From: Jason James on 10 Mar 2010 08:18 "John Tserkezis" <jt(a)techniciansyndrome.org.invalid> wrote in message news:4b9774ed$0$19545$afc38c87(a)news.optusnet.com.au... > On 10/03/2010 12:28 PM, Jason James wrote: > >> Software bugs should be pretty rare these days. > > Dunno about that. Across all areas, it appears to me that software is > suffering more now than in the past. That's depressing. > Not something I can explain, I don't know if it's dealing with vast > volumes of code to squeeze in all the features demanded of the market, > management pushing projects out sooner than they should, or programmers > are just getting plain lazy. The inimitable programmers are slacking off? I would never have believed it!! > Either way, code is getting sloppier. A full turn of the wheel, where the techs iron out design bugs during install.. Jason
From: Doug Jewell on 10 Mar 2010 16:17 John Tserkezis wrote: > On 10/03/2010 12:28 PM, Jason James wrote: > >> Software bugs should be pretty rare these days. > > Dunno about that. Across all areas, it appears to me that software is > suffering more now than in the past. > > Not something I can explain, I don't know if it's dealing with vast > volumes of code to squeeze in all the features demanded of the market, > management pushing projects out sooner than they should, or programmers > are just getting plain lazy. I think a bit of both. Back in the late '80s when I first started working with embedded systems, everything was Z80 or 6800 CPUs, and all the programming was done by hand assembly from scratch. Also, everything was single tasking, with response to data inputs done by polling. These days quite a bit of the stuff is using i386 or similar architecture, which allows for multi-tasking and multi-threaded programming. Response to inputs is usually done by interrupts. A lot of the programming is patching together existing libraries. Allows for more capabilities in the program, but introduces more risk of bugs. One of the libraries may have an undocumented "feature" that hasn't been discovered, or a newer version of a library may behave slightly different in some circumstances, or the most common problem is that in a multi-threaded app a library won't be multi-entrant. These types of bugs can be very hard to diagnose, as the events that cause them to show up may only occur once in a blue moon. For example the problem may only occur if input B occurs within 5 microseconds of input A, and causes a library to re-enter code that is not re-entrant. This type of fault may be what happens when a priarse is driving down the highway. It could be something as obscure as the driver puts his indicator on at the exact moment the computer decides it needs to put the airconditioner on, that throws the computer into an error. > > Either way, code is getting sloppier. -- What is the difference between a duck?
From: Doug Jewell on 10 Mar 2010 16:22 Noddy wrote: > "Trevor Wilson" <trevor(a)rageaudio.com.au> wrote in message > news:7vntnpFs0iU1(a)mid.individual.net... > >> **By those comments, I assume that you will not purchase any car >> manufactured after (say) 2015 - 2020. Electronic throttle control will be >> ubiquitous by that time. Same as electric power steering, electrically >> operated valve trains, electrically operated clutchs/gearboxes, etc, etc. >> Like it or not, electronic controls allow manufacturers considerable >> freedoms and considerable potential to lower manufactured costs. > > That doesn't mean they're great. > > Drive by wire as currently fitted to some vehicles is *appallingly* bad, and > electric steering assistance has already shown itself to be less than > stellar in the reliability stakes. There is no reason for it to be bad. Aircraft have had fly-by-wire for decades now, where it is controlling many more variables than just the throttle. Of course there is the old adage "if it ain't Boeing i'm not going", but Airbus's issues are to do with the programming of the FBW, not the reliability of the hardware. > > -- > Regards, > Noddy. > > -- What is the difference between a duck?
From: Doug Jewell on 10 Mar 2010 16:29 Jason James wrote: > "Milton" <millame23(a)yahoo.com> wrote in message > news:4b9619b7$0$27789$c3e8da3(a)news.astraweb.com... >> Oh dear! Mr Toyota must be shaking in his boots. >> >> http://news.ninemsn.com.au/world/1025243/Toyota-Prius-in-runaway-drama-on-California-freeway >> >> A Toyota Prius has accelerated out of control on a busy California freeway >> before police intervened to bring the vehicle to a standstill, police >> said. >> James Sikes, 61, was driving on the busy Interstate 8 freeway outside San >> Diego when he noticed his car was starting to accelerate of its own >> accord, the California Highway Patrol said. >> The terrified motorist was helpless as the car hurtled out of control >> along the road at speeds more than 90 miles per hour. >> However Sikes was able to call police, and officers using a loudspeaker >> were talked the driver through the process of slowing down by using his >> emergency brake and then turning off the engine. >> Police then pulled in front of the car as it decelerated and rolled to a >> stop and put the rear bumper of the squad car against the front of the >> Prius. >> The incident came as Toyota staged a technical demonstration to attack >> allegations by a vocal critic that problems with its electronics may cause >> its cars to speed out of control. >> In recent months, Toyota has recalled more than 8 million vehicles >> worldwide due to acceleration issues. >> The latest incident in California was a chilling echo of the incident last >> August where off-duty California Highway Patrol Officer Mark Saylor was >> killed along with his wife, her brother and the Saylors' 13-year-old >> daughter when the accelerator of the Lexus ES350 they were in got stuck. >> Minutes later, the Toyota-manufactured vehicle slammed into the back of a >> sport utility vehicle at about 100mph, veered off the freeway, overturned >> and burst into flames. All four family members died. > > Is the throttle "fly by wire". On that issue alone, I wouldnt buy such a > car. The issue isn't FBW on it's own. I'm sure you're aware aircraft have used FBW for ages. The issue as I see it with these Toyotas, is that they have a programming fault that shows up occasionally, locking out human input - even to the point that it locks out turn-off. Of course most cars allow the ignition switch to be turned to the off position, but the keyless ignition of the toyotas requires pressing and holding the start button. Quite possible that if there is a computer fault, that the press & hold won't work either. At the very least it is a non-intuitive means of shutdown. I personally think they should have a master switch on the dash that shuts down all electrical systems other than lights. > > Jason > > -- What is the difference between a duck?
From: John Tserkezis on 10 Mar 2010 16:30
On 11/03/2010 5:50 AM, Jason James wrote: > There are parameters which deal with this type of thing ie MTBF > (Mean time between failures)which have been around for ages. Ahem. Doesn't hold much. The system used to calculate the number is different for each manufacturer, and sometimes between products. And certainly doesn't translate to serviceable life either. In other words, unless you're comparing same brands, you're not going to get anything useful on which is more reliable. In any case, it doesn't give you a real-life figure of X-years N-months from new before it's likely to fail figure anyway. |