About the F-22 software bug
in [ADA]
|
Prev: Ada in Boeing 787
Next: Private or public task ?
From: Dmitry A. Kazakov on 5 Feb 2010 05:31 On Fri, 5 Feb 2010 01:02:40 -0800 (PST), Martin wrote: > On Feb 5, 8:52�am, Ludovic Brenta <ludo...(a)ludovic-brenta.org> wrote: >> Martin wrote on comp.lang.ada: >> >> >> >>> On Feb 5, 6:42�am, Dennis Lee Bieber <wlfr...(a)ix.netcom.com> wrote: >>> > On Thu, 04 Feb 2010 19:46:15 +0100, Pascal Obry <pas...(a)obry.net> >>> > declaimed the following in comp.lang.ada: >>> > > I know that the F-22 is 90% of Ada. Is there some public information >>> > > about this bug? Is that a design bug? >> >>> > � � � � So far as I recall -- from some years ago -- it was an algorithm >>> > problem handling position information wrap-around from crossing, as >>> > mentioned, the International Dateline... -180.0 to +180.0 deg longitude. >> >>> > � � � � I didn't hear that they had to follow the tankers back -- was under >>> > the impression once they managed to cross back heading east a reboot of >>> > the navigation system started working again... >> >>> > -- >>> > � � � � Wulfraed � � � � Dennis Lee Bieber � � � � � � � KD6MOG >>> > � � � � wlfr...(a)ix.netcom.com HTTP://wlfraed.home.netcom.com/ >> >>> For some reason (that I don't get at all) lots of systems define long >>> as -180 <= x <= +180 degrees. >> >>> Having the potential to alias a position seems like a bad idea for a >>> start, so when I've been coding such systems up, I've always spent a >>> bit of time making it convert positions into the range -180 <= x < >>> +180 degrees and using a proper ADT. >> >>> I wonder if it was anything to do with that?... >> >> I would have thought a longitude was really a mod 360, shifted by -180 >> for display purposes? For fractional degrees (i.e. minutes and >> seconds), make that mod (360 * 60 * 60), shift by -180 * 60 * 60 and >> split in degrees, minutes and seconds when displaying. >> >> -- >> Ludovic Brenta. > > No...it's -180 <= x <(=) +180...always - check any map / globe!! > > Lat is always -90 <= x <= +90 deg - no doubt about that one :-) > > Adding "shifts" would make understanding any problem very hard... And shift does not solve the problem anyway, if that existed. The potential problem is that the angle is not a real number. It could be represented by one, but then the operations like +, -, *, /, =, /= must be replaced and ones like <, >, <=, >= disallowed. With that done the value +180 would do no harm, because -180 = +180 in terms of proper ADT operations. You simply would not be able to distinguish them (without tricks like Unchecked_Conversion). -- Regards, Dmitry A. Kazakov http://www.dmitry-kazakov.de
From: Martin on 5 Feb 2010 06:18 On Feb 5, 10:31 am, "Dmitry A. Kazakov" <mail...(a)dmitry-kazakov.de> wrote: > On Fri, 5 Feb 2010 01:02:40 -0800 (PST), Martin wrote: > > On Feb 5, 8:52 am, Ludovic Brenta <ludo...(a)ludovic-brenta.org> wrote: > >> Martin wrote on comp.lang.ada: > > >>> On Feb 5, 6:42 am, Dennis Lee Bieber <wlfr...(a)ix.netcom.com> wrote: > >>> > On Thu, 04 Feb 2010 19:46:15 +0100, Pascal Obry <pas...(a)obry.net> > >>> > declaimed the following in comp.lang.ada: > >>> > > I know that the F-22 is 90% of Ada. Is there some public information > >>> > > about this bug? Is that a design bug? > > >>> > So far as I recall -- from some years ago -- it was an algorithm > >>> > problem handling position information wrap-around from crossing, as > >>> > mentioned, the International Dateline... -180.0 to +180.0 deg longitude. > > >>> > I didn't hear that they had to follow the tankers back -- was under > >>> > the impression once they managed to cross back heading east a reboot of > >>> > the navigation system started working again... > > >>> > -- > >>> > Wulfraed Dennis Lee Bieber KD6MOG > >>> > wlfr...(a)ix.netcom.com HTTP://wlfraed.home.netcom.com/ > > >>> For some reason (that I don't get at all) lots of systems define long > >>> as -180 <= x <= +180 degrees. > > >>> Having the potential to alias a position seems like a bad idea for a > >>> start, so when I've been coding such systems up, I've always spent a > >>> bit of time making it convert positions into the range -180 <= x < > >>> +180 degrees and using a proper ADT. > > >>> I wonder if it was anything to do with that?... > > >> I would have thought a longitude was really a mod 360, shifted by -180 > >> for display purposes? For fractional degrees (i.e. minutes and > >> seconds), make that mod (360 * 60 * 60), shift by -180 * 60 * 60 and > >> split in degrees, minutes and seconds when displaying. > > >> -- > >> Ludovic Brenta. > > > No...it's -180 <= x <(=) +180...always - check any map / globe!! > > > Lat is always -90 <= x <= +90 deg - no doubt about that one :-) > > > Adding "shifts" would make understanding any problem very hard... > > And shift does not solve the problem anyway, if that existed. The potential > problem is that the angle is not a real number. It could be represented by > one, but then the operations like +, -, *, /, =, /= must be replaced and > ones like <, >, <=, >= disallowed. With that done the value +180 would do > no harm, because -180 = +180 in terms of proper ADT operations. You simply > would not be able to distinguish them (without tricks like > Unchecked_Conversion). > > -- > Regards, > Dmitry A. Kazakovhttp://www.dmitry-kazakov.de Yup, instead you end up with functions like "Is_Between (Anlge, First, Last : Longitude) return Boolean" Although because have 'clever' construct functions "Longitude (From : Float)" that always put things in the 'correct' range (-180 <= x < 180), we could allow "=" (and "/="). Cheers -- Martin
From: Hibou57 (Yannick Duchêne) on 5 Feb 2010 11:50 On 5 fév, 11:31, "Dmitry A. Kazakov" <mail...(a)dmitry-kazakov.de> wrote: > And shift does not solve the problem anyway, if that existed. The potential > problem is that the angle is not a real number. It could be represented by > one, but then the operations like +, -, *, /, =, /= must be replaced and > ones like <, >, <=, >= disallowed. I would have thought for reals, operations like +, -, *, /, <, >, /= changed and =, <=, >= disallowed Is there something I don't understand ?
