3(3) p. 33

Matching Software and Curriculum: A Description of Four Text-Analysis Programs

Joyce Kinkead

The great search for software is on as writing program administrators ask the question, "What is the best software for writing centers?" The answer? It depends. It depends on money, use, environment, and staff. Almost any word processor makes revision simpler as students find it easier to delete, yank and put, and add--although teachers must remember easier revision does not necessarily mean better revision. Anyone looking for software needs to ask one all-important question: "What are the goals in a writing course and what software meets those goals?"

Four schools have answered that question with four different answers: one computer writing lab uses ALPS software, another GRAMMATIK/COMMENT; a third college has WRITER'S WORKBENCH (WW), and a high school designed its Writing Room around UCLA's WANDAH (marketed as HBJ WRITER). Of these, all are on microcomputers except for WW, which is accessible on a mainframe.

p. 34

Weber State College of Ogden, Utah, started in 1983 to match software with writing programs. Unsuccessful in their search, they collaborated with ALPS (Automated Language Processing Systems) of Provo to develop a text-analysis program for use with their composition program. Originally designed for Burroughs Convergent Technologies terminals, ALPS transferred the program to the Macintosh and released it on the market in 1985. In doing so, they revised the program, deleting a readability analysis and a style chart.

The result matches the goals of Weber's writing program, centering on freshman composition and focusing on expository writing. Set up in one room, the Macintosh terminals are linked to an Apple Lisa (with hard disk) and a laser printer. In an ALPS brochure, they advertise that such a lab with 15 terminals can support up to 500 students.

Using the MACWRITE word processor, students find problems in their essays as identified by MACPROOF; its four programs include USAGE, GRAMMAR, STRUCTURE, and SPELLING. The USAGE program contains "sexist," "racist," "confused," "vague," "overworked," and "discouraged" word analyses. In the "confused" program, students find answers to problems such as their/there/they're or to/too/two. The "vague" program looks at abstract words while the "overworked" program focuses on slang words and clichés. For the "discouraged" program, faculty include words--like copout--that they find problematic in student writing, an important feature for the writing faculty who demand control over standards choices.

The "GRAMMAR" analysis reviews "be" verbs and

p. 35

"nominalizations," first explaining why passive voice, expletives, and nominalizations may not constitute effective style. The STRUCTURE program focuses on sentences: in the first part, "abridgment," the first and last sentences of each paragraph are printed to check cohesion of paragraphs and the entire essay; in the second part, "expansion," sentences are printed out separately with each numbered to check fragments, run-ons, parallel structure, variety, and length. STRUCTURE also shows numbers of words, sentences, and paragraphs. Finally, a spelling checker, with a dictionary of 41,000 words, finds potential misspellings; however, ALPS advertisements insist that the dictionary is effective for more than 86,000 words because "words in the text are reduced to their base form before checking and since there are, on the average, more than two inflected forms for every base word in English."

For these text analyses, students choose "highlight," "list," or "interpretation." The "highlight" option prints the paper with the word underlined. For example, if the student is working on nominalizations, then the essay is shown on the screen with the nominalizations underlined. Or, the student might choose "list" instead; in that case, the nominalizations are simply listed on the screen. Or, if the student does not understand nominalizations, then "interpretation" prints an explanation of nominalizations.

Simple to learn, the ALPS program has its drawbacks. Even though punctuation checkers are still rudimentary, ALPS does not locate problems with punctuation and quotation marks, unbalanced parentheses or quotes, capitalization, or punc-

p. 36

tuation and spacing. As for sentence structure, the program fails to identify sentence beginnings or sentence types. However, the biggest drawback for the writing program on a tight budget is money. The ALPS annual licensing fee is not inexpensive, and the networked microcomputer set-up amounts to a good deal of cash.

