A shared data program is no different from a shared text program
except that the magic number 412 specifies that the shared area is
read-write. Separate shared datas are distinguished by a char in the
a.out header unused word. Thus when a shared data program is executed
it will link to the shared area that matches if it exists otherwise
it will be established.
It is however a problem to produce code from c that is in the data
or non shared segment. Here is where the program 'lked' comes in.
Its syntax is as follows..

lked id share.o -(ld flags) nonshare1.o nonshare2.o library -lc ..

Thus the first file's text area becomes the shared area with id 'id'
and the rest of the code is put in the nonshared or data area.

For examples of its use see the file 'pgm.m'.
