Good questions.
My test setup is exactly how I will be using TRS19 in the future. I have replicated my dataset and work on the same routes in there just as I would in T:ANE if I were using that. I maybe an exception because I came along with a big package to begin with to test with. The process I went through to adopt my current data is not for the inexperienced, and I would not recommend doing it because it introduced a lot of other 'features' aka bugs into the works which I need to be aware of, but it also gave me a chance to see what kinds of content issues we will face and I've been reporting these to members of the Content Repair Group (CRG), and in the forums as I come across them.
Recent updates brought along some additional bugs/snafus with setups due some missing dependencies. These dependencies, however, are available on the Download Station (DLS) and an update of a single asset missing this dependency updated 378 others I had installed. There is one dependency which is not available and that is driver Therese. If you use that driver anywhere, you will need to modify your sessions to accommodate this by replacing the driver. Outside of these issues, most of the built-in routes should be fine unless other script changes were made that could cause other problems.
If you are concerned that your setup is polluting the test dataset, the alternative is to setup a new one. You do this by removing the current install path in the Trainz settings on the Launcher, and restart when prompted. You will need to reenter your login information and reconfigure your preferences, but you will have a fresh dataset to work with. If you setup a new data folder, I recommend naming it to something like TRS19-BETA-Build xxxxx - where the xxxxx is the current build number. By doing this, you will know which dataset is what, and this will ensure you are using a fresh dataset. These extra data folders can always be cleaned up later on when the full program is released.
Hope this heps.
