The cottage industry is ideally suited to small, specialist production runs where a single task is performed by one operative on a whole but small batch of components, with the input coming from a basket at one side of the take and the modified output being placed in a completed (so far) basket at the far side.  

Here's an example of "cottage industry" code (in this case written in Perl) where each process takes a list (collection) input on the right hand side of the = sign, processes it and saves it in another list on the left had side, with a series of processes one after another.



But this is a less than ideal approach where there's a lot of product to be processed.  It gets long winded and in coding terms many intermediate lists are set up which might hang around for quite a while.  A better approach - for the right tasks is a production line approach ...

foreach (sort (keys (%roles))) {

Here, intermediate lists are not saved into named variables but, rather, are temporary and are released after they're no longer used.   But in fact this example does not eliminate the intermediate lists completely; that can't be done simply with anything that needs to be sorted as you have to generate all the records before you can output the first - just in case it's a case of last in, first out.

A true production line scenario would be set up using loops iterating through each element in turn with no intermediate storage of more than a single scalar value.   Perl's approach to that, using the example above, would be via the each function and a while loop (and we could not sort);  in Python, we could take the same approach as in Perl or we could use one of those marvellous generator functions!

---