From: Dmitry A. Kazakov on 5 Feb 2010 13:19 On Fri, 5 Feb 2010 08:50:31 -0800 (PST), Hibou57 (Yannick Duch�ne) wrote: > On 5 f�v, 11:31, "Dmitry A. Kazakov" <mail...(a)dmitry-kazakov.de> > wrote: >> And shift does not solve the problem anyway, if that existed. The potential >> problem is that the angle is not a real number. It could be represented by >> one, but then the operations like +, -, *, /, =, /= must be replaced and >> ones like <, >, <=, >= disallowed. > > I would have thought for reals, operations like +, -, *, /, <, >, /= > changed and =, <=, >= disallowed > Is there something I don't understand ? You can compare angles, but cannot order them, at least while keeping transitivity of ">": a > b /\ b > c => a > c Also a combination of operations like yours is impossible: a /= b <=> not (a = b) through the law of excluded middle (if not in the intuitionistic logic). Then: a > b \/ a = b <=> a >= b So ">" and "/=" would carry the rest with. -- Regards, Dmitry A. Kazakov http://www.dmitry-kazakov.de
From: Phil Clayton on 5 Feb 2010 18:39
On Feb 5, 9:02 am, Martin <martin.do...(a)btopenworld.com> wrote: > On Feb 5, 8:52 am, Ludovic Brenta <ludo...(a)ludovic-brenta.org> wrote: > > > > > Martin wrote on comp.lang.ada: > > > > On Feb 5, 6:42 am, Dennis Lee Bieber <wlfr...(a)ix.netcom.com> wrote: > > > > On Thu, 04 Feb 2010 19:46:15 +0100, Pascal Obry <pas...(a)obry.net> > > > > declaimed the following in comp.lang.ada: > > > > > I know that the F-22 is 90% of Ada. Is there some public information > > > > > about this bug? Is that a design bug? > > > > > So far as I recall -- from some years ago -- it was an algorithm > > > > problem handling position information wrap-around from crossing, as > > > > mentioned, the International Dateline... -180.0 to +180.0 deg longitude. > > > > > I didn't hear that they had to follow the tankers back -- was under > > > > the impression once they managed to cross back heading east a reboot of > > > > the navigation system started working again... > > > > > -- > > > > Wulfraed Dennis Lee Bieber KD6MOG > > > > wlfr...(a)ix.netcom.com HTTP://wlfraed.home.netcom.com/ > > > > For some reason (that I don't get at all) lots of systems define long > > > as -180 <= x <= +180 degrees. > > > > Having the potential to alias a position seems like a bad idea for a > > > start, so when I've been coding such systems up, I've always spent a > > > bit of time making it convert positions into the range -180 <= x < > > > +180 degrees and using a proper ADT. > > > > I wonder if it was anything to do with that?... > > > I would have thought a longitude was really a mod 360, shifted by -180 > > for display purposes? For fractional degrees (i.e. minutes and > > seconds), make that mod (360 * 60 * 60), shift by -180 * 60 * 60 and > > split in degrees, minutes and seconds when displaying. > > > -- > > Ludovic Brenta. > > No...it's -180 <= x <(=) +180...always - check any map / globe!! > > Lat is always -90 <= x <= +90 deg - no doubt about that one :-) > > Adding "shifts" would make understanding any problem very hard... > > "So the position coming in is (+40, -100) but what's that inside the > code again????" Interesting to see that 'cyclic fixed point types' are under discussion: http://www.ada-auth.org/cgi-bin/cvsweb.cgi/ai05s/ai05-0175-1.txt?rev=1.2 http://www.ada-auth.org/ai-files/minutes/min-0911.html#AI175 Phil |