|
|
Forwarded by popular demand (Ed'd popular and he demanded it 8*)
{Several typos corrected}
-------- Original Message --------
Applaud! Applaud! Great insight, Ernest. Should post this on the Rotary list just for our general info.
Ed
Ed Anderson
Rv-6A N494BW Rotary Powered
Matthews, NC
eanderson@carolina.rr.com
----- Original Message ----- From: "Ernest Christley" <echristley@nc.rr.com>
To: <aeroelectric-list@matronics.com>
Sent: Monday, June 26, 2006 10:44 AM
Subject: AeroElectric-List: Re: AeroElectric-List Digest: 26 Msgs - 06/24/06
--> AeroElectric-List message posted by: Ernest Christley <echristley@nc.rr.com>
AeroElectric-List Digest Server wrote:
The AI NEVER stabilizes there is a slow tumble until I see either nothing but sky or ground. NEVER STABILIZING!
Has anyone ever run across this problem and if so what was the cause and cure?
It has been suggested that I perform two (2) operations:
1 - Upgrade the Software from Ver 2.21 to Ver 2.64
2 - Hook up the keyboard and do a Warm System reboot from the keyboard
What are your thoughts? Does this sound like a cure? Of course the software upgrade is required and will be done.
Barry
"Chop'd Liver"
I work in the software industry as a Quality Assurance Engineer, ie. I test software. I've written a lot of software. I've met a lot of software engineers. I've worked for several software companies. There is a wide range of attitudes when it comes to software.
At one end of the spectrum is the companies/engineers that treat software and the requisite hardware as an appliance. The software is tied intimately to the hardware, and a breakage in one signifies a breakage in the other. A breakage in either signifies a breakage in the organization. Code is meticulously maintained, and every last detail of the software/hardware combination is tested as much as humanly possible before shipping to the customer. The engineers tend to be very experienced and 'slow moving'. The code is not expected to break when delivered to QA. The best example I have personal experience with is IBM's mainframe networking division.
The other end of the spectrum is what I like to call "The Microsoft Generation". Code is whipped out using long tool chains built on top of libraries written by someone else. I meet lots of these engineers building database frontends to generate reports to management. The criteria for these engineers is that they generate a disposable app quickly and move on to the next project. They are conditioned to value a new glitzy widget over solid engineering. Code unit test is an afterthought at best, and if it does happen it consist of being able to generate a report on a couple sets of data at most. These guys deliver stuff to QA with the idea that the test team will tell them what is wrong with the code. I call it the Microsoft Generation, because that company has been the forerunner in preaching that hiring competent, seasoned engineers is uneccessary with their software. Software is complex and should be expected to break, they preach, but their tool set will enable a cheap college grad that works around the clock for Coke and pizza to spit out polished applications to run the enterprise. Widgets, "new features" and just more eye-candy can be added ad nauseum, quickly and easily.
The Microsoft Generation is OK for generating reports to management. I'll even abide letting them build a video game or two. But I want real engineers writing any software that I will be letting my butt ride on.
I talked to Blue Mountain and Dynon at Sun-n-Fun. The question I had in mind was, "If I were an QA engineer at this company, what would the development team deliver to me for testing?" The Dynon unit booted quickly and showed a simple display. I got the "feel" that the software was written for the hardware, and the hardware was designed for the software. The BMA unit seemed to take forever to boot up, complete with splash screen to keep the user occupied reading copyright information while it did a digital dance behind the scenes. (BTW, a "splash screen" is a red flag that someone from the Microsoft Generation is behind the scenes. How is it helpful, except to show more eye-candy? And copyright? What am I going to do, run the stuff on my PC?) I got the distinct feeling that BMA engineers would expect me to tell them what was wrong with it, while the Dynon folks would only expect me to verify that it works as they designed it.
My thoughts? The fact that the BMA is designed for "quick upgrades" is a glaring red flag. The thing is a limited function device. It should work out of the box. The in-field "quick upgrade" tells me that BMA is using you as a beta tester. The fact that it doesn't work out of the box tells me that either the hardware or software is broken. Being that this is the real world, I can accept hardware being broken. Things break in shipping...not every IC is tested off the assembly line..etc. But the fact that a 'software fix' is available gives me the thought that the development organization needs a fix.
-- ,|"|"|, Ernest Christley |
----===<{{(oQo)}}>===---- Dyke Delta Builder |
o| d |o http://ernest.isa-geek.org |
--
,|"|"|, Ernest Christley |
----===<{{(oQo)}}>===---- Dyke Delta Builder |
o| d |o http://ernest.isa-geek.org |
|
|