Right now a task can only resolve if one input (with the same pattern, e.g. *.txt) is given, if pattern within input renders more than one possible hit task will not properly finish. However some in order to take multiple inputs task must:

• Control the computer resources in order not to exceed the number of available CPUs and memory. For instance, when we pass more than one input and give it to a program that uses multithreading, if many inputs are given resources may be exceeded and odd behaviors may happen.
• There are two types of usages in which this might be useful and that render very different results:

1. When task returns one output per input file
2. When task returns one output for all input files given.

Note: output may in fact be several files, here I mean by output the 'results of a task'.