Alexandre Oliva wrote:
[kernel and kernel inspection and filesystem utilities, no more]
Per your definition, UNIX wouldn't have ever been regarded as just an
operating system.
And then, UNIX philosophy would probably be limited to "everything is
a file" because, by this definition, everything else we understand as
UNIX wouldn't apply to the operating system, but rather to this
broader concept for which we don't seem to have a name.
I look at at this way: a unix-like operating system is the part that
makes everything look like a file and applications that are portable
across them only need an API of creat(), open(), read(), write(), and
ioctl() with all needed permission/device/socket/fifo magic handled by
the OS during open(). That's not completely true and of course there
are library layers above that, but it is the concept that distinguishes
unix. It's closer to 'everything is a file descriptor' than everything
is a file since they are often inherited without the app even knowing
about the associated file/device/socket.
Nothing like "many small programs, each doing a single simple task
very well, that can be combined through pipes and a powerful shell
programming language" would be part of the UNIX philosophy, because,
well, these small programs wouldn't be part of UNIX per this narrow
definition.
That's a good idea under any OS, not particularly unique to unix.
Nothing like "the same low-level programming language usable all the
way from the guts of the kernel to applications, and a well-defined
system API available to that programming language" would be part of
the UNIX philosophy, and it wouldn't have a C library (and a C
compiler) as fundamental building blocks.
The c library isn't unique to unix by design.
--
Les Mikesell
lesmikesell@xxxxxxxxx
--
fedora-list mailing list
fedora-list@xxxxxxxxxx
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
