Common OS/kernel code between freebsd and linux

[Eric Brunner-Williams]

Howdy Anjali,

I was paid for over a year to read libc implementations and the
supporting kernel code, and wrote XPG/1 as a specification of what my
employers at the time, Bull, ICL, Siemons, Olivetti and Nixdorff,
circa 1986, were intending to implement, from their various v7, PWB,
SysIII and SySV starting points.

I was paid for over a year to sort of do the same thing and co-wrote
XPG/4.2 as a specification of what the big installed base thought were
the syntax and semantics of the kernel entry points and library
routines, even commands, that each assumed was present on a host
operating system.

What you propose is a butt load of work, in effect, the academic
exercise of specification of an abstract platform and its interface to
a network stack.

So who would pay? The correct answer should be of the form "the cost
of not doing this to some parties who can (a) become aware of this,
and (b) pay, is greater than the cost of doing it.

So how much code does either camp care about being portable, and isn't
the best answer to "portability" simply tossing another pizza box at
the problem and running the application native?

You can have fun, but you should have realistic expectations about the
uptake of your work by working programmers and program managers and
product planners and corporate strategic platform planners.

Eric