I mean, I get that argument, but I actually ran BeOS full-time for over a year. It had a great Works-style office suite (GoBe Productive) written by the people who wrote ClarisWorks, a few good graphics programs including an
amazing competitor to Macromedia Fireworks (e-Picture), a solid BBEdit-like programming editor (Pe), a few music programs that did things that I’m not sure I’ve seen other systems do
to this day like SoundPlay’s wacky ability to act as a mixer, with speed control, between multiple MP3 files or ObjektSynth’s…object-oriented synthesizer (it’s very hard to describe). There was a stage control system for live performances whose name I forget now—the company is still around, as far as I know—that was used,
running on BeOS, for several Broadway shows and Circue de Soleil installations. And an animation program that started on BeOS, Moho, is still around today.
The engagement was certainly starting, and I think there’s a chance—a small one, to be sure, but a chance—that if Be, Inc., hadn’t clearly decided that carving out a comfortable niche just wasn’t enough, BeOS might have succeeded. (Instead they decided to go all-in on “Internet Appliances,” which ended up dealing them the death blow rather than a big success. Ironically, I think that market effectively succeeded a decade later, but in the form of the iPad.)