More information on FANGORN, its use and its development is available in the following article:

Anderson, Peter Bøgh. \"FANGORN: A Special-Purpose Language for the Humanities.\" Computers and the Humanities 15.4 (1981): 227-242. Web.

Program Overview

FANGORN was a language written particularly for literary and linguistic analysis developed by Peter Bøgh Andersen of the Nordic Institute of the University of Aarhus in the 1970s (Andersen 227).

In Andersen's words, \"FANGORN, a special-purpose language for linguistic and literary analysis, was designed because I encountered certain descriptive problems that were not due to the linguistic research I did, but to the languages I used\" (227). Andersen had encountered difficulties particularly with most languages' inability to recognize linguistic constructs as being simultaneously subject and object (227), and so, \"FANGORN was designed in order to avoid the descriptive problems generated by the subject/object dichotomy\" (Andersen 228).

FANGORN programs consisted of networks of nodes with a known location, name, and a process related to other nodes in the network (228). Each node and node relationship corresponded to a value in the source text (Andersen 229). According to Andersen, \"[t]he interpreter is programmed in modules, each node roughly corresponding to a SIMULA class declaration. Because of the modular structure, it is easy to remove or add new nodes\" (Andersen 241). It was applied to a variety of text analysis functions:

The language has been used for minor, rather different purposes: morphological analysis and synthesis, testing sentence grammars and narrative grammars by synthesizing sentences and stories, semantic analysis, plan construction, question answering, and parsing.\" (241)

At the time of Andersen's 1981 article \"FANGORN: A Special-Purpose Language for the Humanities,\" FANGORN was in its third version (241). Andersen implemented the language via \"an interpreter written in the SIMULA 67...running under the NOS operating system on the CDC Cyber 173,\" which was available at University of Aarhus' Regional EDP-Center (RECAU) (241). In terms of performance, Andersen noted that the interpreter \"is reasonably fast, but it consumes far too much space\" (241). He offered a user manual, available by contacting him directly at the University of Aarhus (241).

Andersen evaluated FANGORN, though prefaced with the disclaimer that \"[e]very father loves his children, and that goes for programming languages too. Objective evaluation is difficult\" (Andersen 242). From the perspective of its developer, Andersen found that \"[w]hat I like most in FANGORN is the 'bodily' character of the algorithms: it is possible to have the program build and modify algorithms in rather sophisticate ways. This is extremely useful in programs dealing with historical and developmental issues\" (242).

He also acknowledged its weaknesses:

The main drawback in the language is the syntax. FANGORN programs are as unreadable as LISP programs, and I always use a graphical notation when I write them. One solution, and possibly the best one, is to provide FANGORN with graphical input facilities. Another solution is to design a linguistic front end that is easier to read and write for humans. The problem is that the bodily nature of the algorithm, i.e., the network that implements it, must be completely transparent from the program text; else the user could not change it...the syntactical problem is not merely a technical problem. It arose because we have no suitable framework or categories in which to describe self-modifying processes; at least, I have no such framework. (242)

However, in Andersen's view, \"FANGORN was built as a tool for experimenting with exactly those processes\" (242).

Source

Anderson, Peter Bøgh. \"FANGORN: A Special-Purpose Language for the Humanities.\" Computers and the Humanities 15.4 (1981): 227-242. Web.