Status Updates From A Small Matter of Programmi...

The answer includes task specific programming languages, visual application frameworks, and support for collaborative computing practice

A problem with the task-specific approach is to determine just how task-specific a language should be → this depends on the circumstances under which it will be used. Nardi suggests the activity theory and distributed cognition as two possible frameworks to use to study the context but my guess is they will be useless in establishing boundaries — what should be a part of the system and what shouldn’t. TBD

Wynn made the interesting point that the ambiguity and impreci­sion in mundane conversation itself may, in many situations, be functional and intentional; as she said, “The messiness, potential ambiguity, implicitness... of natural conversation serve many of the purposes that actors have, including those of intimacy and mutuality, through means of less and less explicit surface discourse”

“Spreadsheet users are productive with a small number of functions because the functions are high-level, task-specific operations that do not have to be built up from lower-level primitives.”

I wish I'd read this before I went to the effort of writing a blog post about how great an IDE Excel is.