Does the time it takes to recalculate the definition increase with higher vertex counts of externally loaded meshes when being translated?
I think it does but I don’t exactly understand, why. Wouldn’t this be an action that only affects stuff in the viewer itself?
Computation time of the definition does not depend on externally loaded meshes, only the url passes through the definition. Bigger meshes will take more loading and processing power in the viewer, maybe this explains the difference you are experiencing?
Alright, thanks for clarifying.
I was hoping that loading power for external meshes would only be required when the viewer gets initialized. Do these external meshes get reloaded when being translated, then?
For now they are (although likely using local caching if the same mesh is reloaded in a translated position). The reason is that the translation request still goes through the definition and the scene is recomputed from the response of the servers. This will change in the future for static assets in general, i.e external geometry but also materials; as of now, the behaviour you described is similar to changing a material color for example.
Could you give an indication on when this will be implemented? For one of my models this will make all the difference.
This is a major change in the communication between ShapeDiver servers and clients embedding the viewer. We do consider this to be high priority but it will take a significant effort and I can’t give a release schedule at the moment.
Thanks for the clarification!