Discover Research Tools for Textual Study
Micro-EYEBALL is a microprocessor version of the English language parser developed in the 1970s. This historically important legacy tool could also provide statistical descriptions of many linguistic features of a text. This tool is no longer available, and does not have a web presence.
More information on Micro-EYEBALL, its applications and development is available here:
Ross, Donald and David Hunter. "Micro-EYEBALL: An Interactive System for Producing Stylistic Descriptions and Comparisons." Computers and the Humanities 28.1 (1994): 1-11. Web.
More information about the original version of EYEBALL is also available here:
Ross Jr., Donald and Robert H. Rasche. "EYEBALL: A Computer Program for Description of Style." Computers and the Humanities 6.4 (1972): 213-221. Web.
Micro-EYEBALL was a revision of the mainframe program EYEBALL, which "parses English language texts and provides a statistical description of many linguistic features which are of interest to those involved in stylistics" (Ross and Hunter 1). It was a microcomputer implementation of the original, which dated to the 1970s, so the program could be of continuing use to humanities scholars who had not participated in mainframe computing (1).
EYEBALL, developed by Donald Ross Jr. and Robert H. Rasche, was an information system for students of literature with particular applications for the study of literary style, literary criticism and the history of ideas (Ross and Rasche 213). They began development in the late 1960s:
The general design of the project was outlined in December 1969; by July 1970, the basic VLII programming for all stages had been completed. Current work is in two areas. First, the PL/I programs have been "translated" into the Fortran language, in versions for IBM 360 and CDC 6600 computers. This step has made EYEBALL more widely useful; we have found also that Fortran is markedly faster (and cheaper). Second, we have designed batch-processing algorithms of Phase 2 and Phase 4 for computing systems which do not support interactive processors. (Ross and Rasche 216)
EYEBALL consisted of a series of computer programs that first surveyed natural language text and then marked wherever possible the text's syntax and vocabulary use properties (Ross and Rasche 213). The intermediate processing stage "indicat[ed] the ends of phrases (prepositional, infinitive, and so on) and of clauses and locations of subjects, predicates, and objects within clauses" (213). EYEBALL attempted to navigate the ambiguity of the English language while "maximizing accuracy and minimizing the man-hours involved in the analysis". Its error tolerance was in the range of 2%, which Ross and Rasche found to be "quite acceptable" to science and social science standards, particularly compared to the up to 10% error rate Ross calculated for his own hand counts (213). It did not require a large computer dictionary, nor did it require users to apply markup to their texts prior to processing (Ross and Rasche 214).
The microprocessor edition, completed in 1994), was a major overhaul of EYEBALL (Ross and Hunter 1). Ross and Hunter wanted to ensure Micro-EYEBALL would run on widely-available computers to remove the cost and accessibility barrier for even poorly-funded researchers (Ross and Hunter 2). It ran on IBM/PC or compatible personal computers and required only 512K of memory and DOS 3.02 to run, and though a hard disk install was preferable, Ross and Hunter believed that it should also work from floppy disks (2). They kept Micro-EYEBALL's constituent programs in FORTRAN and designed the whole "to be ported to other systems with little or no alteration of the source code" (2).
Micro-EYEBALL accepted ASCII text files and worked best with samples of 500-3000 words (2).
Its programs could accomplish common tasks such as generating word lists and parsing (Ross and Hunter 4), and could also systematically characterize distributions of categories and functions (Ross and Hunter 4). Ross and Hunter aimed for Micro-EYEBALL to be straightforward and to accomplish necessary tasks without requiring a large investment of time or effort of its users:
The parser does not identify "deep" or "remote" structures, transformational histories, nor is it delicate enough to analyze nominal or verbal phrases beyond their main constituents. Unlike a parser which might emulate the psychology of the human reader, both versions of EYEBALL have worked within the technical limits of readily available, inexpensive hardware and software. The aim of the project is not to develop a computer program which teaches or learns English grammar. Our goal is to get to an accurately tagged text with the least time and effort on the researcher's part. We have decided upon straightforward syntactic and structural descriptions since a more elaborate system would produce descriptive classes too small for the sample size we recommend. (Ross and Hunter 3)
In their view, EYEBALL enabled researchers to quickly and easy test hypotheses about relationships within texts (Ross and Hunter 10).
TAPoR v.2.5 | Copyright © 2014 TAPoR Team, University of Alberta.