Because the primary goal of the Weber writing program is to clean up student writing, MACPROOF meets the needs of identifying surface problems and suggesting improvements. Teachers feel that with many mechanical errors taken care of, they are "free to focus on issues of contents and organization." Finally, teachers like the control they have in a centralized writing lab, where tutors can answer both computer and writing questions. (Table 1 lists the characteristics of ALPS and the other three programs to be discussed.)

The second college--Utah Tech at Salt Lake--which has many of the same goals as Weber, uses a combination of GRAMMATIK and COMMENT, marketed by Digital Equipment Corporation, in a centralized writing lab for use with their DEC Rainbow terminals. GRAMMATIK was developed by Wang originally for business users who wanted to proofread documents, and Tommy Barker at Texas Tech wrote the COMMENT program to explain and elaborate on the GRAMMATIK messages.

After entering text using EASY RIDER, or another word-processing program, the student puts the essay through a spelling checker--the RANDOM HOUSE DICTIONARY version; then the paper is analyzed by GRAMMATIK. The student receives a short analysis (usually 1-2 pages) that focuses on three areas: typographical errors, style, and statistics on word and sentence

p. 37

                                   TABLE 1

                      A Comparison of Software Features

   Program            ALPS          GRAMMTIK/COMMENT      WW	     WANDAH

Spelling	     41,000 words  needs proofreader  l00,000 words 39,000 words
Sexist/gender           x                 x               x           x
Racist                  x
Vague                   x                 x               x
Overworked words        x                 x               x
Confused words          x                 x               x           x
Slang/jargon            x
Improper usage                            x               x
"Be" verbs              x                 x               x           x
Passive                                                   x
Nominalizatons          x                                 x           x
Organization            x                                 x           x
Sentences separated     x
Expletives                                x (th- only)                x
Transitions                               x               x
Pronouns                                                              x
Prepositions                              x
Readability                               x               x
Abstract                x                                 x           x
Sentence variety                  (only short sen.)       x           x
Sentence openers                                          x
Double words                              x               x
Double punctuation                        x               x
Capitalization                            x               x           x
Punc. and ""                              x               x           x
Balanced "" ()                            x               x           x
Split infinitives                                         x
a, an                                                     x
counts (words, sen.,
  essay length)         x                 x               x           x
Paragraph development                                     x           x
                       - - - - - - - - - - - - - - - - - - - - - - - - - - -
Comment                                                               x
Prewriting aids                                                       x
Word processor                                                        x

p. 38

length. The typographical errors include double words, double punctuation, inconsistent capitalization, balanced quotation marks and parentheses, and placement of punctuation marks inside quotation marks. STYLE picks out archaic usage, gender-specific terms, informal usage (ain't), awkward usage (and/or), commonly misused words (affect), overworked or trite words, redundant phrases (seldom ever), improper usage (must of), vague adverbs, and wordy phrases (a number of). Additionally, it counts sentences and words, the average sentence length, the average word length, and the number of questions and imperatives; it identifies sentences of shortest and longest words plus the number of short sentences (defined as under 14 words) and long sentences (over 30 words). Finally, it counts "be" verbs, prepositions, vague words, "th-" openers, transitional words, and subordinating conjunctions. Figure 1 shows a GRAMMATIK print-out. Although there are no suggestions in this last summary section, the "punctuation" and "style" sections include revision possibilities; for example, the suggestion under "wordy phrase" might be "this phrase often indicates wordiness" or "omit if followed by a verb or adjective."

