|
A
reader from Finland who wishes to remain anonymous wrote...
"How
many of you right now have a half thumbled book on UML lying on
your desk" was a punchline that made me read through your article
with care ;) As you wrote in uidesign.net article there's ongoing
progress to start using UML with rational rose.
I'm
working currently as a usability architect and I just read a book
"Use case driven object modeling with UML" and your article. I'm
of the impression that people are trying to save my company's software
process by applying uml+rose to design software. I'm also worried
it would make UI design retired. How would UI design fit into a
process described in a book I just read? I'd need some instructions
and material to find out more
uidesign.net
reply...
A
very good question. You will note that my article is now 21 months
old but it still resonates as "new" and important. As luck would
have it I have been sitting in a Rational class today on the RUP
and Use Cases. [ My staff are learning RUP and OO modeling.] I have
to conclude that the movement towards a more rigorous [RUP] process
which includes the notion that you elaborate Use Cases using Activity
diagrams, then build the activity diagram as a Control class, as
part of a model consisting of Boundary, Control and Entity classes,
does indeed make UI design redundant.
Rational have continued to ignore UI design as a discipline and
ignore the creative side of the web experience. This is primarily
because they have no expertise in that area and are intent on selling
tools. So they continue to pretend that it is not important and
focus their sales efforts on technical minded geeks.
In RUP, various Rational authors suggest that, the UI design must
be completed prior to the Use Cases being written and they use this
as an excuse to say that UI design is being considered in the process.
However, elsewhere in the process they stress the iterative nature
of the work and the flexibility of the RUP to cope with this iterative
nature. However, it is true to say that Rational are advocating
a purely waterfall method for UI design, as the UI design cannot
change as soon as the Use Cases are written.
The mechanism of Boundary, Control and Entity classes is in my opinion
truly the death of UI design within the software engineering profession.
You have prompted me to write a new and overdue article probably
entitled, "RUP and the death of UI design - Revisited" which I will
be publishing soon.
I
actually met one of the key players in this new Boundary-Control-Entity
model at the UML 2000 conference a few weeks ago. We met at a UML
Workshop on Web Application Architecture. I have to say that we
disagreed on what was acceptable as good design and I was struck
by how much the Rational people simply don't "get it" where UI design
is concerned.
I recently learned that Rational have submitted the new Activity
Diagram, Boundary-Control-Entity Class model to the OMG as a proposal
for the "official" way to model Web Applications. Clearly this is
wrong. It completely locks out the Interaction Design and the Information
Architecture communities from web design. I intend to use my vote
as a "contributing member" of the OMG to vote against this Rational
proposal. However, I have to say that the OMG is dominated by Rational
and its influence and therefore the proposal is likely to be accepted
regardless of my views.
If it makes you feel any better, I can offer that my own employer
is also going through the same process of adopting Rational Unified
Process and I too fear that it is the death of UI design.
Cheers
David Anderson
Editor
|
