The reason I would like to be able to load an endpoint configuration for a VSTi, is that I use different sound libraries for the same type of instrument at the same time. I do it both for layering (a common practice of having different libraries combined to make a single sound), for using complementary libraries, and for extending the amount of techniques for a single instrument (for example, to add extended techniques to a basic set of articulations).
a) I want to use VSL Synchron Strings Pro and Spitfire Chamber Strings to create a complex, layered sound in two separate groups of staves (SSP Strings, SCS Strings). With the current system, the playback template will automatically assign the first endpoint configuration it finds in its list to both groups. Programming the second one must be done manually, instrument by instrument. Reassigning the playback template would scramble manual programming.
b) I want to use a set of Orchestral Strings for tutti, one of Chamber Strings for divisi. Again, the playback template will always recognize each occurrence of Violins as the first endpoint configuration (in this case, Orchestral Strings).
c) To create a complex instrument, made of Violins + Violins Extended I have to create two different instances of violins, each one referring to a different sound library. But the playback template will always assign the first one found in its list.
Being able to assign a separate endpoint configuration to each VSTi would allow mixing different plugins. Any complex routing would be allowed. Loading of techniques would be more granular, since each sound library could be associated to a set of techniques.