I find I am in agreement with Craig on most of his outline.
I have been using Gramps now since somewhere around 2010, after I had trialled several other products, and I needed something that gave me more control, and flexibility of output. I had also grown to a database then of several thousand individuals, now very much more than that.[About 4/5 of my database is there so we can confirm they are NOT connected, or only on an extremely distant branch.]
I have always avoided the various commercial offerings like the plague. And for practically all of my work I will as default opt for open source software.
I had an academic background, with extensive experience in applications for scientific and spatial analysis on a wide range of operating systems. When I ended up a university manager I also had to grapple with the inadequacies of the corporate systems our staff and students had to interact with…
Yes, getting Gramps to work for you is a quite big learning curve, but for people who want to do serious things with substantial data sets, there is a significant return on the investment of time and effort to learn how to get the best out of it. I still consider myself only a partly familiarised Gramps user, but I have progressed to the state that I could not do without it.