Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.misc > #17647 > unrolled thread

McDonald's Kiosk failure

Started bySylvia Else <sylvia@email.invalid>
First post2019-03-28 15:48 +1100
Last post2019-03-28 16:21 +0000
Articles 4 on this page of 24 — 12 participants

Back to article view | Back to comp.misc

Contents

  McDonald's Kiosk failure Sylvia Else <sylvia@email.invalid> - 2019-03-28 15:48 +1100
    Re: McDonald's Kiosk failure mm0fmf <none@invalid.com> - 2019-03-28 07:37 +0000
      Re: McDonald's Kiosk failure Roger Blake <rogblake@iname.invalid> - 2019-03-31 00:19 +0000
        Re: McDonald's Kiosk failure Adrian Caspersz <email@here.invalid> - 2019-04-01 18:16 +0100
          Re: McDonald's Kiosk failure Adrian Caspersz <email@here.invalid> - 2019-04-01 18:18 +0100
          Re: McDonald's Kiosk failure Roger Blake <rogblake@iname.invalid> - 2019-04-01 18:02 +0000
            Re: McDonald's Kiosk failure Adrian Caspersz <email@here.invalid> - 2019-04-03 13:37 +0100
              Re: McDonald's Kiosk failure Roger Blake <rogblake@iname.invalid> - 2019-04-04 01:20 +0000
          Re: McDonald's Kiosk failure Eli the Bearded <*@eli.users.panix.com> - 2019-04-04 20:57 +0000
    Re: McDonald's Kiosk failure Andy Burns <usenet@andyburns.uk> - 2019-03-28 08:27 +0000
      Re: McDonald's Kiosk failure Sylvia Else <sylvia@email.invalid> - 2019-03-28 22:55 +1100
    Re: McDonald's Kiosk failure Richard Kettlewell <invalid@invalid.invalid> - 2019-03-28 08:39 +0000
      Re: McDonald's Kiosk failure Sylvia Else <sylvia@email.invalid> - 2019-03-28 22:56 +1100
    Re: McDonald's Kiosk failure Adrian Caspersz <email@here.invalid> - 2019-03-28 10:43 +0000
      Re: McDonald's Kiosk failure Sylvia Else <sylvia@email.invalid> - 2019-03-28 23:00 +1100
        Re: McDonald's Kiosk failure Adrian Caspersz <email@here.invalid> - 2019-03-28 12:38 +0000
          Re: McDonald's Kiosk failure Sylvia Else <sylvia@email.invalid> - 2019-03-28 23:47 +1100
        Re: McDonald's Kiosk failure kludge@panix.com (Scott Dorsey) - 2019-03-28 10:02 -0400
          Re: McDonald's Kiosk failure Huge <Huge@nowhere.much.invalid> - 2019-03-28 14:17 +0000
          Re: McDonald's Kiosk failure Ivan Shmakov <ivan@siamics.net> - 2019-03-28 14:30 +0000
        Re: McDonald's Kiosk failure Andy Burns <usenet@andyburns.uk> - 2019-03-28 15:30 +0000
    Re: McDonald's Kiosk failure Dan Espen <dan1espen@gmail.com> - 2019-03-28 09:45 -0400
      Re: McDonald's Kiosk failure Rich <rich@example.invalid> - 2019-03-28 14:18 +0000
        Re: McDonald's Kiosk failure Adrian Caspersz <email@here.invalid> - 2019-03-28 16:21 +0000

Page 2 of 2 — ← Prev page 1 [2]

#17663

FromAndy Burns <usenet@andyburns.uk>
Date2019-03-28 15:30 +0000
Message-ID<gg47koFphkiU1@mid.individual.net>
In reply to#17654
Sylvia Else wrote:

> The kiosk BSofD seemed just a symptom of the wider problem

Seems the ELO touchscreens have an API, so maybe they use that instead 
of just mouse emulation, and with no network it sent rubbish values that 
crashed the driver?

[toc] | [prev] | [next] | [standalone]

#17658

FromDan Espen <dan1espen@gmail.com>
Date2019-03-28 09:45 -0400
Message-ID<q7ij5t$7q3$1@dont-email.me>
In reply to#17647
Sylvia Else <sylvia@email.invalid> writes:

> My local Maccas was temporarily unable to take orders today because
> there was some kind of upstream network failure. I took this
> photograph of one of their kiosks.
>
> https://www.dropbox.com/s/lsmg6ux84svt2ne/WindowsCrash.jpg?dl=0
>
> What does this image tell us?
>
> 1. McDonald's is using Windows to run a system that takes orders for
> burgers and coffee.
>
> 2. Said system cannot handle an upstream failure without crashing due
> to problem with a driver.
>
> 3. The driver appears to be related to the touch screen.
>
> <https://www.scribd.com/document/254188226/Elo-Touch-Solutions-Multi-Touch-Driver-User-Manual-6-4-0>
>
> 4. Somehow, and upstream failure can break a touch screen driver.
>
> The mind boggles.

I don't find this all that mind boggling.

This could be 2 things failing at once, or one thing failing, driving
the software into a poorly tested error path to show the second error.

Many times, I'd see an error occur creating a cascade of other things
failing.  Things that you'd think were unrelated.

-- 
Dan Espen

[toc] | [prev] | [next] | [standalone]

#17661

FromRich <rich@example.invalid>
Date2019-03-28 14:18 +0000
Message-ID<q7il2o$mdh$1@dont-email.me>
In reply to#17658
Dan Espen <dan1espen@gmail.com> wrote:
> Sylvia Else <sylvia@email.invalid> writes:
> 
>> My local Maccas was temporarily unable to take orders today because
>> there was some kind of upstream network failure. I took this
>> photograph of one of their kiosks.
>>
>> https://www.dropbox.com/s/lsmg6ux84svt2ne/WindowsCrash.jpg?dl=0
>>
>> What does this image tell us?
>>
>> The mind boggles.
> 
> I don't find this all that mind boggling.
> 
> This could be 2 things failing at once, or one thing failing, driving
> the software into a poorly tested error path to show the second error.

This..  Sadly, much software (and nearly all 'enterprise' software, as 
this McD system would fit) is seldom tested by simulating error 
conditions to then test the error recovery paths.

So, over time, as band-aids for bug fixes/new features get added, one 
ends up with a codebase where if an error is returned from a call that 
normally never returns an error, that the error path itself encounters 
another error it can't recover from, and the whole house of cards comes 
falling down.

Or, it could be that some of this McD system was implimented by their 
contract Indian coders the same way the new enterprise software at work 
was built (also by contract Indian coders): the API calls to legacy 
systems have never return an error while we were prototyping, so we 
don't bother to even check if they returned an error during production.  
Much fun then ensued the day one of the legacy systems was down for 
maintence and the front end for the API call returned an error code to 
that effect.

[toc] | [prev] | [next] | [standalone]

#17666

FromAdrian Caspersz <email@here.invalid>
Date2019-03-28 16:21 +0000
Message-ID<gg4ajnFq6gdU1@mid.individual.net>
In reply to#17661
On 28/03/2019 14:18, Rich wrote:

> 
> Or, it could be that some of this McD system was implimented by their
> contract Indian coders the same way the new enterprise software at work
> was built (also by contract Indian coders): 

The company in question is probably this bunch. Not Indian.

https://www.rdisoftware.com/



-- Adrian C

[toc] | [prev] | [standalone]

Page 2 of 2 — ← Prev page 1 [2]

Back to top | Article view | comp.misc

csiph-web