From this summary section, the student moves to the COMMENT analysis. First, the program asks such questions as how many sentences are in the essay; how many words; how many questions. The student, using the information from the GRAMMATIK analysis, types in the correct numbers. From these statistics, COMMENT makes generalizations about the writing sample. The "readability" score rates the sample on a scale of 0 (hardest) to 100 (easiest); for example, a

p. 39

-------> .
*At sentence 2 - `P' - Punctuation error
*Suggestion: needs separator
-------> spacing
*At sentence 2 - `C' - Capitalization error
*Suggestion: 1st word of sent
-------> one of the
*At sentence 3 - `W' - Wordy phrase
*Suggestion: one, a
-------> this
*At sentence 7 - `E' - Error - unspecified
*Suggestion: vague pronoun: specify referent, as in "this x"
-------> there are
*At sentence 9 - `W' - Wordy phrase
*Suggestion: this phrase often idicates wordiness
-------> which are
*At sentence 9 - `W' - Wordy phrase
*Suggestion: omit if followed by a verb or adjective
-------> one of the
*At sentence 11 - `W' - Wordy phrase
*Suggestion: one, a
-------> in order to
*At sentence 12 - `W' - Wordy phrase
*Suggestion: to
-------> this
*At sentence 12 - `E' - Error - unspecified
*Suggestion: vague pronoun: specify referent, as in "this x"
-------> it is
*At sentence 13 - `W' - Wordy phrase
*Suggestion: this phrase often idicates wordiness
-------> this
*At sentence 14 - `E' - Error - unspecified
*Suggestion: vague pronoun: specify referent, as in "this x"
-------> this
*At sentence 17 - `E' - Error - unspecified
*Suggestion: vague pronoun: specify referent, as in "this x"
-------> that are
*At sentence 19 - `R' - Redundant phrase
*Suggestion: omit if followed by verb, adverb, or adjective
Summary for B:EXAMPLE.EZF / Problems detected: 13

# sent: 22 ; # words: 0384
avg sent len: 17.4 ; avg word len: 5.1
# questions: 1 ; # imperatives: 0
short sent (< 14 wds): 7 ; long sent (> 30 wds): 1
longest 39 wds at sent ll 13; shortest 0 wds at # 1
to be's: 22; prepositions: 58
User category totals:
#4: 6 #5: 5 #6: 3 #7: 6

Figure 1. GRAMMATIK Print-out

p. 40

score of 45 suggests that the text is "difficult to read." Next, the "to be's" are counted, and the program suggests that good papers have 30% or less of these verbs and recommends avoiding passive voice by replacing "to be's" with more specific verbs. Consultations with the instructor or writing-lab personnel are recommended for more help. After that, prepositions are in the spotlight, where a score higher than two indicates that the writer may have used too many vague terms or nominalizations; a sample sentence and its revision are listed. In the "continuity" portion, the presence of transitional words is noted to show how well the ideas flow one after another. According to COMMENT, good writers use transitional words in about 20% of their sentences. An analysis of "expletives" comes under "th- openers," which suggests using under 9%; over 9% of "the, there, and this ... invites wordiness." Under "vagueness," COMMENT recommends a score of below 5%; a higher score may indicate "that the writing lacks clear details" and, consequently, a low interest value. Finally, under sentence variety, the text is evaluated for short sentences because "good writing on the freshman level contains, on the average, about 30% short sentences." If it contains more, then GRAMMATIK says the language may sound choppy and suggests revision using sentence combining.

From this feedback, the student is advised to revise the new file created by GRAMMATIK, which flags the errors with a ?# symbol. GRAMMATIK, like ALPS, focuses mainly on superficial problems; unlike ALPS, GRAMMATIK is not as simple to use. The print-out is cluttered, making it difficult for the writer to distin-

p. 41

guish among the error (at the --->), the identification of the error (at the first *), and the suggestion for revision (at the second *). An easier location indicator would be a line reference rather than a sentence number; but, errors can be found by the ?* symbol. Because the errors are not seen in the context of the text, they are also hard to understand.

Although COMMENT is presented in a more attractive and less-complicated format, the steps to get to COMMENT--typing in statistics--are awkward. (See Figure 2 for a portion of a COMMENT print-out.) On reading COMMENT, students are reminded that "some of these results may not apply to your work." Deficiencies in COMMENT include the following: (1) the expletive portion identifies only the th- openers, not "it openers"; (2) only short sentences are identified in "sentence variety," not sentence types; (3) nominalizations are alluded to in the "preposition" program but without explanation or identification in the text. Moreover, COMMENT is limited to running on CP/M even though GRAMMATIK runs on PC-DOS or CP/M microcomputers.

