Abstract: Attribute grammars are a powerful specification
paradigm for many language processing tasks, particularly semantic
analysis of programming languages. Recent attribute grammar
+
+
+
systems use dynamic scheduling algorithms to evaluate attributes
by need. In this paper, we show how to remove the need for
a generator, by embedding a dynamic approach in a modern,
object-oriented programming language to implement a small,
lightweight attribute grammar library. The Kiama attribution
-
-
-
library has similar features to current generators, including cached,
uncached, circular, higher-order and parameterised attributes,
and implements new techniques for dynamic extension and
@@ -1051,7 +1068,7 @@