Answers are grounded in this paper and cite their sources.
If a compiler is to generate efficient object code, there are several different kinds of optimization which should take place. Each of these optimization procedures has a preferred domain: that is, some algorithms prefer to operate over the input string, others over the tree which describes the syntax of the string, others over the macro-instructions which are generated from the tree, and so forth. In an earlier paper, one of the present authors pointed out the necessity for employing the tree form in particular as a natural domain for optimizers which consider syntactic context and suggested that, just as Irons and others had built general-purpose table-driven parsing algorithms, one could also build a general-purpose table-driven program for getting from trees to macro-instructions. The final compiler design presented here is the result of pursuing that kind of thinking somewhat farther.
No takes yet. Share an insight, caveat, or question.
Warshall et al. (1899) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: