I know lots of you are in IT, so probably already very familiar with The Daily WTF (for the stunningly squeamish, it stands for "Worse Than Failure.") During an archive binge I discovered the Hangar Management System story. Maybe I'm underestimating the cost of writing enterprise software, but when you look at the fines, reaccommodation and ferry costs, the accumulated delays of Klaus and his minions looking up all the procedures in unindexed PDFs, then looking them up again in the update manuals must warrant a real system. When accessing that information is an onerous multi-step procedure and time is a factor, how can they not be tempted to not look everything up. Assuming you know the correct part or procedure on an airplane can have even more disastrous consequences. There were no deaths or passenger injuries in that incident, but it could easily have been a hull loss with no survivors.
The fact that a visiting college student was the one who discovered that the buzzword compliance of the hangar management system went only as far as the folder names suggests that there isn't a high level of interest in either efficiency or safety there. You can see mechanics sometimes working on airplanes with little CD readers: you'd think they were catching up on their netflicks queue until you saw that today's feature was the technical specs of the airplane. Sounds high tech and expensive, but I imagine the time saved makes up for the cost of the devices pretty quickly. They could have enhanced reality goggles for instructions on carrying out tasks. They could have built in checklist management, parts inventory and timekeeping functions, too. Aviation is an industry where time is a lot of money and mistakes are lives, so it's not beyond the bounds of reason. I won't be too surprised if someone tells me that such a system is already in use for the B787. But sadly I won't be too surprised if a reader has firsthand knowledge of the horrific code lurking inside modern avionics.
5 comments:
I love TDWTF. I can't speak directly to avionics software, but I think most people outside of IT would be shocked to learn how ridiculously poor most commercial software is (including medical systems which are arguably in the same class -- that I do know from firsthand knowledge). Mostly because you can rarely actually "see" software any meaningful way.
Imagine your HSI being driven by a myriad of bulky, complex and sometimes unused or non-functional gears and linkages -- all constructed by different mechanics working for different companies who had no idea what the other ones were doing or what the overall goal was. That's pretty much where most software is at.
I can only hope aviation software is better than everyone else's seems to be.
@Anonymous
Given how difficult it must be to get changes certified, I can only imagine that Avionics software is probably much, much worse.
When someone who knows their stuff comes along and sees the "mess" it must be awfully hard to get permission to fix it -- even if it is obviously wrong. Think of all the paperwork it would take.
Agreed with the comment that most enterprise software is pretty bug riddled.
Now, in this case, we aren't really talking about software that puts peoples lives at risk. We're talking about software that organizes the garage. So it should be a simple matter of rebooting the iPad if the MyFBO app goes down and you can't access all the records and manuals you want to with a swipe of the finger.
The sad thing is, even though the above is true, there is another truth (at least in the US). Once something bad happens the fingers will point until they find someplace to stop. The passenger is going to point at the pilot, who is going to point at the owner, who is going to point at the mechanic, who is going to point at, you guessed it, the software provider.
How legitimate the case is and where the fault really lies is almost secondary. The cost of insuring the software provider, coupled with the fact that most maint organizations are not really high-tech operations, means that the cost of creating that software business is close to, or more than, the revenue that can be expected.
I'll put it another way: I've been in IT for over 20 years, have started four companies and am thus arrogant enough to think I could probably start another one. I wouldn't touch this idea with a ten foot pool even though I think it *should* happen.
There do exist sophisticated parts-tracking systems, so my ranting was directed at the one airline rather than the industry in general. Using the wrong part really can be life or death.
Funny title for mission critical s/w? I mean WTF? (sorry).
RW2:
Why would this system have to be certified? There is a parallel in FDA systems for IND management. There are companies which just provide the "file cabinet" management software for pieces of paper that need to be signed by humans. Without getting to 21CFR11 standards for electronic signatures, they exist (and profit) in a highly regulated environment.
Why not make a document management system that just gets stuff scanned into it and can produce documentation on demand? Failure of the system just means the document won't be produced. It's still up to the A&P or IA to make sure the document meets the regs. No different than a microfiche reader in a way.
Take it on-line and you might make some money (or have some fun). Use an open source DB/OS/etc system architecture for low overhead.
I'm thinking GA here. Part 135, 91. Not 121. MyFBO is already doing some of this stuff tracking pilot records and I can't imagine it's certified.
I'm just thinking out loud...
Actually a spark of an idea...
Aviatrix, you're great.
--paul
Post a Comment