Three categories of computer software are particularly useful in producing discourse: the word-processing program, style and mechanics checkers, and invention programs. While the debate continues about whether, why, and how word processing improves student writing, virtually all writers who become comfortable with a word-processing program perceive their tasks as being made easier by technology. Style and mechanics checkers have their place in final-stage editing of conventions. More directly related to developing quality writing are several invention programs that have become commercially available. In general, these programs have been created by pioneer scholar-teachers who use computers in composition classes. Recognizing each program's strengths and limitations, composition teachers may find the various categories of invention effective in bringing inexperienced writers into academic and professional discourse communities.

The simplest level of computerized invention is inserting prompts into a word-processing program. A related approach is to develop such prompts within the hierarchic structure of a computer outliner/idea processor or the uniquely electronic hypertext software. Most sophisticated are the branching, open-response programs, such as Burns's ground-breaking TOPOI (unpublished) and H. J. Schwartz's SEEN and ORGANIZE, based in current rhetorical theory as applied to teaching composition.

Decisions about what, if any, invention software to use for inexperienced and developing writers depend on many factors; among them, the institutions' and individual teachers' rhetorical emphases and the kinds of hardware, funding, and computer time available for the writing program. Two warnings are fundamental: First, computerized invention will not teach writing; the planning programs' use will not be intuitively obvious to inexperienced writers. Second, these programs are not intelligent; they cannot evaluate good or mediocre writing, ask perceptive questions, or suggest global changes to the author. Spitzer (1989) points out that such "programs try to build on what students already know, and they try to stimulate students to see the relationships among those thoughts" (p. 206). When novice writers are using such software, human explanations and responses are crucial. But within these limitations, invention software can help in the teaching of writing. As H. J. Schwartz reported in 1984, "With computer aids . . . we can combine the systematic coverage of a writing class with the individualization of a writing workshop" (p. 239) by providing a directive environment in which inexperienced writers may develop their skills while taking some risks.

First, consider the simple process of creating what Sudol (1987) calls a rubric or template file for prompting in a word-processing program. In TEXTFILES, students are directed to open a word-processing file and enter six rhetorical categories: situation, purpose, message, audience, persona, and design (p. 54). This template is saved on the student's disk and called up whenever that student needs to begin exploring a topic. When the exploration is finished, the file is saved under a different name to preserve the template file. Material from the preliminary writing file can be reviewed on screen, printed, and/or transferred into the word-processing essay file.

Such rubric files offer writing teachers maximum flexibility for prompts. The rubric can be as simple or as complicated as the instructor wants and can be keyed in by the student or copied from a prepared disk (the simplest sense of authorable). There are no additional costs or hardware compatibility problems. Further, such prompting is completely portable; students can use whatever word-processing system they have access to, whenever it's convenient. No special licenses, hardware, or copying arrangements are necessary, nor do students have access to the planning material only in a specially designated computer lab or classroom. Such preliminary writing files also mean less software teaching time: The student learns only one word-processing program for both prompting and creating the discourse.

Additional advantages accrue, which generally also appear in the true prompting software. Students who are bothered by scribing errors can easily correct them. More important for generating the text, the student can easily review the material and insert, delete, or move the text as ideas occur. Inserting responses to the prompts depends on word wraparound, so responses can be as extensive as the student is willing to provide.

However, some features of prompting software cannot be duplicated in word-processing programs. No branching or hidden areas are possible. All the prompts appear on screen, without the directive focus of branching programs. No help screens are available. And, perhaps most important, there is no illusion of dialogue between the writer and the prompter; open-response software often inserts specific fields either into later prompts or on subsequent screens, directing the student in the chosen topic--unintelligently, of course, but as Kemp (1987) argues, such direction can be very powerful (p. 38).

Outliner prompting shares some features of word-processing templates in that any content is provided by the student writer or the instructor rather than by the program. Unlike word-processing programs, though, outliners present inherent hierarchic structure for organizing prompts and responses. (Sometimes electronic outliners are called idea organizers or thought processors; hypertext or nonsequential idea organizers are examined later in the hypertext section.) The outline window shows a traditional outline indenting pattern as indicated in Figure 1.

Figure 1. Example of electronic outliner.

