A Brief History of Intergrams

Intergrams was written during the period 1988 - 1992 and originally published by Eastgate Systems on floppy disk. It is my earliest fully interactive work, and was developed, like my other works of that period, in the Macintosh-only environment called Hypercard. However the origins in my work of many of the concepts in Intergrams go back much further. Early experiments with the diagram notation (done with pen and paper) date back to 1968. The motivation for deploying an “external” diagrammatic syntax came from a desire to translate into poetry a concept that composers were using with what seemed to me like complete ease: tone clusters. Hearing the term ‘tone cluster’ triggered a virtual explosion in my head: I simply had to do word clusters! But this is much easier said than done. When a painter puts down color on top of color, the result is still colored space. When a composer puts sound on top of sound, the result is still sound. But putting a word on top of another word does not result in any kind of language object that would be familiar — it certainly isn’t a word — and is likely to be mostly illegible. In addition, there is a structural problem. The part of speech of a word is partly brought by the word itself, and partly brought by context. What is the “part of speech” of a word cluster? This question loomed as a real conundrum. An obvious solution to this dilemma was to indicate syntax explicitly, with an external notation. So began the Diagram Poems, which have become the core of my lifelong work. As far as I can recall, the earliest public showing of any diagram poems came with an installation at The Kitchen in New York called Permanent and Temporary Poetry 5/75 which was done in conjunction with a performance. Diagrams Series 3 was published in 1979, and Diagrams Series 4 in 1984. While by Diagrams Series 3 I was already working on a computer, and the poems were reproduced as computer printouts, these were still completely works on paper; the original concept that started the whole thing, word clusters, was implemented only very approximately, using graphical positioning of what I would rather have made be layers on top of one another. Approximately in 1987, while playing with a small drawing program in an environment called Smalltalk V — which actually ran under MSDOS! — I realized that by means of interactivity I could finally actually implement the original motivating concept: word clusters. At this point I had two possible pathways: spend years learning the class libraries of Smalltalk and implementing my own classes to do word clusters, or buy a Macintosh (which I didn’t own at the time) and simply hope that Hypercard — which I had heard about but never used — would do the job. It didn’t take long to work this out: if I went the Smalltalk way I would spend years writing code and not actually making poems. So began Intergrams.

As originally implemented, Intergrams required Hypercard to run, so was only available on the Macintosh. In 1996, Intergrams was released for Windows. This was accomplished using an environment called Oracle Media Objects (OMO). OMO had no ability to read Hypercard project files (called “stacks”), but had an almost identical basic model, and could understand simple Hypercard scripts almost unchanged. It was relatively simple to create a script in Hypercard’s programming language (called Hypertalk) which would walk my stacks, dump the graphics to graphics files, and create a description of the behavior of each “object” (in the loose sense of the word) as a text file in a simple file format designed for this project. A second OMO script would read these files, import the graphics, and recreate the stack in OMO. This was possible (and in fact not difficult) because both Hypercard and OMO may be said to be “self-describing”: it was possible to write a script in Hypertalk which would walk through e.g. every card, or button, and these were objects whose properties were accessible in Hypertalk. (And likewise for the programming language of OMO.)

Oracle discontinued support for OMO, and more grievously, Apple did not port Hypercard to OS X, and discontinued support for MacOS Classic. Hypercard became a completely stranded platform. However, the methodology used to port Intergrams from Hypercard to OMO showed how the same materials “dumped” from Hypercard could be interpreted in Squeak to create a portable cross-platform version of Intergrams that will run on both new and older computers.