unix - why fork and exec are kept 2 seperate calls -

i understand differences between fork, vfork, exec, execv, execp. pls dont rant it. question design of unix process creation. why did designers think of creating 2 seperate calls ( fork , exec ) instead of keeping 1 tight call ( spawn ). api design reason developers had more control on process creation? because of performance reason, delay allocating process table , other kernel structures child till either copy-on-write or copy-on-access?

the main reason separation of fork() , exec() steps allows arbitrary setup of child environment done using other system calls. example, can:

• set arbitrary set of open file descriptors;
• alter signal mask;
• set current working directory;
• set process group and/or session;
• set user, group , supplementary groups;
• set hard , soft resource limits;

...and many more besides. if combine these calls single spawn() call, have have complex interface, able encode of these possible changes child's environment - , if ever added new setting, interface need changed. on other hand, separate fork() , exec() steps allow use ordinary system calls (open(), close(), dup(), fcntl(), ...) manipulate child's environment prior exec(). new functionality (eg. capset()) supported.