Software Engineering Stack Exchange is a question and answer site for professionals, academics, and students working within the systems development life cycle. It only takes a minute to sign up.
Sign up to join this community
William Cook in a tweet wrote that:
"UML is the worst thing to ever happen to MDD. Fortunately many people now realize this ..."
I would like to know the reasoning behind that claim (apparently, I'm not referring to his personal opinion).
I've noticed that many people out there don't like UML that much. Also it is worth mentioning that he is in academia, where UML is preety much the holy grail of effective design and modelling.
Well, I'm the academic who posted the original tweet. Tweets are not meant to be scholarly articles. They are advertisements, and I think they can also be controversial. Here are my follow-up tweets:
1) UML was created to model OO designs. It effect you are modeling the code of a system, not the system's behavior. UML is at wrong level.
2) the idea that 7 (or 13) diagram formats in UML can cover everything is crazy. What about GUIs, web wireframes, authorization, etc. ???
3) UML has encouraged the idea that models must be graphical. Ridiculous! Text and graphic models are both useful and often interchangeable
4) UML is at once too large and complex and at the same time very limiited. Stereotype and profiles are not effective for usable extensions.
Note that I'm not necessarily saying UML is bad. I'm simply saying that it is not helping the goal of "model-driven development", which is what I'm interested in. I don't understand the comment about "holy grail".
UML is the equivalent of taking a screwdriver and a hammer and taping them together and calling it a "Universal Fastening Tool." In theory it can be used to represent a ton of things in great detail, in practice its a bunch of poorly integrated tools claiming to be a single tool, that makes doing any one task far more difficult than having a proper tool to begin with.
I think there's also a case to be made that MDD is the worst thing that happened to UML (why else would we have the UML2 that we have?), but ignoring that for the moment...
MDD = Model Driven <Design|Development>. The idea is to be able to develop solutions at a level of abstraction appropriate to the problem domain - that is, it is an attempt to express solutions to problems in the most natural syntax for expressing those solutions. The problem domain itself is characterized with an operational model (that is, by a model that can be executed by computer). So, MDD can be a very attractive approach, albeit one with two major requirements:
It's my understanding that the UML2 effort was intended to address point 1, probably under the belief that industrial experience with UML showed that point 2 was satisfied for some large subset of problem domains. Unfortunately, and this is what I think William Cook was getting at, UML does not satisfy point 2 for anywhere near the scope of problems that was thought. I don't speak from personal experience, but I think the industrial experience of using MDD with UML has two common outcomes:
In either case, the promise of MDD is unfulfilled. UML might be considered the worst thing that happened to MDD because it occupied the attention of MDD tool developers to the exclusion of models that might actually work (albeit for a smaller set of software problems).
UML is great as long as it is just a modeling language. If you try to connect MDD to UML in order to get a graphical view then it is useless. MDD would be great without UML as well as UML without MDD.
Let's say UML and MDD have divorced today in order to have a better life not together anymore :-)
