It seems to me that algorithms written in text don't match these ideas at all well. And, I believe that a lot of people have trouble thinking strategically about putting together sequences, so algorithms are best avoided. (MIT's Scratch team is doing interesting things with children without abandoning algorithms by using the Nassi-Schneiderman notation.)

I am seeking a concrete model of software that we can match up to the idea of putting our hands on it. This has led me to an object-flow model, and in the 1990s I spent a lot of time making that work in a practical way.