So right now I am a software guy for a fairly major -atleast major to me- project that could become an international product come February. They want this to be deployed in around the US and Canada. Because I don't want to mention what exactly it is I'm making, I won't say anymore than: It's designed to save dozens of lives per year, and the more that are installed the fewer fatalities per year will be seen. This sucker is supposed to save lives. Because its embedded, Its just has to work. No help from me, its not like a school project. Its a full blown Product at the end. It-has-to-work.
This is where I'm getting a little nervous. There is only one software guy -me-. Which means all the behaviour that this thing does, all the implementation, is going to be done by a fresh out of undergrad EE. With no real world programming experience. I know my stuff, and I think I can get it to run. ...but get it to run perfectly for years at a time without ever faulting once? Shit I'm happy when my little projects run minutes at a time without screwing up.
What I'm getting at here is that, in the case of safety critical programming. It should really be done as a team, of at least 2 qualified people. My way of doing things might work, but might not be fail safe-enough.
If you knew the device that might save your life was tied back to some 22yr old punk E.E. who did this himself in his room late nights on weekends. Would you trust it???
Worries me is all.
-DK
No comments:
Post a Comment