Is UML dead? & What is the future of UML?
March 9, 2009 8 Comments
Once upon a time UML is a hot buzz word in software development industry. If you know what is aggregation, abbreviation, etc. you will get job easily as an architect. Tools like Rational Rose are costly only a few company had that tools.Now UML silently dead. People were thought that UML based design will be better and it will make less rework during coding. Not many companies/developers complained about UML approach but they could not adapt on day to day development work. eventually UML died.
At the same time my intuition says UML is not dead but in future UML will come back with slight variance with difference buzz word. All it need is marketing people; most of the marketing people are busy with Agile process promotion.
Perhaps I encourage viewer to share their thoughts on UML’s future.
Venkat Sadasivam 
Hahahahaha….. u r right upto some extent… :-)
Venkat,
I don’t think UML is dead. I believe the industry still uses it but the buzz on UML has got significantly reduced. I believe this is mainly because of the maturity level today.. Knowing UML doesn’t mean that the person is good designer or architect, 2-3 years back that was the case when somebody say they know UML everybody in the top mgmt thought he is a good architect/designer but that’s not the case anymore. As u rightly mentioned in ur Design patterns post, Design or Architect skill is totally different to depicting/communicating it to others. UML & Design pattern helps u in doing the communication part and not on the thinking part. Its the individual ability to think which matters most. Industry has now understood that to decent extent and that’s the reason for less hype on that word.. I believe the same will happen Web 2.0 or 3.0 technologies in near future.
~Muthu
Muthu you are right the buzz is gone down. UML has be simplified to make it more practical and easy to adopt for day-to-day technical discussion.
Venkat,
It provides so many diagramatic representations out of which I find Sequence/Interaction & Use case diagrams can come in to our day to day activity. Deployment diagram is quite good for the admin person. Other than this I don’t find any major use of doing others. They are good for documentation purpose but on a day to day usage, I don’t think they are adding any value to us…
Is that what u mean by simplification? Having lesser representations…
~Muthu
As per the original idea of UML is model based design. But whereas we use UML to represent our completed design. As code/design evolves we cannot keep the model upto date always.
I found UseCase and class diagrams are useful. Sequence diagram and activity diagram I don’t like as we can do similar thing with simple flow chart notation.
I haven’t seen or found deployment diagram useful because most of the case architects tend to choose their own creativity to represent the deployment model.
Hi Guys, I’ve been working with UML for years and its always been a bit hard to sell because there is so much of it, it keeps growing and there’s no real methodology behind it. Having aid that people still draw use case diagrams and class diagrams. Activity diagrams are really a form of flow chart as Venkat quite rightly points out. I’m a business analysis and I find that use cases and activity diagrams are really useful for me and then I’ll do a database schema and everything else as variations on a class diagram, but I use the UML a lot and I use less than half of it! Most people only use the diagrams they need, it’s a modeling standard, not a religion, and I can’t see anyone with a competitive standard. That means I think things will remain much the same until a lot more people in IT start using modeling and UML diagrams, and I don’t know what would cause that.
UML is a common language for expressing key system perspectives to others. I find that UML is more recognized amont developers and architects than any other notation, so that in itself makes UML a valuable addition to any architect’s toolbox.
In my view the two primary reasons for the failed widespread acceptance of UML are as follows
1. UML tried to follow the philosophy of one size fits all or the panacea for all problems of software design and evolved into a behemoth of umbrella specifications covering diverse diagrammatic notations. For a large complex system this proved to be too cluttered and cumbersome for practical use.
2. UML does not add any value to the later stages of software development lifecycle such as code implementation or testing. It is only good for static or structural design. It does not in any way help in understanding the internal working of the system like the way CAD/CAM designs help in understanding the internal working of a mechanical systems.