I'm experimenting with Smile & Genie - and enjoyed it very much

However, during my experiments I found some (minor) issues in Smile (Java JNI Layer) that affect the visualization in Genie we want to use for our tool. Maybe these issues can be solved in a later version Smile? Here is the list of our (minor) issues:
OutcomeDescription-Limitations (dots and other special characters):
We're planning to use genie to visualize the results of a code recommender engine. Thus, in our app '.', '#' and braces '({[<>]})' are important characters and replacing them with underscores reduces comprehensibility of our models presented to a user. Could these character restrictions be relaxed further?
Saving node visualization as Bar Chart
Serializing a network (using network.write()) results in plain networks. To my knowledge (TMK), there is only a reduced set of functionality in the JNI API that can be used to affect the visualization in Genie. It would be a nice feature if a visulization hint for Genie could be stored for each node - like show as Bar Chart (maybe there is already one using a user property which I don't know yet?)
Integrating the use of predefined Layouters in Smile
TMK, computing the layout information for each node of a network is a manual task. There exists a predefined set of Layouters in Genie which could be used to create a predefined layout easily. Maybe the integration of yet existing layout-engines could be an option (http://www.manageability.org/blog/stuff ... on-in-java) ? Likely, this feature should not be a part of the Smile API but adding some documentation on a Wiki page how to use external layout engines would be a nice feature.
Saving a flag Indicating to infer probabilities immediately after loading(instead of using Ctrl-F5)
Really, really a minor issue

That's all I have in mind yet

Best wishes,
Marcel