In contrast to the writing programs at Weber and Utah Tech that focus only on first-year composition, the program at Utah State University is a vertical one, meaning that students are required to complete writing courses beyond first-year composition; in fact, the university requires both first- and second-year writing classes while departments campus wide often require one or two more classes of upper-division writing--such as technical writing. This writing program almost demands a decentralized writing lab with computer access at several stations (including the English Department

p. 42

                               C O M M E N T

                         FIRST ANALYSIS OF example

       REMEMBER:  Some of these results may not apply to your work.
                  Use them as a guide only.  If you have questions
                  about what this program tells you, ask your professor
                  or visit the Writing Center (room 308).


EXAMPLE has a Readability score of 45.35 (scale: 0=hardest; 100=easiest).
This indicates that your text is `difficu1t to read'.


Your text used some form of `to be' verb in 70.00% of your sentences.
Papers rated excellent had 30.00% or fewer `to be' verbs. Since yours had
more, you may have overused passive voice.  To get help avoiding the pas-
sive voice, or replacing `to be's' with more specific verbs, consult with
your professor or visit the Writing Center (Room 308).


A count of the number of prepositons (of, to, in, for, etc.) in your sen-
tences shows that you have used about 3 prepositions per sentence, on the
average. A better score would be 2 prepositions per sentence.

Your score is excessive and may indicate that you also use too many vague
terms and nominalizations (nouns made from verbs).

Consider revising your sentences to eliminate strings of prepositional


     `IN this version, the sentence is IN need OF fewer prepositions.'


             `As is, this sentence needs fewer prepositions.'


The presence of transitional words, (`also', `on the other hand', etc.)
often indicates how well your ideas flow one after another. Good writers
use transitional words in about 20% of their sentences.

Figure 2: COMMENT Print-out

p. 43

Writing Center, which houses 21 terminals); consequently USU looked for a mainframe text-analysis program to suit first-year writers as well as technical writers. The answer? WRITER'S WORKBENCH.

Originally developed for Bell Laboratories technical writers, WW was revised at Colorado State University by Kiefer and Smith for academic writing (marketed as the Collegiate Edition by AT&T). Some schools purchased the original WW version rather than the revised one; missing from the original version are the DEVELOPMENT, VAGUENESS, and CHECK programs (although DICTION and SUGGEST may seem to be new programs, they are actually the old WORD CHOICE). Moreover. the prose standards suggested for WW users in the original are those standards for Bell Lab technical writers; consequently, first-year composition students are compared to on-the-job technical writers rather than to other first-year composition students unless the school--as the University of Kansas has done--revises the standards of its own program. (A school must own the compiled version of WWto alter standards.)

Divided into sections like COMMENT, WW is easier to read than GRAMMATIK although both offer similar analyses. Using a dictionary of 100,000 root words, WW identifies possible spelling errors (students can add words not in the dictionary); for example, students' names will appear as possible spelling errors because their names are not listed in the dictionary. Technical words absent from the dictionary can also be added to the student's personal dictionary, a plus for USU writing students, often engineering or science majors. The `punctua-

p. 44

tion" program, describes the punctuation of the text, noting double quotes, single quotes, apostrophes, and parentheses. WW notes typographical errors, such as spacing, and serves as an instructional prompt for apostrophes, a reminder for the student to check apostrophes for correct use. In this section, sentences with incorrect punctuation or spacing are printed, followed by the revised sentence:

	line 3
	OLD:  Jeffrey Quest,or Jeff Request, depending on his mood.
	NEW:  Jeffrey Quest, or Jeff Request, depending on his mood.

	line 6
	OLD:  "The lights finally dim, and the band appears on stage".
	NEW:  "The lights finally dim, and the band appears on stage."

