version and model information in output of server

[andrewufrank]

Stanford coreNLP improves from version to version - would it be possible to include the information about the version and the models used for training in the output xml (and json) files? The tagsets used for some languages change from version to version and confusion could be avoided if the relevant information would be included in each output file - at a very small cost!
thank you!

[manning]

This does seem like it would be a good idea.

We're trying to ramp down changes before releasing a new version now, but a good thing for the future.