Universalization.

At the dawn of the current generation of computers (ie, almost 35 years ago), it was fashionable to talk about universalization. It was exactly how the universalization tool was used to talk about the purpose of the OS (in any case, it was so deposited in my head). With the era of FAT16, INT21, and Windows that came to light, they forgot about it. In the late 90s, and in general, many talked about it as a utopia. For the last few years, I began to notice that people are again flocking back to the universalization square by devious paths and gardens. Already with the burden of compatibility, with huge bales, never knowing the refactoring of libraries and links ... But still.
1. Different operating systems start experimenting with connecting each other's drivers in small steps. FreeBSD can do some Windows NDIS drivers, Linux, too, FreeBSD is developing a wrapper for Linux drivers. Xen Xen has become increasingly used as a universal interface peripherals.
True, iron is becoming more and more closed. Apparently, you need the emergence of a new "IBM PC". Cheap and open.
2. Appear and get along with unspoken standards. MVC there everyones, OpenDoc, PDF, Flash, SiteMap, browsers, at last, came at least to the oral agreement.
3. Baytkod. Platform-independent bytecode. This is almost a fad, the peak of fashion. Yes, why is there baytkod? Distribution in a heterogeneous environment, mobility and compatibility.