We are able to visualize intermediate values that arise from transient selections in the convolution example, albeit with problems related to layout and interaction. These problems represent the next selection of tasks to be picked off.
In the convolution example, we want to be able to see the intermediate matrix used to compute a given cell of the output. To support this kind of use case, we need (the beginnings of) a notebook-like presentation of the program as a graph of cells. (This graph is in some sense an abstraction of the DDG.)
See also
We are able to visualize intermediate values that arise from transient selections in the convolution example, albeit with problems related to layout and interaction. These problems represent the next selection of tasks to be picked off.
In the convolution example, we want to be able to see the intermediate matrix used to compute a given cell of the output. To support this kind of use case, we need (the beginnings of) a notebook-like presentation of the program as a graph of cells. (This graph is in some sense an abstraction of the DDG.)
See also