In-Depth
What’s next?
- By Deborah Melewski
- June 18, 2001
Meanwhile, the UML RTF has been busy this year. The activity does not indicate, however, that the specification is unstable, noted Kobryn. "The scope of the RTF's responsibility is to fix things that are broken and to clarify things that are currently unclear," he said. "If we determine that something is outside the scope of a minor revision, but find that it has merit, we will make a recommendation for further RFPs."
The RTF has been systematically going through UML issues, which can be submitted and viewed by UML users at the uml.shl.com URL address. Approximately 400 issues have been received. Areas in behavioral work - for example, activity modeling and collaborations - are being finalized. The next technical revision, UML V1.3, will be out in November.
Many organizations are just now get-ting their feet wet with the UML standard and have not used the full notation or all of its diagrams. Many are finding a definite learning curve associated with its implementation. Are perceived complexities keeping companies from adopting UML?
That could be the case, according to Greg Schottland, president of Advanced Software Technologies Inc., Highlands Ranch, Colo. "The UML is not a simple notation by any stretch, it's very complex," he said. "We've seen tremendous value for customers using UML, but that needs to be promulgated. The number one goal should be to get the message of its value out to people, and to take away the complexity.
"For example, UML provides a common language to capture and document thoughts and designs," continued Schottland. "Users need palatable solutions that are easy to learn and work with. The rest of the UML is valuable, too, but most people aren't ready for it yet."
"Some of the UML model elements may be overkill and unnecessary for a particular problem," said EBG Consulting's Gottesdiener. She noted, for example, that there is no need to do state diagrams for most applications, only for those at a very high level.
"It can be overwhelming for those who don't have a computer science background or those who aren't working in realtime," she said, adding that a lot of the models are also geared to software engineers, with nothing at the business level.
Admitted Booch: "If a person today picked up the semantic document of UML as submitted, they would find it very complex. In practice, though, it is not different than how people deal with languages like Java or C++."
Booch explained that UML is very expressive, and that certain subsets can be used where applicable. "Given a typical development team, one person can't be an expert on everything," he said. "Technologies are too diverse, for example, so each person may use different diagrams."
MCI Systemhouse's Kobryn agreed, but added that UML users will typically progress through three stages: walking, jogging and sprinting. "It's a natural progression," said Kobryn. "A typical starting point is static structural modeling. The user may then move up to performing dynamic behavior, and then progress to model management."