The context for this discussion is the upcoming software for the New York Textile Lab, where their network takes wool from sheep and alpacas and, through several processes, converts it into garments for people to wear..

Planning forward means to start with wool and find out how much yarn can be made from the wool on offer.

Planning backward means to start with requests for yarn and see how those requests could be fulfilled.

The Pattern of the Textile Flows

Using the Input-Process-Output Resource Flow pattern.

The processes include:
* Pick Up and Combine Lots,
* Scour (clean the wool),
* Spin,
* Split up yarn, and
* Knit.

With transportation between the processes, so the overall pattern is:
* Pick Up and Combine Lots,
* Ship to Scour,
* Scour (clean the wool),
* Ship to Spin,
* Spin,
* Ship to Knit
* Split up yarn, and
* Knit.

The algorithm

Each process has inputs and outputs. Assuming a Valueflows Recipe or something like it, the quantities of the inputs in the recipe will be required to produce the quantities of the specified outputs, for example, 1 measure of apples to produce 1 pie.

If planning forward, start with the offered wool.
* add each of the input resources to the process's input list,
* and use the process's output calculations to create the process's outputs.
Then take the outputs and make them inputs to the next process.

If planning backwards, do the same thing in reverse, starting with the requested yarn,

“If you are supply-driven and want to know how much yarn could potentially made from wool on offer”