I do own a TMG license and I have played around with it a bit. But I have never really been a TMG user. Curiously, I was inspired to purchase and play around with TMG only after the demise of TMG as a supported product and only after RM developed a direct import for TMG. There were many very interesting posts on this forum from former TMG users who were switching to RM, and I wanted to get a TMG license to see what the fuss was all about. I ended up getting sufficiently into TMG that I was even invited to make RM presentations to former TMG users.
TMG was a very interesting product with a lot of interesting and rebellious ideas that I tend to like and to share. One of TMG's greatest strengths was its great power to produce wonderful sentences and wonderful reports. However (and this is where I probably differ with the TMG world), TMG accomplishes these wonderful sentences and these wonderful reports by tightly coupling the actual genealogical data with the metadata about how to construct sentences from that data. The coupling is so tight that it makes it very hard to extricate the data from the metadata. There was a time that I would have been fully on board with the concept of coupling the genealogical data with the metadata of how to present the data. But now I really think that the data and the metadata need to be completely uncoupled. My point form printed reports exemplify that concept.
Since you are from the TMG world, you are probably aware of a product called SecondSite that would do a direct import of a TMG database and produce wonderful Web sites from that data. Since the demise of TMG, the author of SecondSite has developed a new product called GedSite which will produce wonderful Web sites from GEDCOM. Because it uses GEDCOM as its input, GedSite can make wonderful Web sites using data from RM or any of RM's competitors.
One of the things that GedSite tries to do is to use the GEDCOM input to interpret all the metadata it can about how the data should be presented. For example, I can export GEDCOM from RM, import the GEDCOM into GedSite, and Gedsite will produce sentences on the Web site it produces for me that are identical to my sentences in RM itself. Happy, happy, happy. Except that curiously, I decided not to do it that way. The Web offers some opportunities in presenting data that simply don't work well for printed reports on paper. In addition to taking metadata about sentences from GEDCOM, GedSite includes its own sentence definition language which is somewhat similar to that of TMG and SecondSite. So I do my GedSite sentences that way, and I'm delighted with the results. The presentation of the data is superior that which would be achieved simply by using my RM sentences in GedSite. In other words, my RM sentences are optimized for paper and my GedSite sentences are optimized for the Web. The underlying genealogical data is identical.
I only have a few people online at a prototype site created by GedSite, but much more is coming "real soon now". You can see my prototype at http://jerrybryan.co...e/g1/p1.htm#i14 I'm especially proud of the Burial GPS fact in RM that actually becomes a link to a live Google map of the burial site. The link is towards the bottom of the GedSite page. Neither RM nor GedSite supports this burial GPS link in exactly the way I want it supported, but there are always tricks and workarounds.
Once you are on the page to which I have directed you, you can click around a bit with the menus at the top of the page to see a few other people for whom I have prepared RM data for display via GedSite.
In the highly unlikely event that I ever switched from RM to some other desktop software as my main genealogy software, I wouldn't care and I really wouldn't want my RM sentences to be imported into my putative RM replacement. I would expect to define my sentences in the new software - full separation of data from metadata.