Rafal...Great post even if I am nearly 2 years late to share.
Over the years as a software architect consultant, I have discovered that UML has become "nearly" obsolete as a standard tool to communicate architecture. In the early days (I am dating myself), we used UML to demonstrate an understanding of the business requirements, and then to share intent with software engineers. It also served as documentation. Today, with Agile, Scrum, storyboards, and the like, it appears the input energy required of UML doesn't warrant what it can output.
Today, we use UML for new client projects for the purposes stated above, to then feed into our project generator to create the first pass of an application and all it's other bits. We also use existing UML created for legacy apps (eg: JEE and 4GL apps, etc..) to assist in generating the groundwork in app refactoring.
I am not sure it will see a revival, but we have discovered value in UML when it serves as input to a larger more valuable output.
I'm focused on developing and expanding my knowledge and skills. Enjoying new challenges. I'm assuming that there are no stupid questions, there are only silly answers.
Rafal...Great post even if I am nearly 2 years late to share.
Over the years as a software architect consultant, I have discovered that UML has become "nearly" obsolete as a standard tool to communicate architecture. In the early days (I am dating myself), we used UML to demonstrate an understanding of the business requirements, and then to share intent with software engineers. It also served as documentation. Today, with Agile, Scrum, storyboards, and the like, it appears the input energy required of UML doesn't warrant what it can output.
Today, we use UML for new client projects for the purposes stated above, to then feed into our project generator to create the first pass of an application and all it's other bits. We also use existing UML created for legacy apps (eg: JEE and 4GL apps, etc..) to assist in generating the groundwork in app refactoring.
I am not sure it will see a revival, but we have discovered value in UML when it serves as input to a larger more valuable output.
Hi Steven,
Thank you for your comment. I agree that the entry threshold is often too high, and sometimes using a UML diagram is simply an overkill.
BTW I'm glad that almost two years after the post has been published is still alive 😉