In my case, I've got dozens of pieces, most of them 3,000+ measures with 50-page parts, which would require considerable effort to port to another platform, even with MusicXML. I still need to fix little problems I find from time to time, so need everything ready to do that in case something comes up.
I spent at least a year learning Finale well enough to be really productive, and really many years becoming truly expert. These other programs are likely easier to learn, but would still require substantial effort and Q&A on help forums, likely months before really getting up to speed, which would detract from doing actual productive work.
That's why, in my case.
Jari's and Jan's work is amazing and brilliant and I'm very grateful for it, but the fact is, by definition as plug-ins they're bags on the side of Finale, even if MakeMusic were to distribute them itself as official plug-ins. Though they make up for some of Finale's shortcomings, it's not the same as having that functionality built in.
What I see at Dorico is a hard-working and enthusiastic open-minded team eager for bug reports and suggestions wanting to make the product better and better. I could be wrong, but what I see at MakeMusic looks like a reluctance to change anything for fear of breaking something, an inclination to write off bug reports as "not reproducible" or "only affects a small number of users," and considering recreating a problem file by running it through MusicXML a valid solution to a problem. That new releases always seem to introduce new bugs and "oops--didn't think of that" demonstrates that their fear is justified, I guess. I'm not sure what they can do about it, but these blog posts and the last few releases leave me very worried.