Documentation
Attributes
- Type of analysis: Uncategorized Programming
- Background Processing: Doesn't Run in Background
- Ease of Use: Moderate
- Historic Tool (developed before 2005): Influential No longer in active development
- TaDiRAH Goals: Creation
- TaDiRAH Methods: Programming Uncategorized
List this tool appear on
Rating
Suggested Tools
Comments
More information on FISHER and its use is available in the following article:
Allwright, Christine. \"Software Review: FISHER: A String-Handling Package.\" Computers and the Humanities 10.5 (1976): 297-298. Web.
Program Overview
FISHER (FORTRAN Implemented String Handling and Editing Routine) was a package of routines for text handling in FORTRAN developed in the 1970s by A. Colin Day (Oakman 181). It was distributed on 1500 punch cards and available by contacting Day at the University College London's Computing Centre (Allwright 297).
Robert Oakman, commenting on FISHER in his book Computer Methods for Literary Research, noted that \"using FISHER's commands resembles programming in PL/I; the process of putting them together in order to carry out a major job like a word index of a text is similar to programming in a more textually compatible language\" (Oakman 181).
FISHER was also reviewed by Christine Allwright in the journal Computers and the Humanities. In her examination of the package, she found that
FISHER consists of a set of Fortran subroutines and functions which, it is claimed, in effect add several new data types to Fortran....The three new data types of greatest interest to the user are: the character held as a small integer, the string of characters represented by an integer value and the marker pointing to a particular position within a character string, also represented by an integer value. Besides being able to create items of the new data types, FISHER provides facilities for their efficient use in many text handling situations, for example, comparing strings, searching for particular characters in strings, removing sections, joining two strings together, inserting substrings, etc. (Allwright 297)
The FISHER package was written on an IBM 360, a mark of cross-system compatibility (297). Allwright was able to run it successfully on a CDC 6600 using two other FORTRAN compilers (297). The package included documentation comprised of three sections, a User's Manual, a Technical Description, and an Implementation Guide (297). The User's Manual included descriptions of FISHER's text-handling, error-handling facilities and explanations of error messages, plus \"a comprehensive index and glossary\" (297). The Technical Description expanded on FISHER's various aspects (297), such as its routines and data structures (Allwright 298).
Allwright favourably compared FISHER's capabilities to SNOBOL, a contemporary language designed for string handling (297). In her review trials, it proved time-efficient when running small and medium programs (Allwright 298). However, she assessed it to be of most value not to beginners, but rather \"to assist the experienced Fortran programmer...given that, it is relatively straightforward to use\" (Allwright 297). Oakman also cautioned potential users that \"[e]ven with the helpful user's manual, the inexperienced scholar will find the going rough at first\" (Oakman 182).
Allwright's final assessment of FISHER was positive:
As an apparently useful set of routines for anyone wishing to handle characters and character strings, FISHER provides the programmer with facilities which otherwise he must expend time and energy creating fro himself each time he sets out to write a text-handling program. It would be a useful addition to a program library wherever character-handling computer users may be found. (Allwright 298)
Oakman agreed on FISHER's value to FORTRAN users, though he retained reservations with regard to using FORTRAN for text processing:
Compared to a program package like JEUDEMO or CLAS, in which the computer can be instructed in a few commands to do a complex index, in terms of convenience FISHER comes out in second place, although it is probably more machine efficient. On the other hand, for people whose only option is FORTRAN, Day has already done much of the hard work of modifying the language for text processing. For those scholars, FISHER could make the difference in being able or not being able to carry out any analysis on the computer. (Oakman 182)
Sources
Allwright, Christine. \"Software Review: FISHER: A String-Handling Package.\" Computers and the Humanities 10.5 (1976): 297-298. Web.
Oakman, Robert L. Computer Methods for Literary Research. Athens: University of Georgia Press, 1984.