Page 78 - Software and Systems Requirements Engineering in Practice
P. 78

50   S o f t w a r e   &   S y s t e m s   R e q u i r e m e n t s   E n g i n e e r i n g :   I n   P r a c t i c e



                                                Bicycle
                                               Attractive to
                                               Customers
                                    +
                                                   +                    +
                         High Quality           Low Cost            Low Weight
                              +                                        +
                                                   –
                                                Titanium
                                                 Gears

                      FIGURE 3.5  Simple goal model fragment


                      the optimal goal set can be calculated. However, the reality is that the
                      contribution of many high-level requirements cannot be calculated for
                      a variety of reasons, including changing demographics, rapid shifts in
                      technology,  etc.  Sometimes,  difficulties  associated  with  conflicting
                      goals are not recognized until the requirements have gone through a
                      complete review cycle. The refinement of nonfunctional requirements
                      can  bring  to  light  issues  that  may  otherwise  remain  hidden.  The
                      importance  and  impact  that  nonfunctional  requirements  can  have
                      warrant their consideration and elicitation as early as possible in
                      the product development cycle.
                         Goal  models  can  be  as  simple  or  as  complex  as  necessary.
                      Figure  3.6  shows  some  of  the  goals  for  a  nuclear  power  plant
                      simulator.  Such  simulators,  mandated  by  regulation,  are  used  to
                      train  the  operators  of  nuclear  power  plants  and  must  have  high
                      fidelity and reliability. The figure shown identifies quality assessment
                      methods, or QAMs, that are used to determine how well the business
                      goals meet the desired quality [Cleland-Huang 2005]. For example,
                      QAM 5 states that when any action is taken, the simulator indicator
                      light response shall be within 200 milliseconds of the response in the
                      real plant. That is, if a button is pressed in the power plant closing a
                      valve and an indicator light comes on in three tenths of a second,
                      then in the simulator, that light must come on within three to five
                      tenths  of  a  second.  The  actual  QAM  was  evaluated  by  randomly
                      connecting  an  oscilloscope  to  button/light  pairs  (there  were
                      thousands of such pairs) in the simulator and determining that the
                      response was within specification by measuring the step wave on
                      the oscilloscope. Goal models with QAMs can be used as checklists
                      to ensure that important nonfunctional requirements have not been
                      overlooked.  If  a  QAM  cannot  be  defined  for  a  nonfunctional
                      requirement, then it may not be possible to test that the requirement
                      has  been  met,  and  the  requirement  should  then  not  be  part  of  a
                      contract or requirements specification, as it may not be feasible to
                      implement.
   73   74   75   76   77   78   79   80   81   82   83