Re: "the dodgy Google/Libre duo"
"[...] or even see if you can reimplement your application in Libre Office. "
That is what I intend to do. Each tentative trial port is a test to see what will transfer cleanly. With 2020 not far away then next year will be the time for a serious attempt to make the port to Libre Office on Linux.
After 50 years of programming in various OSs and applications I believe I have the skills to improve LibreOffice itself. The alternative is to recode at least some of the VBA part of my application in C.
There is already a tentative development to improve the processing speed under W7 Excel. An attempt to produce a VB.Net binary of some functions failed when the conversion program failed to run after an apparently successful install. As always with such situations - which path is most likely to be most productive? I have a support programmer's mindset - I always look to build on existing blocks rather than always wanting to reinvent wheels.
In the meantime a large chunk of my time is spent keeping the application running on W7 Office 2007.
It is an "alligators in the swamp" situation - running the application to a deadline every week is currently a higher priority than doing the port.