Gnisrap

[page 66, column 1] The Parsing-Gnisrap model is a model of coding, rather than of programming, that is, unlike most earlier models, ours starts at the stage when the programmer has already solved the main problems of how to write the program, and wishes to turn these into code. […] [page 69, column 2] Similarly we believe our model could handle the different ‘psychological complexity’ levels of certain drawing programs to which Kiss and Pinder draw attention. In fact, our approach has much in common with theirs at the gnisrap end, but they do not consider the parsing side. So much for the output side – the ‘gnisrap’.

[M]any deviations are made from linear development, where programmers leave gaps in the emerging program to be filled in later. Green et al (1987) have proposed a model to account for this finding. Their Parsing/Gnisrap model introduces a working memory component into the analysis of coding behaviour which forces the model to use an external medium (eg the VDU screen) when program fragments are completed or when working memory is overloaded. This means that programmers will frequently need to refer back to generated fragments in order to recreate the original plan structure of the program which may have only been partially implemented in code. The parsing element of the model describes this process, while gnisrap (the reverse of parsing) describes the generative process.

A similar account is offered by the ‘parsing/gnisrap’ theory of computer programming Green et al. (1987). Programming, they argue, involves two processes: parsing, the generation of external code structures from internal cognitive structures, and gnisrap (the reverse of parsing), which involves the recreation of internal structures from code during code comprehension.