This talk is a work-in-progress report, overviewing the behavioral theory of interactive algorithms,
being developed by Y.Gurevich, A.Blass, B.Rossman and the first author, and its implications for the practice
of concurrent and distributed programming, being explored by the second author.

We explain and motivate the
distinction of ordinary vs. general interactive small-step algorithms, and the splitting of the latter notion to
stand-alone and component cases. We show how the notion of wide-step (parallel)
algorithm  extends to an interactive situation. Each of these classes of algorithms has an
appropriate ASM syntax capturing the class  exactly.

The derived syntax implies an extension of the current AsmL language to an explicitly interactive version
AsmL-I. The work of the second author on the definition and implementation of AsmL-I has met exactly the same
problems as other ongoing attempts at introducing more transparent and efficient paradigms of concurrent/distributed
programming. The solutions arrived at seem, by first benchmarks of first prototypes, to be remarkably efficient.  Together
with the capability of AsmL-I to express communication and finely grained concurrency/parallelism transparently,
on a high level of abstraction, this has led a part of the current provisional AsmL-I syntax and pragmatics to be considered for
adoption as a general purpose commercial programming tool.