The Model Driven Software Network
Raise your level of abstraction
This is an interesting survey and I agree that your conclusion about adoption of UML in the early phase of the lifecycle seems accurate. Given the result, I am surprised that class diagrams are one of the most commonly used notations, I can only assume the results show class diagrams being used during the analysis phase. Maybe the survey should be repeated with an additional question about where in the lifecycle the elements of UML are being employed.It would appear that UML is not being fully adopted during the detailed design and implementation phase. I wonder why?At Objektum we find significant benefit in terms of time and cost, when adopting UML to drive our code generation. In fact all our development uses UML and code generation techniques which improve our productivity, quality and overall maintainability. I wonder if the reason that others are not adopting a similar strategy is owing to a lack of training in code engineering as well as misconceptions about the maturity of modern UML tools.I am confident that if other developers adopted an approach similar to our own, we would see a significant shift in the survey result. I would be interested in hearing from others on the use of code generation, particularly those who have not had a positive experience.
Add a Comment
Comment by Andreas Leue on February 15, 2011 at 19:07 > What do you make of the survey?
Figure 1 is more or less what I would have expected. Figure 2, class diagrams, too. I mean, "UML == class diagrams" is common notion. I'm surpirsed by use cases & activity being next. Activity? I thought everybody uses BPMN (not that there's a huge difference, but I thought people think so). Use cases might be due to agile/user stories or formal specs - would be interesting to get more details. Figure 3 might be just that people really use "UML == class diagrams"; and when asked about expertise they remember that there are somehow a dozen more diagrams and they have next to no idea how they work.
Comment by Derek Russell on February 15, 2011 at 13:02 Hi Jordi,
Thanks for your comment. I agree, it is impossible to draw a scientific conclusion but I think the survey mirrors what we are seeing.
Bob says of the survey:
Responses were solicited from numerous online groups / people. The study was promoted on related LinkedIn groups, Twitter, my blog and website. This is by no means a scientific study. There were 44 respondents from the various online groups that would have some experience / interest in this topic. However, subsequently I have had multiple people contact me and tell me that their research, or the research of others, aligns with these results.
Kind regards,
Derek
> What do you make of the survey?
Without info on the number and background of the participants is difficult to draw any useful conclusion from a survey.
© 2012 Created by Mark Dalgarno.
Powered by
You need to be a member of The Model Driven Software Network to add comments!
Join The Model Driven Software Network