Question for US Rules committee on AH capability within LX NAV computers?

This statement that you have written a "patch" disabling LK8000 competition
mode is quite untrue.
You are only enabling back the TRI.
On our website we clearly state:

THE ONLY APPROVED AND GRANTED TO BE COMPLIANT VERSIONS FOR COMPETITIONS ARE
DOWNLOADABLE FROM THE OFFICIAL RELEASE AUTHORITY OF THIS SOFTWARE, WHICH IS
lk8000.it .

DOWNLOADING THE SOFTWARE FROM ANY OTHER SOURCE DOES NOT - DOES NOT - GRANT
ANY KIND OF COMPLIANCY.

THE CRC/MD5 CHECKSUM FOR THE EXECUTABLES MUST MATCH THOSE OF THE EXECUTABLES
RELEASED BY THE SOFTWARE AUTHORITY.

This is also required by the US RC, and I believe it is a good approach.
In fact, your faked version will not pass the CRC MD5 check, and the user
will be banned as a cheater for the rest of his life.

So your statement
Given the existence of this patch, contest organizers cannot be sure
whether a pilot's PNA runs an approved LK8000 version or a fake
full-featured version with my patch.

is not true. You must do something more than that: fake the CRC MD5
checksum, and make the TRI it fit inside the same number of bytes of the
code.
And by the way, the compiler generating the code for LK8000 has been
recompiled for the purpose, so I doubt you can regenerate the same code.

Too optimistic, I guess.
paolo



"Max Kellermann" ha scritto nel messaggio
...
Paul Remde wrote:
This implies that the AHRS is completely disabled and can't be
re-enabled for the 14 days.

What if I publish a proof-of-concept patch that adds a horizon to
LX8000 and others, circumventing this switch?

I think I can assemble one for LX8000/9000 (from LXNav) and the
upcoming LX Zeus (from LX Navigation) in a matter of a few days. It
will not be detectable without special equipment. And it will work
without the new AHRS hardware.

(Pilots interested in such an exploit may send me a private message)

If mainline XCSoar gets banned, I will demonstrate that most other
products must be banned, too. Better keep an old first-generation
electronic vario at hand when you attend a contest, to avoid surprises
;-)

I have already written a patch for LK8000 that pretends to be
"LKCOMPETITION" but doesn't actually disable the horizon:

...

PCool wrote:
is not true. You must do something more than that: fake the CRC MD5
checksum

It seems you do not understand the nature of CRC. "Faking" a CRC is
trivial, and is the most basic property of CRC. Better remove the
mention of CRC from your web site, it's embarassing!

Faking a MD5 checksum is not impossible nowadays, but still harder
than winning a contest. And not required at all for "cheating".

Now you tell me how the contest will verify the MD5 checksum
(practical example, not some theoretical contest organized by some
uber-geek), and I tell you how to get around it easily.

(There are enough catch-all cheats that not even the aforementioned
uber-geek will notice, but I'm curious how you imagine the
verification procedure will work in practice)

and make the TRI it fit inside the same number of bytes of the
code.

Don't be silly. That one is just as trivial as "faking" a CRC.

Max

Thanks, Professor.

"Max Kellermann" ha scritto nel messaggio
...
PCool wrote:
is not true. You must do something more than that: fake the CRC MD5
checksum

It seems you do not understand the nature of CRC. "Faking" a CRC is
trivial, and is the most basic property of CRC. Better remove the
mention of CRC from your web site, it's embarassing!

Faking a MD5 checksum is not impossible nowadays, but still harder
than winning a contest. And not required at all for "cheating".

Now you tell me how the contest will verify the MD5 checksum
(practical example, not some theoretical contest organized by some
uber-geek), and I tell you how to get around it easily.

(There are enough catch-all cheats that not even the aforementioned
uber-geek will notice, but I'm curious how you imagine the
verification procedure will work in practice)

and make the TRI it fit inside the same number of bytes of the
code.

Don't be silly. That one is just as trivial as "faking" a CRC.

Max

By the way,
http://git.xcsoar.org/cgit/master/xc...cpp?id=v6.2.6c
inside xcsoar 6.2 there is really a Horizon, and US RC require you to remove
that code even if unused, like we did.
Even if unused, the Horizon code is included by the Makefile
http://git.xcsoar.org/cgit/master/xc...ile?id=v6.2.6c

So the 6.2 version is not compliant to US RC rules.
Instead of trying to help people cheating, I suggest you release a
competition version of xcsoar too.

regards
paolo




"Max Kellermann" ha scritto nel messaggio
...
Paul Remde wrote:
This implies that the AHRS is completely disabled and can't be
re-enabled for the 14 days.

What if I publish a proof-of-concept patch that adds a horizon to
LX8000 and others, circumventing this switch?

I think I can assemble one for LX8000/9000 (from LXNav) and the
upcoming LX Zeus (from LX Navigation) in a matter of a few days. It
will not be detectable without special equipment. And it will work
without the new AHRS hardware.

(Pilots interested in such an exploit may send me a private message)

If mainline XCSoar gets banned, I will demonstrate that most other
products must be banned, too. Better keep an old first-generation
electronic vario at hand when you attend a contest, to avoid surprises
;-)

I have already written a patch for LK8000 that pretends to be
"LKCOMPETITION" but doesn't actually disable the horizon:

PCool wrote:
So the 6.2 version is not compliant to US RC rules.
Instead of trying to help people cheating, I suggest you release a
competition version of xcsoar too.

