This follows from a previous posting I made about lack of a clean test machine for software installations. I'm doing a bad job of explaining how DLL dependencies work and how some machines might not have the right libraries at the time of installation. The problem is that it's being viewed as a defect with the build process. I'm trying to educate the higher ups that it's not the build process per se but rather the installation process which is to blame.
Here's a quote from my boss relating subcontractor work to our work to put it into perspective:
I'm not a software person. All I see is that when they hand something to us it just works but when we hand something to the client there's all sorts of problems. There must be something wrong with how you're building the code.
[edit]When the subcontractor applications arrive, they can be copied right onto our development machines because our machines contain all the right Microsoft runtime libraries. However, on the customer machines, not all of them contain these dependencies.[/edit] It's very easy to see how someone who is smart (scarily smart) could come to the wrong conclusion. So how would you explain the whole DLL dependency issue?