Hello:

It is not very clear to me how or where I can store a function that will be used in future scripts? Do I store them in a specific directory, profile, etc??

Depends which shell. You can generally have a number of functions in your .profile which would give you the performance benefit of these being memory resident when you come to use them.

You could also store them in any old file(s) which you could use as a functions library to source (.) into your scripts when you require them.

In addition, in ksh, you can use the autoload feature. This uses the FPATH environment variable to search for files matching the function you have attempted to call. If it finds a matching file name, it assumes this hold the fucntion definition and 'autoloads' it. To turn this functionality on, use typeset -fu.

Damian

Depends which shell.

currently using the bourne shell.
============================
You could also store them in any old file(s) which you could use as a functions library to source (.) into your scripts when you require them.

Yes ..... I would like to follow the suggestion above.

dir villan
dir scripts
dir fctns
file AB <ftp function>
file BC <width function>
file CD <count function>

Would the above directory structure work?

Does the shell immediately know it is a function and where to look when the (.) is used within a script

Do I have to put the above tree structure in the $PATH variable?

Sourcing a file/script simply executes that script in the current shell as opposed to forking another. This means that any variables or functions declared in the sourced file will be available in the calling script.

You don't need to do anything with the path (unless you want to) as you can always supply it when you source the file.

e.g.

. /some/directory/containing/yourFunctions/functionsFile


Also, from memory, some implementations of the Bourne shell will require the keyword source rather than the '.' syntax to source a file.

HTH