Thanks for the nice suggestion, but I suggest you actually read and
understand the code before drawing (the wrong) conclusions.

Max

I did.
I could read:
This feature of having a backup artificial horizon based on inferred
orientation from GPS and vario data is useful, and reasonably well
tested, but has the issue of potentially invalidating use of XCSoar in
FAI contests due to rule ref Annex A to Section 3 (2010 Edition) 4.1.2
"No instruments permitting pilots to fly without visual reference to
the ground may be carried on board, even if made unserviceable." The
quality of XCSoar's pseudo-AH is arguably good enough that this
violates the rule. We need to seek clarification as to whether this
is the case or not.

And this feature is included in the 6.2, it does not matter if disabled or
enabled.
The US RC is requesting that such code is not inside the executable.



"Max Kellermann" ha scritto nel messaggio
...
PCool wrote:
So the 6.2 version is not compliant to US RC rules.
Instead of trying to help people cheating, I suggest you release a
competition version of xcsoar too.

Thanks for the nice suggestion, but I suggest you actually read and
understand the code before drawing (the wrong) conclusions.

Max

+1 Max. Its a bit like the first attempt at communication with the alien mother ship of Close Encounters.

http://www.youtube.com/watch?v=tUcOaGawIW0

Its going to take some time to understand eachother!

Paul, I truly appreciate your post on the firmware version and your efforts to smooth the sands. The timing is (shockingly) perfect for LXNAV to release this news. I'll be happy when the USRC makes a specific statement about the requirement for special firmware to ensure that LXNAV's modern instruments AH capability is absolutely inaccessible, just as is required for the others. Firmware and app versions is a fairly weak method of enforcement, FWIW. That said, my posts are intended to engage a broader argument. I hope its starting to sink in a little. Its a pain in the ass for everyone! Fairness and consistency comes into play...

The inconvenience these AH ban rules are increasingly producing (for all of us...dealers, software and hardware manufactures, RC, contest staff and especially pilots who heaven forbid have bought updated systems) greatly outweigh the competition or safety value of banning the flood of AH type functionality in modern (racing level) soaring instruments. I assume the vast majority of US pilots do not cheat by cloud flying. Even with the existing rules...the intelligent cheater is going to EASILY be capable cheat no matter what is written. This is the fact that it most troubling.

It is definitely not enough, in my opinion, for LXNAV customers to "say" that they don't have the AH box installed. Its capability is FAR MORE THREATENING than, for example, mobile phone based systems. The rest of us (Butterfly, LK8000 and XCSoar) are required by USRC to build/install special versions of firmware and software for our products. If Paul's post is correct, add LXNAV to that list assuming the USRC requires it and enforces it. Great. But what does all this effort really gain us?

At a basic level, who is going to enforce the firmware and app versions and confirm them? This is more difficult than it sounds. I could easily switch my SD card on my phone for example with XCSoar. So could a young child. To be effective in any meaningful way these inspections need to be on a daily basis as firmware can be changed in 5 minutes and SD cards can be swapped out in seconds. Etc. The cheating pilot is far more creative than this ban and even strong enforcement levels can detect... The ban does literally nothing to prevent actual illegal cloud flying. It is a very minimal deterrent in general given the level of tools available.

So what is it going to be in regards to LXNAV? USRC required contest firmware version or daily box checks inside the panel? We cant have it both ways, can we? Who will enforce this at contests and to what level & frequency (daily I hope with random spot checks). Can competing pilots request to check other pilots gliders at contests if they suspect cheating?

I suggest that it should NOT be an honor system with LXNAV's $1700 AH system. The LXNAV system is an extremely capable product and should require a much higher standard than useless mobile systems which HAVE ALREADY HAVE BEEN REQUIRED by you to produce US Contest legal versions of their firmware and software.

Sean
F2

On Tuesday, April 3, 2012 10:26:15 AM UTC-4, Max Kellermann wrote:
Paul Remde
wrote:
This implies that the AHRS is completely disabled and can't be
re-enabled for the 14 days.

What if I publish a proof-of-concept patch that adds a horizon to
LX8000 and others, circumventing this switch?

I think I can assemble one for LX8000/9000 (from LXNav) and the
upcoming LX Zeus (from LX Navigation) in a matter of a few days. It
will not be detectable without special equipment. And it will work
without the new AHRS hardware.

(Pilots interested in such an exploit may send me a private message)

If mainline XCSoar gets banned, I will demonstrate that most other
products must be banned, too. Better keep an old first-generation
electronic vario at hand when you attend a contest, to avoid surprises
;-)

I have already written a patch for LK8000 that pretends to be
"LKCOMPETITION" but doesn't actually disable the horizon:

http://git.xcsoar.org/cgit/max/lk800...5859495c5818d2

Given the existence of this patch, contest organizers cannot be sure
whether a pilot's PNA runs an approved LK8000 version or a fake
full-featured version with my patch.

I'm not trying to support cheaters, I just want to make clear that
banning new technology is not a useful measure to prevent cheating.

Max

To Sean's point, will there be x-ray machines and random strip searches on the grid before each day's launch to ensure no one has a PNA hidden in their jockey shorts?

How is this for a logo for "not XC Soar" without the totally unusable 1 cm^2 "AH" box.

https://plus.google.com/photos/10776...52900309699361

Sean Fidler wrote:
How is this for a logo for "not XC Soar" without the totally unusable 1 cm^2 "AH" box.

https://plus.google.com/photos/10776...52900309699361

LOL, I would "+1" this if I had an account :-)