The material can usually be viewed and/or printed in a variety of formats from the standard Roman numeral/alphabet system to the bullet series to the decimal numbering patterns. Subcategories are indented appropriately, and most outliners have collapse/expand features.

Outlining programs for personal computers are readily available, relatively inexpensive, and fairly easy to use. Although many are stand-alone programs, some of the true prompting programs include outliners, and some word-processing programs integrate outliners. Outline prompts are not quite as portable as word-processing rubric files because students may not have access to outliners outside of a computer lab. Outliners usually require some additional expense and teaching time as well--although most students come to college with a basic understanding of outlining and its uses--so instruction can probably be limited to technical operations.

Outliners can be used with rubric, teacher-developed prompts or as blank files in which writers generate their own content. For writers who work well with such structures, outliners could be used to stimulate early invention, but for some novice writers, such a formal structure might be inhibiting. Outliners might be used more profitably by many writers at a later stage to organize material from a less structured preliminary writing or even to aid in draft revision. Students often welcome outliners because the terror-inducing outlining conventions are handled for them by the software, and they are left to concentrate on content.

Figure 2. MACWRITE essay and ACTA outline windows on single screen.

As with word-processing prompt files, outliners permit easy correction of scribal errors. It is easy to add, delete, and move material; moreover, response length is unlimited. Material produced in an outliner usually can be easily imported to a compatible word-processing program and (depending on the software) often can be displayed in a window simultaneously with a word-processing file. In Figure 2, the MACWRITE essay window and the ACTA window appear on the screen at the same time; the writer can make the partially hidden window active--in the front, fully visible--by clicking with the mouse arrow on the window. Transfer of material is quick and easy through the Macintosh cut-and-paste clipboard system. Visual access to preliminary writing files during drafting might well encourage novice writers to make more use of their planning materials.

A most valuable, uniquely electronic feature of outliners is the collapse-expand capability. Users can go back to collapse, or hide, some of the sublevels to highlight the surface structure. Returning to expanded sublevels is generally quick and easy. Such collapsing can be done in most word-processing programs, of course, though not very efficiently, and some prompting software, such as HBJ WRITER, incorporates collapse-expand features. HBJ WRITER 's outliner, presented as a revising tool, will collapse an essay. The program selects the first sentence of each paragraph--or permits the user to select one sentence per paragraph--and displays these sentences in outline format (Friedman, 1987).

Electronic outliners can be useful tools for invention, but most applications of current rhetorical theory are not inherent in the software. Any branching or hidden areas would have to be designed by the instructor and would not necessarily be used by novice writers in any intended sequence. Finally, as with word-processing software, outliners have no built-in illusion of dialogue.

In 1986, Langston called for "developing new, uniquely electronic paradigms for future aids" (p. 261) to invention. Probably the closest thing to a uniquely electronic environment for planning has come through hypertext programs, nonsequential environments where text, graphics, and access to other media such as videodisks coexist harmoniously. Some hypertext environments are already available commercially. In addition, HYPERCARD gives nonprogrammers the tools with which to create hypertext/hypermedia stacks on virtually any subject.

Hypertext programs such as GUIDE and LEARNING TOOL are, like outliners, form without substance. LEARNING TOOL, for example, provides a master-list window next to a note-card window. The note cards may be grouped, linked with lines, and stacked. Stacked cards, whose titles appear as indented headings in the master list, can be seen graphically in a submap. Each note card has a detail window which can contain extensive text. Such software, as Figure 3 illustrates, may be adapted as a writing aid for novice writers.

Figure 3. LEARNING TOOL windows.

Figure 4. STORYSPACE screen.

Joyce's STORYSPACE, a visual structural editor that includes graphic and textual environments, specifically emphasizes the writing process. Joyce uses the program in his classroom primarily for brainstorming. He believes that the "ability to graphically view connections between ideas helps his students to understand the basics of developing an argument" (cited in Kaplan-Neher, 1990, p. 6). The STORYSPACE screen in Figure 4 shows only its simplest linking and text spaces; print cannot effectively represent complex electonic linking.

Speaking of hypertext generally, Joyce (1988) notes that these "tools offer the promise of adapting themselves to fundamental cognitive skills which experts routinely, subtly, and self-consciously apply in accomplishing intellectual tasks" (p. 10). STORYSPACE is designed to foster "constructive use . . . of hypertext as an invention or analytic tool" (p. 11).

