The WebUML initiative wants to set UML free , that is, to release it from the desktop and bring it to the world wide web, where online technologies enable a more social way of modelling.
The claim that it is very difficult to achieve some minimal degree of distributed collaborative modeling is true. No need to repeat now the usual complaints about the XMI interchange format .
What I’m NOT so sure IS about the solution they propose: move UML TO the web. This IN itself IS NOT a bad thing (e.g. the Gravity tool looks promising) but ONLY AS a complement TO existing desktop modeling tools . I’d like to be able to interchange models between different tools regardless the tool is a desktop tool or a web-based one.
Modeling is much more than drawing UML diagrams so it is not enough to provide a browser-based modeling tool for UML we need to port to the web all tools required to manipulate UML models (e.g. model transformation tools).
But even more important than that, the main reason I’m convinced about the problem but NOT so much about the solution IS that I don’t see the programming community going in that direction. Many of us write collaborative programs as part of distributed teams but I don’t know anybody that uses a browser-based IDE FOR that (although they exist, eg. CodeRun ). TO me the solutions IS MORE about stablishing once FOR ALL an interchange format (NOT ONLY FOR the model content but also FOR the layout) AND USING Subversion AND similar TO control the model versions AND merge changes done by different modelers (TRUE, we ARE NOT there yet, especially FOR the merge, diff,… operators ON models but there IS promising research going ON).
Meanwhile, the WebUML initiative can provide SOME relief TO the CURRENT situation. ALL initiatives that try TO improve AND expand modeling ARE MORE than welcomed.
FNR Pearl Chair. Head of the Software Engineering RDI Unit at LIST. Affiliate Professor at University of Luxembourg. More about me.
Recent Comments