It is evident from the above discussion that computational origami has a number of serious disadvantages. For example, if simple, fine-grained processors are used, either a lot of hardware or long delay lines are required to execute a complex program. It is impractical to implement an origami array in silicon because the long delay lines required are not area-efficient when implemented using standard techniques. In addition, even with the maximum number of useful processors available, the system will have a high latency. Finally, the systems that have been characterized here can only run programs that can be implemented without indeterminate recursion, since the data flow must be causal.
However, origami has a number of advantages. Origami systems are easy to lay out and wire. There is a nearly perfect processor/speed tradeoff, as opposed to most parallel processing architectures which get bogged down in communication costs as the number of processors increase. There is excellent throughput when enough processors are available. Programs are independent of the number of processors available as long as sufficiently long delay lines can be constructed, and need not be recompiled as the number of processors changes. And complex problems can be performed with as little as a single processor if necessary.