The DOUBLE WORDS program (not in the Collegiate Edition) finds typographical errors. It is followed by a WORD CHOICE program (renamed DICTION and SUGGEST in the Collegiate Edition), which lists possibly wordy or misused phrases:

	beginning line 185
	Computer writing labs *[which]* use micros may charge 
		a fee for use.

p. 45

beginning line 15 On the [inside of] the building, there are many large colorful paintings of Indians covering the walls.

For each bracketed word or phrase WW suggests a substitution. For the bracketed which, WW provides the following advice: "Use `that' when clause is restrictive"; "use `of which' for `of that'"; or "use `when' for `at which time'." For the second example, WW suggests that inside should be used for inside of. The teacher must emphasize the fallibility of any text-analysis program--the WW analysis identifies problem words from a list of about 500 words and phrases without regard to context. Thus, the student must decide if a suggestion is appropriate. In the which example, only the first revision works; for the inside of example, WW offers inappropriate advice. In the original version, an explanatory program, WORDUSE, offers further instructions for the writer who wants to find out more about individual phrases.

Next, split infinitives are identified (in the Collegiate Edition, "split infinitives" and misuses of a and an are detected in the GRAMMAR portion). The STYLE section of WW offers a readability score; percentages of types of sentences (i.e., simple, complex) in its "variation" section; and then gives the percentage of passives (the program notes it can err on identifying passives), nominalizations, and expletives. Unlike GRAMMATIK, WW gives a detailed explanation with this analysis. Under "variation," it might advise how to rewrite for better sentence variety in both types and

p. 46

lengths. Or if the writer has used too many expletives, the writer will get an explanation of expletives; however, if the percentage of expletives is acceptable, then there will be no explanation. Likewise, if over 3.36% of nominalizations are used, then sentences with these nominalizations starred are printed. If the writer uses an acceptable number of nominalizations, the program commends him or her. An option of WW includes personalized style tables. The program, in its original version, compares student writing to technical writing, which is fine if the student is in a technical-writing class; it is not so fine for the first-year student writing a narrative. The instructor may opt to put in several good examples of narratives and set up a style table specifically for narratives. Other options include getting a shortened version of the prose analysis or matching style tables.

The "organization" program prints the first and last sentences of each paragraph so that the writer can check the flow of ideas. Another way to test the ease of reading the text is to use the "abstract" program, which analyzes the percentage of abstract words used and offers an explanation of the differences between concrete objects and abstract ideas (the list of 314 abstract words is modifiable). As good writers prefer concrete to abstract in most instances, they also prefer active verbs to passive verbs; consequently, the next program flags all "be" verbs and prints a formatted, double-spaced copy of the essay; in the new version, essays with under 25% "be" verbs do not have these verbs highlighted.

p. 47

An original program not nearly as useful--and wisely deleted from the Collegiate Edition--is the PARTS OF SPEECH one, which prints the entire essay with corresponding parts of speech.

The "sexist" language program--in DICTION in the Collegiate Edition--is helpful in making students aware of the problems of sexist language. Another program that was deleted in revision defines "syllables" of words so that students can print out a list of monosyllabic words, a list of two-syllable words, or n-syllable words; its value for students is questionable. Likewise, "topics," a listing of keywords in the text, is not in the newer WW. However, a statistical table that may be more helpful is the "style" chart, which sums up the essay by number of sentences, words, types of sentences, word usage, and sentence beginnings, more detailed than the similar GRAMMATIK summary.

The GRAMMATIK/COMMENT analysis print-out runs only a page or two; the WW print-out--on even a short essay--may be as long as 15 pages and may overwhelm the student. Again, students require instruction on interpreting the analysis and using it for revising. AT&T, in marketing the Collegiate Edition, offers an option to the mainframe--the 3B2 Computer, a stand-alone unit designed for the centralized writing lab approach to WW, allowing for terminals to be used in a network. In this scenario, students type in text and run "draft," which offers the following analyses:

development (points out short paragraphs)

p. 48

check (finds troublesome word pairs)

After revising the text, students print the "final" program: the final copy of the paper plus "organization," "style summary," "vagueness index," and "abstract index"--a packet ready for instructor-student conference and evaluation.

