Actually, some standardization has occurred along the way. In the beginning, C did
not have a standard library at all; programmers always had to ``roll their own''
utility routines. After several abortive attempts, a certain set of library routines
(including the str* and stdio families of routines) became a de facto standard,
at least on Unix systems, but the library was not yet a formal part of the language.
Vendors could (and occasionally did) provide completely different routines along
with their compilers.
In ANSI/ISO Standard C, a library definition (based on the 1984 /usr/group standard,
and largely compatible with the traditional Unix library) was adopted with as much
standing as the language itself. The Standard C library's treatment of file
and device I/O is, however, rather minimal. It states how streams of characters
are written to and read from files, and it provides a few suggestions about the
display behavior of control characters like b, r, and t, but beyond that it is silent.
(Many of these issues are, however, addressed and standardized in Posix.)
If the Standard were to attempt to define standard mechanisms for accessing things
like keyboards and displays, it might seem to be a big convenience for programmers.
But it would be a monumental task: there is already a huge variety of display devices,
and huge variation among the operating systems through which they are usually accessed.
We cannot assume that the years to come will offer any less variety.