The Burroughs large system architecture had (still has, although it's now emulated on x86) tagged memory, so that the hardware could distinguish references to fundamental types and ensure consistency; it also had non-page-oriented virtual memory (maybe one of the tag bits for a pointer said non-resident?). It worked, and it had a reputation for security - banks tended to use those for decades. Think of it as a bit analogous to a JVM on hardware, an instruction set designed for the high level language (ALGOL). It didn't even have an assembler per se (although the system implementation flavor of ALGOL had the added ability to do what amounted to embedded assembler). It also had typed files: only an executable typed as a compiler was allowed to change a data file to a (read only) executable, and only an administrator could change some other file type into a compiler . So in principle, nothing could run on it except trusted compilers and code produced by them. And mere applications programmers didn't have the privileges to use the system version of the compiler, just the application programmer version, which couldn't do the more dangerous things. It supports or supported quite a range of languages in addition to various flavors of ALGOL. At one time, it had a flavor of PL/I; and presently, it has a C compiler and (limited) POSIX interfaces. The design is unusual even today, and was very innovative in 1961! Something a bit beyond that could arguably provide a degree of hardware-level support for an object-oriented language, perhaps making it safer, perhaps making it a bit easier to do syscall wrappers. On the flip side, even on such an unusual architecture, a C compiler was possible, although one might imagine that the architecture makes certain ways of shooting oneself in the foot more difficult to get away with.

For movies, in many cases the production has been beyond the scope of small organizations or individuals, although even this is changing. But for music, and even more for books, anyone with talent and some fairly modest equipment can produce and can electronically distribute. I would think that there's a real opening for a new class of agents, that know how to wheel and deal with Amazon, B&N, and other online distributors. Given that, the traditional publishers and distributors _ought_ to die IMO, and make room for a more efficient system that does less gatekeeping on the many authors and artists with real talent but little name recognition; and one that, at least once they gain some recognition, puts a lot more of the profits in the hands of the creator (and the agent that directly serves them) rather than in the hands of traditional physical publishers that just don't add enough value to justify their costs.

The AT&T 7300 (aka "Unix PC") from the '80s could take messages as email attachments,
if it had the "Voice Power" (voice-grade digital audio and touch-tone recognition)
board and accompanying software installed. I don't think it did Caller-ID (indeed,
I think it predated widespread availability of Caller-ID), but it certainly provided
random acess to messages.

I would expect that commercial call center equipment and software has had such capabilities
(including something like Caller-ID) well before they were widely available to the consumer.