design - Is it worth it to model applications with UML/other diagrams? -
i mean uml diagrams method works viable. had subject @ university modelling uml , seemed more effort actual benefit realise might because i've never created huge complex system. worth while , types of diagrams/methods usually* best?
*of course, many times concrete tools need chosen concrete problems maybe there patterns.
firstly
that clever uml question in stackoverflow see far :-)
it seems have no uml fever :-) welldone ...
check following link , read grady booch comment @ end of paper death uml fever
secondly
mda (model driven architecture) - executable uml utopia now. have limited area such system can modeled state machines etc. tools , techniques not mature enough.
so still dream produce diagrams producing working software "abracadabra" [ tool vendors still try sell ]
finally
code [source code] still king.our actual models not uml or other diagrams, our real models source code write , execution. if use old analogy real architect's blueprints not uml or other diagrams, blueprints map our source code.
so modeling bad? death?
of course not dead , bad. not practical-efficient-effective model whole systems on table coding laborious activity.
so do?
uml , other diagrams give visual thinking tools. apply difficult-tricky part of design explore alternatives ,to judge design , communicate design ideas others.do not apply whole system, every details.
which diagrams best?
depends on context. drawing uml diagram not free painting activity. if develop uml diagram should have purpose-reason : thinking problem, communicating ideas team members, try judge design, explore alternative designs etc.
but careful understand system should know dynamic behavior static structures[ classes , structural relationships in class diagram]. not forget use dynamic behavior diagrams [ such sequence diagrams show messaging between components ] when necessary.
Comments
Post a Comment