LEARNING TOOL shares some features of outliners. As with the general hypertext programs, the instructor has to design the prompting pattern and any hidden areas, and these may be altered intentionally or in error by the user. The program cannot force the user to adhere to a sequence, nor can it ensure that the user covers what the instructor might consider essential. When taking protocols for the NCRIPTAL study in 1988, I observed that many inexperienced writers failed to look at prompts in detail cards, possibly because they either were not adept at using the program or because they were not convinced that thinking about audience, for example, was helpful in planning the writing.

At this stage, there are few writing-specific hypertext alternatives, though some faculty are developing HYPERCARD stacks for their writing classes. Hypertext/hypermedia environments have the potential, according to Barrett (1989), to "support the social construction of meaning that characterizes understanding and communication in the larger world beyond the computer screen" (p. xiv), particularly when users work--as he strongly recommends--on a network (p. xvi). Great advantages in such environments include the nonsequential access to an array of material, the audience-driven nature of such access, and the easy access to different source media. Carlson (1988) and Younggren (1988), among others working with technical documentation, laud hypertext because the user determines the "paths through the content[:] [n]ovice users [could] be given one possible route to use, experienced users another" (Younggren, 1988, p. 78). Hypertext planning environments can be more or less directive and are, as Slatin (1988) notes, "beautifully appropriate for collaborative work" (p. 128). Some care needs to be taken in designing writing-specific stacks, of course, to avoid being prescriptive or overly complex for the inexperienced writer. As a result, the hypertext environments may facilitate the kind of creative planning Arms (1988) saw in her engineering students who began with drawings and wrote around them (pp. 43-44).

Branching, open-response, or true prompting programs provide electronic versions of current rhetorical planning models. In general, such programs have significant advantages over their paper-and-pen counterparts. Their uniquely electronic features build effectively on traditional invention strategies, making planning more accessible to the inexperienced writer.

Prompting programs present users with various heuristics--systematic and unsystematic, generalized and tailored to specific kinds of writing tasks. The software ranges from programs like HBJ WRITER --which integrates word- processing, some text-analysis features, and several invention activities--to stand-alone programs such as ORGANIZE by Helen Schwartz. Most of the stand-alone prompters permit transfer of the generated material to a word-processing program.

Depending on the program, almost any current rhetorical invention model is available electronically, from free-writing to generalized prompts to prompts developed (usually by the instructor) for specific assignments. WRITER'S HELPER incorporates programs for all three kinds of prompting. The BRAINSTORMS program is free-writing. It measures typing speed, prints periods when the writer hesitates (to prompt more writing), and does not allow use of cursor movement keys or backspacing to edit. The program then asks focusing questions about the three-to-ten-line free-writing. A second kind of prompting illustrated in WRITER'S HELPER is general but systematic; "Three Ways of Seeing" gives the user twelve to fifteen questions in 3 groups based on the Young, Becker and Pike tagmemic scheme. Finally, WRITER'S HELPER has an authorable exercise for exploring subjects. The program has a set of built-in, one-line general questions, and an instructor may author additional files of up to twenty short questions tailored to a specific assignment.

Most of the prompting programs present a variety of approaches from which the user may choose. WRITER'S HELPER, with more than ten exploring activities, is the most comprehensive. Such software helps to individualize the classroom; as Strickland (1984) indicates, a variety of prompts supports individual composing styles. Further, having the options on a screen and using menus or commands to get to the invention system provide a focus that printed material presented on several pages or in various chapters cannot.

What programs are available for what hardware at what cost changes frequently. The "Invention Software List" at the end of this paper includes many of the programs now available. It details some of their features, but it is no substitute for trying the software to determine whether the commands are overly cumbersome and to see whether the rhetorical approach suits the instructor and potential student users.

Electronic prompting programs have many of the advantages of word-processing prompts with additional features. Editing functions are generally adequate though occasionally limited, depending on the program. In UNC-WRITE (now published as the WRITING TUTOR), for example, the word- processing window comes on the screen for the user's responses, so full-feature editing is always possible. Yet in some other programs, cursor movement keys cannot be used for editing; changes can be made only by deleting everything from the end of the text back to the place where the change is required. In some cases, also, no editing is permitted when it is pedagogically appropriate, as in WRITER'S HELPER 's BRAINSTORMING. The editing scheme for the software should probably be discussed with students before they use the software; if not, they may expect full-function editing and equate the software's effectiveness with its editing capabilities rather than with its purpose.

