I guess your message has escaped our attention, sorry for the response delay...
Thank you for your kind comments, your feedback is really appreciated!
I'll try to address all your remarks as precisely as I can (I already forwarded the title problem 9) to our webmaster).
First of all, a few of the features you mention are already in our plans and should be implemented soon, 6) and 7) for instance. We will keep you updated about them.
Some other things will be easily added, say soon-ish, as they are not too much work, but we might have a forum discussion with the users about them regarding details. This is the case for 2), 3) and 8) for instance. In those cases, the question is whether most of the control should be confined to the grasshopper definition only (with eventual help from additional ShapeDiver components), or if the viewer should be able to define views, visibility or lock/recompute abilities regardless of the definition. We will start discussions in the forum, as users' experience is key regarding these functionalities.
Many users already asked about plugins/dlls. This is more of a mid-term priority at the moment. Obviously, we have to test all plugins and dlls with our system before installing them on the Grasshopper servers. We have some work to do before iterative and/or recursive plugins like Galapagos or Anemone can be included, but overall we plan to enable some widely used plugins, as it is relevant for them to be included directly on the platform. Several of these are already being considered, and we welcome suggestions (math.net dll noted). For more confidential ones, it is likely that a future version of shapediver features professional accounts with dedicated servers allowing more customization (which will go far beyond custom plugins and dlls).
About internalising all data (item 5)), we would also have to start a discussion. Could you expand a little bit on what you mean? We will soon add the possibility to download CAD files for the models displayed in the viewer (3dm, dxf...), but I think you have a more general vision and it sounds very interesting.
Now for the last and most challenging piece: we have thought very early about sharing a session between several users (item 4)). I can say that it will be possible in the mid to long-term, but I can't say when yet. This is a very exciting possibility and it would be a shame if ShapeDiver didn't offer it. This is where we have a chance to take parametric modelling one step further. Once again, we will soon start a big discussion about this topic, and the way users would like this feature to work. Should only one user at a time be an "emitter" and the others "receivers"? In this case, should we lock the stream to the emitter's view, or only broadcast the parameter changes? How should people collaborating on a model switch emitting control between each other?
This general topic goes hand in hand with a symmetrical feature we have thought about: the possibility of having co-dependent models that influence each other, or even meta-parameters sharing their influence to several models.
Here you have it, ShapeDiver's future from the next, concrete steps to mid- and long-term dreams. This should give you an idea of where we stand and what our goals are.
We hope these possibilites can spark discussions, ideas and excitement for the future.