic no-effects in login screen options.
                  - Uninstalling nvidia "recommended" driver in "additional drivers".
                          - Then fixed by installing (via Synaptic) and enabling
                      (in Jockey/additional drivers) xserver-xorg-video-nouveau
                      driver (which he reports as being faster than Nvidia's
                      proprietary driver.)
                      (Something about moving-aside /etc/X11/xorg.conf)
Mon May 21 21:47:14 PDT 2012

  - Continued debugging:
        - Was able to get it working by selecting a 3D session.
        - Tried disabling Nvidia's driver.
              - This ended up with a system that didn't display X screen asking
                  for full-disk password.
              - Got around that by rebooting, which got grub options to
                  try recovery mode and using that.
              - Disabling driver apparently works by:
                    - putting nvidia in /etc/modprobe.d/blacklist-local.conf (only entry)
                    - putting nvidiafb in /etc/modprobe.d/blacklist-framebuffer.conf
                  so modprobe wouldn't load it. But taking it out didn't fix boot.
            - /boot/(image) was fresh: Looks like it's a snapshot with the
                broken config embedded.
              - Was able to get live by:
            - Recovery mode
                    - Removing above blacklists.
            - Removing (putting aside as .neo) xorg.conf (which was apparently
            - go to live mode.
            - sudo modprobe nvidia, sudo modprobe nvidiafb
            - text login as [my i.d.]l
            - startx.
              - Reinstalled nvidia-current and went back to using the 2d version
                  for now.
            > (Try fixing properly next weekend. Too much to do now.)

  - Firefox is working.
  - Can get to email on yahoo.
  - Libre Office will open resume. Good.

  - Flash not working:
        - Stuff about HTML5 trial at YouTube, but opting out didn't help.
        - Downloaded flashplugin-downloader & flashplugin-installer.
        - That was enough to bring it live WITHOUT actually running it...
            (Probaly either autoran on triggers or fixed missing piece.)

Comment Re: We'll be fine. (Score 1) 201

So the model S isn't very good around the track.

You're interpreting "not being designed for the track" as "has bad handling", as if the two are at all the same thing. The Model S has superb handing, and reviews are almost uniformly in agreement on this. It's not a track car because it's not designed to handle track cooling loads, having nothing to do with handling.

The track car market is much smaller than the luxury sedan market, so obviously it isn't their target. That said, they do plan to make an actual track car, which will be their next generation Roadster (the first generation, like Tesla's other cars, was a road car, not a track car). It's also targeting a bone-crushing sub-2-second 0-60, too - they're calling the drive mode "Ultimate Plaid" ;) And that's stock, with stock tires, etc.

Comment Re:Newer Linux, tablets, Win 10 on VMWare, OSX (Score 1) 483

Each time a new build comes through the system boots into a black screen and that's the end of it.

There is/was a bug in some video drivers that caused the screen to come up black, though the system was otherwise working fine. (I ran into that several years ago installing Ubuntu on a Toshiba laptop. They had just gone to Unity and the chip in the laptop wasn't quite initialized correctly.) There was a workaround (with an argument to the bootloader?) for startup, to get an initial live screen, then another to make it work automatically and persistently.

Don't recall it. But I was able to find it on the net by googling with the laptop model and verbiage about the problem. (Since you're on tablets, if it's a similar problem it might be different in detail, so my experience wouldn't apply directly.)

If you've got some older tablets gathering dust, you might want to try installing a later-patch-release version of a recent release of some linux distro - and if they don't come up, google for the symptoms. You might end up breathing new life into them.

IMHO it's not that the developers are deliberately dropping support or killing older machines. It's just that, as they work on new stuff, sometimes they break older stuff and don't have enough thousands of different machines to discover it before release - so it goes out and gets fixed (if at all) only after somebody complains.

Slashdot Top Deals

Trap full -- please empty.