Response lengths vary in the invention software, often depending on the pedagogical purpose of the prompts. Several of the programs impose artificial limits to encourage focusing or expanding, establishing maximum or minimum response lengths. With ORGANIZE, in the "Basic Terms," the program asks for one idea about the writing and limits the response to one line. Kappel's WRITING IS THINKING at one point asks for a single sentence expressing an opinion on the topic and will not let the writer proceed without writing more than one line. Clearly such limits are not intelligent, and it is incumbent on the instructor to clarify the pedagogical purposes for the novice writer.

True prompting programs also may help the user to focus by presenting prompts in sequence on separate, progressive screens. The writer is led through a series of questions or topics that may branch, depending on the writer's responses. Also common is a quasi-intelligent interactive routine, which inserts fields from earlier responses to personalize the later prompts. In SEEN, for example, user responses are incorporated in subsequent prompts, giving the illusion of dialogue and encouraging the user to focus. The user keys in responses to three prompts: "Name a character," "In what work?" and "What hypothesis?" The subsequent screens insert the answers in later prompts, such as "What does [the character] do that shows she/he is [hypothesis]?" Obviously, this uniquely electronic linking of fields is artificial, even to the least sophisticated student user. Kemp (1987) acknowledges that such programming appears "modest and unexceptional" (p. 38), but Kemp finds the software "instructional" in the best sense: After going through a prompt series a few times, the user begins to internalize the "self-questioning process" (p. 36).

In many prompting programs, on-line explanations and help are available for the student who is unclear about what responses might be appropriate. Unlike paper prompts, the explanations are isolated for the writer's viewing and linked directly to the appropriate prompt screen; the student does not have to search pages or paragraphs of text for examples. In most of the software, the author and/or programmer has prepared these materials, but Helen Schwartz's SEEN makes even the help screens authorable by instructors without programming expertise. Of course, examples and explanations are not unique to computers, but the student's access to the information is focused and facilitated by the electronic display.

Another popular electronic feature in prompting programs, which is becoming more common as college computer laboratories install networks, is the electronic bulletin board. SEEN allows for such a bulletin board without physically linked computers. Langston (1987), as an ENFI project participant, has described using COGNOTER and the CB software to support on-line invention activities for groups. Kemp (1988) and Taylor (1988) have implemented real-time conferencing programs that can be used for invention activities, and the Daedalus Group (1988) has developed INTERCHANGE to permit the collaborative activities of text sharing and CONTACT for electronic mail. Whatever the network sharing scheme, computerized collaborative invention activities can enfranchise students who might otherwise remain silent in an oral group.

One feature touted as a benefit for most invention programs is the ease with which writers can import materials into word-processing programs. For example in INVENT, which Spitzer (1989) has been developing, the "disk files . . . can . . . be loaded as text files into . . . word processors [, so] students can use the output from the program as the basis for the first draft of their essays" (p. 208). This feature seems to have some advantages for the student, presumably eliminating some retyping. But here, the teacher needs to be vigilant because inexperienced writers often regard anything written as a draft. A series of responses to rhetorical prompts copied and set in paragraph form is unlikely to be a coherent essay, even in an early stage, unless the prompts have been written to result in an essay. And most writing instructors would not condone the prescriptive nature of such prompting. The ease of importing planning responses into word-processing programs seems unimportant, then, in choosing a prompting program for novice writers.

Most invention programs are still stand-alone software, but more integrated programs are appearing. HBJ WRITER; THE WRITING TUTOR; WRITER'S HELPER; Bank Street's WORDBENCH with the utility BRAINSTORMER; and Holt, Rinehart and Winston's PROCESSWRITER, with its note-card planning environment, provide writers with a word-processing program and various invention and revision tools in a single package. Yet another possibility is using HYPERCARD to individualize integration. Such integration allows for movement among activities and has philosophic value in implying the reflexive nature of composing.

