0

I am working currently with the GMF-Framework, created an ecore-Model and a GMF-Editor based on the model. The created editor was integrated in Papyrus (v1.0).

Based on my own DSL I created a model with (round about) 1.750 elements, which are visualized in several diagrams. One of these diagrams is the main diagram and contain all elements (other diagrams are generated with EOL dependent on the associations between the elements). Making some changes in the main diagram slow down the reaction time of the tool (or sometimes there is no more heap space). First question is to speed up the reaction time for a "smooth" use? Has somebody experiences with issues like that?

Sometimes I thought about a change to Microsoft Visual Studio Modeling SDK, but I think, there are not such visualization concerning individualization possibilities like in the GEF-GMF framework but with more stability (I think). Also here I dont have any experiences with the MS Modeling SDK, inclusive the pros and contras between these two frameworks/sdks. Can someone report his experiences with it?

Kind regards Erik

Shounbourgh
  • 258
  • 2
  • 14
  • I'd expect that you can split the model into relatively independent packages, after all it's not only the computer that cannot keep track of this many classes at the same time, it's also your head. In the main diagram, I'd expect only to see the packages, not their actual contents. – flup Oct 02 '14 at 08:32

2 Answers2

0

The most covering empirical evaluation that I’m aware of is conducted by CEA/LIST. Unfortunately they don’t cover MS DSL tools but a number of others. See the report at: http://hal.archives-ouvertes.fr/docs/00/70/68/41/PDF/Evaluation_of_Modeling_Tools_Adaptation.pdf

Scaleability is indeed one relevant feature and the other one that is detected later in real use also is evolution: how models update when the langauge and tool is updated.

Juha-Pekka
  • 97
  • 2
0

I can't speak for experience on either package.

I'd expect the effort it takes to switch, both your learning and all the supporting machinery you've already built, would make doing such a switch simply not worth the trouble. It would likely take you a long period of time to switch over.

If you want faster performance, hold your breath for 1-2 years and work on enhancing your tools. Utility is always a big draw with a tool, and the chip and tool vendors continue to work hard to make things better.

Any solution that comes from sitting on your hands and doing nothing, is always a good one.

Ira Baxter
  • 93,541
  • 22
  • 172
  • 341