There is well known that bad/poor/weak software requirements cause resources/time/money overruns (according chaos reports and others).
I guess we have a problem,
What is the error when eliciting, specifying, analyzing and validating the requirements?
I think there is a knowledge transfer problem (among others), we are failing when passing business knowledge (user needs/beliefs/etc) to a software specification and thus to a software implementation.
I made a little survey to see the software requirements process as a knowledge transfer process, the survey can be answered anonymously if you want, the survey is available at
https://docs.google.com/forms/d/1NaBFQzfSHSpqNEjKH921-iFPRXBnnykKWI6SH3ysjpA/viewform
Thank you very much for your feedback.