The composition teacher whose students have access to computers has a substantial and varied group of planning programs from which to choose. Those who do not want to use any of the traditional, prepared planning prompts can obtain authorable programs or work with outliners or hypertext software. Burns (1987) describes the word-processing program as a "dynamic tool" (p. 378); Helen Schwartz (1989) finds that word-processing programs support "playful risk taking" (p. 198). Computerized invention has many of the same advantages for developing writers. With enough time for students to become adept with planning software and enough instruction for them to learn its uses and limits, today's invention programs can give important assistance to composition students.

Wallis May Andersen teaches at Oakland University.

Arms, V. (1988). The right answer to the wrong question. Computers and Composition, 6(1), 33-46.

Barrett, E. (Ed.). (1989). The society of text: Hypertext, hypermedia, and the social construction of information. Cambridge, MA: MIT Press.

Burns, H. (1987). Computers and composition. In G. Tate (Ed.), Teaching composition: Twelve bibliographic essays (pp. 378-400). Fort Worth, TX: Texas Christian University Press.

Carlson, P. A. (1988). Hypertext: A way of incorporating user feedback into online documentation. In E. Barrett (Ed.), Text, conText, and hyperText: Writing with and for the computer (pp. 93-105). Cambridge, MA: MIT Press.

Deadalus Group, Inc. (1989). INTERCHANGE. Austin, TX: Author

Friedman, M. P. (1987). WANDAH--a computerized writer's aid. In D. E. Berger, K. Pezdek, & W. P. Banks (Eds.), Applications of cognitive psychology: Problem solving, education, and computing (pp. 219-226). Hillsdale, NJ: Erlbaum.

Joyce, M. (1988, November). Siren shapes: Exploratory and constructive hypertexts. Academic Computing, pp. 10-14, 37-42.

Kaplan-Neher, A. (1990, March/April). Emerging technologies and their impact on writing across the curriculum. Syllabus for the Macintosh, pp. 5-8.

Kemp, F. (1987). The user-friendly fallacy. College Composition and Communication, 38, 32-39.

Kemp, F. (1988, March). Using networked micros in the computer-integrated classroom. Paper presented at the meeting of the Conference on College Composition and Communication, St. Louis, MO.

Langston, M. D. (1986). The old paradigm in computer aids to invention: A critical review. Rhetoric Society Quarterly, 16, 261-284.

Langston, M. D. (1987). Invention aids for computer-based writing: Expanding the horizons through collaborative invention. (ERIC Document Reproduction Service No. ED 280 055)

Schwartz, H. J. (1984). Teaching writing with computer aids. College English, 46, 239-247.

Schwartz, H. J. (1989). Creating writing activities with the word processor. In C. L. Selfe, D. Rodrigues, & W. R. Oates (Eds.), Computers in English and the language arts: The challenge of teacher education (pp. 197-203). Urbana, IL: National Council of Teachers of English.

Slatin, J. (1988). Hypertext and the teaching of writing. In E. Barrett (Ed.), Text, conText, and hyperText: Writing with and for the computer (pp. 111-129). Cambridge, MA: MIT Press.

Spitzer, M. (1989). Incorporating prewriting software into the writing program. In C. L. Selfe, D. Rodrigues, & W. R. Oates (Eds.), Computers in English and the language arts: The challenge of teacher education (pp. 205-212). Urbana, IL: National Council of Teachers of English.

Strickland, J. (1984). Problems and promises: Invention on the computer. (ERIC Document Reproduction Service No. ED 273 987)

Sudol, R. (1987). Textfiles. San Diego: Harcourt Brace Jovanovich.

Taylor, P. (1988, March). FORUM: Practical conversation software for the computational classroom. Paper presented at the meeting of the Conference on College Composition and Communication, St. Louis, MO.

Younggren, G. (1988). Using an object-oriented programming language to create audience-driven hypermedia environments. In E. Barrett (Ed.), Text, conText, and hyperText: Writing with and for the computer (pp. 77-92). Cambridge, MA: MIT Press.

Bank Street College Media Group. (1987). WORDBENCH. Reading, MA: Addison-Wesley.

Apple IIe, c, GS (128K); IBM PC (256K). INTEGRATED PROGRAM. Outliner, reference database, note cards. Word processor. Brainstormer "add-on utility": goal setting, nutshelling, invisible writing.

Bolter, J. D., Joyce, M., & Smith, J. B. (in press). STORYSPACE. Cambridge, MA: Eastgate Systems.

