Software Engineering

14.3.1 
Evaluate requirement alternatives
Requirement trade studies should be conducted in the area of requirements to 
ensure that stakeholders and the software engineering team comprehend the impli-
cations of every requirement. There are several motivations for conducting a trade 
study within the requirements domain:
1. 
Requirements as a communication mechanism
. The language used to express 
a requirement is ambiguous by design. Language is inherently ambiguous to 
permit the context in which a term is used to define its meaning. Whenever 
there are formal expectations concerning a product or service, the agreement 
that specifies the deliverable must not be ambiguous. Thus, establishing a clear 
understanding of stakeholder needs and expectations is vital to establishing an 
achievable software requirements baseline.
2. 
Stakeholder needs and expectations may involve conflicting demands
. 
Stakeholders become absorbed with their particular needs and must be made 
to understand how individual demands conflict with other’s needs. To resolve 
stakeholder conflicting demands, it may be necessary to conduct experiments to 
derive a proper balance among stakeholder needs.
3. 
Software products automate business and operational processes
. While stake-
holders may be familiar with their processes, the semantics used by stakeholders 
to describe a process may be overloaded with industrial terminology unfamiliar 
to software personnel. Operational models provide a mechanism to encapsulate 
and translate these process descriptions into a representation comprehensible to 
software-literate personnel.
4. 
Requirements may overstate performance demands
. Requirements impose level-
of-performance demands on data processing transactions. It may be necessary 
to establish operational models to assess the reasonableness of stringent perfor-
mance expectations. The availability of software and computer technology to 
achieve performance expectations must be substantiated by analyzing alterna-
tive operational or computational approaches.