Macintosh (512K). Interactive hypertext for brainstorming, organizing, analysis, presentation. Outlining, visual mapping, interactive notes.

Daedalus Group. (1988). CONTACT. Austin, TX: Author.

IBM, file-sharing microcomputer networks. Electronic mail for Local Area Networks.

Daedalus Group. (1988). INTERCHANGE. Austin, TX: Author.

IBM & Macintosh, file-sharing microcomputer networks. Real-time communications program with private and "publishing" windows.

Daedalus Group. (1988). MINDWRITER. Austin, TX: Author.

IBM & Macintosh, file-sharing microcomputer networks. Invention: interactive "thought prompts" based on Aristotle, Burke, and Pike.

Friedman, M. P., Rand, E., Von Blum, R., Cohen, M., Gerrard, L., Magpantay, A., Cheng, S., Pisano, A., & Mak, L. (1986). HBJ WRITER. Chicago: Harcourt Brace Jovanovich.

IBM (256K). Integrated program. Invention: freewriting, nutshelling, planning argumentation, invisible writing. Word Processor. Reviewing & revising: spelling, punctuation, word use analysis, outliner, nutshell, transitions.

Kappel, M. (1984). WRITING IS THINKING. Dallas, TX: Kapstrom.

Apple II+, e, c; IBM PC, PCjr. Invention: brainstorming, focusing question, structuring/outlining patterns, "fact files," 2 introductory paragraphs, audience review.

Kozma, R., & Van Roekel, J. (1986). LEARNING TOOL. Ann Arbor, MI: Arborworks.

Macintosh (512K). Hypertext "interactive database": outliner and note cards.

Learningways. (1989). PROCESSWRITER. New York: Holt, Rinehart & Winston.

IBM PC (512K & graphics adapter), PS/2; Macintosh 512Ke. Integrated program. Note cards. Word processor.

Leonard, Robert J. (1987, 1988). PROTEUS: The Idea Processor. Stony Brook, NY: Research Design Associates.

Apple II+, e, c; IBM; Tandy; TRS-80; C64. Invention: freewriting, looping, listing, cubing, 5Ws.

Lindemann, E., Wittig, J., & Avery, L. (1989). THE WRITING TUTOR. Chicago: Harcourt Brace Jovanovich.

IBM PC (256K). Integrated program. Invention: Tagmemics, TOPOI, heuristics, problem solving, freewriting, drawing on personal experience. Word processor.

Owl International (1988). GUIDE 2.0. Bellevue, WA: Author.

IBM PC-AT, PS/2; Macintosh (512K). Hypertext system.

Schwartz, H. J. (1989). SEEN: TUTORIALS FOR CRITICAL READING. Iowa City, IA: CONDUIT.

Apple IIe, c, gs (128K); IBM PC, PC-XT, PS/2 (256K), ProDOS & MS-DOS networks. Invention: 11 tutorial questions (authorable) and network bulletin board.

Schwartz, H. J., & Nachman, L. J. (1988). ORGANIZE. Belmont, CA: Wadsworth.

Apple IIe, c (128K); IBM (DOS 3.0+, 128K). Invention: definition, analysis, description, comparison & contrast, narration, process, cause & effect, audience analysis, attitudes analysis, argument, outlines, free writing.

Schwartz, M. PREWRITE. Princeton, NJ: Author.

Apple II, II+, e, c (48K). Invention: brainstorming, questions, journalistic heuristic, purpose & audience analysis, metaphors, freewriting, argument, cause & effect.

Symmetry Corporation. (1987). ACTA. Mesa, AZ: Author.

Macintosh 512K. Desk accessory outliner.

Wresch, W. (1989). WRITER'S HELPER II. Iowa City, IA: CONDUIT.

Apple IIe, c, gs (128K); IBM PC, PC-XT, PS/2 (256K), ProDOS & MS-DOS networks; Macintosh (512K, 1 800K drive). Invention: starters, idea wheel, associations, questioner, lists, brainstorms, crazy contrasts, three ways of seeing, audience, random revelations, connections, teacher. Organize/draft: trees, debating an issue, structure guide, goals, compare/contrast, outliner, develop a paragraph, five-paragraph theme. Revising: structure, wording & conventions.