Still another version of WW is Release 3.0, "an outgrowth of the original version. . . , not an extension of the Collegiate version" (MacDonald, 20). Its features include an interactive proofreading and correcting program that presents errors in context, along with a menu of possible actions.

Again, the objective of a writing program should determine software and hardware. For USU, the mainframe is used not only for its editor and WW but also for electronic mail, a feature that allows for more communication among students and teachers, especially for mailing essays between students for peer critiques. For other colleges or high schools, the writing lab centered around the 3B2 might be more useful.

At the 1985 UCLA Conference on Computers and Writing, a speaker noted that WANDAH did not test well in his technical writing class. It would have been surprising if it had. Developed by Ruth Von Blum and Michael Cohen and marketed

p. 49

by Harcourt, Brace, Jovanovich, WANDAH's original audience was the basic writer in college. Consequently, WANDAH--an easy-to-use program--cannot be upscaled to advanced writing; instead, it performs well at the essay level, which makes it appropriate for first-year composition or high-school writing programs. In fact, the State of Utah has invested hundreds of thousands of dollars in WANDAH for school districts, following the success of the WANDAH pilot program at Logan High School, a school that built its Writing Room around WANDAH, integrating it in the English curriculum (Stoddart, 1985). For WANDAH meets the goals of the high-school curriculum by offering what the three previous software packages do not: an integrated system using prewriting aids, word processor, and reviewing and revising aids.

Within the invention program are instructions for "invisible writing" (the text will not appear on the screen until the student writes 100 words), "nutshelling," "planning," and "freewriting." "Nutshelling" asks the student to think about purpose, audience, and main ideas; "planning," designed for use with argumentative essays, helps writers with an issue's pros and cons. After the student generates and organizes ideas, then he or she types the essay, using the word processor for editing and printing. At any point, the student can use the split screen option to display an outline or even another paper on the other side of the screen.

The third section of WANDAH, "review and revise," identifies problems in mechanics, style, and organization. "Mechanics" include the spelling checker (a separate disk with a

p. 50

39,000 word dictionary), the "punctuation" checker (for problems with parentheses and brackets, quotation marks, ellipses, and question marks), and "word choice" checker (for those troublesome word pairs like accept/except). Possible problems are highlighted in the text, which appears on the bottom half of a split screen; the top half tells the student how to get an "explanation" for the error and how to "replace" the error.

"Style" highlights the following word types: "be" verbs; prepositions; abstract words; -tion or -sion words; and gender-specific nouns. Writers can choose to look for any one or all of these types at a time; the potential problems are flagged on the screen or print-out (see Figure 3). Like the other programs, it, too, offers statistics on sentences and paragraphs (plus "comment" for use in revision) in addition to the percentage of "be" verbs, prepositions, -tion words, abstract words, and gender-specific words. Unlike other programs, it visually diagrams the sentence lengths with bar graphs so that students can see how long or how short sentences are or check the variety of sentence length in addition to checking paragraph length (see Figure 4).

When revising for organization, students can choose "nutshelling," an overview summary "outliner," or a "transition words" analysis. The first choice, "nutshelling," repeats the prewriting technique and then compares its output to the original. The summary "outliner" picks out the first sentence of each paragraph although students have the option of selecting a sentence from each paragraph to compile the outline. In one class students who printed the

p. 51

Analyzing CROONERS for Version 7, pg. 5 "be" verbs b="be" verbs patiently for me to come. I casually look over my shoulder and take one last look at the red neon sign that has been turned off. As I slowly turn the bbbb corner I sadly wonder if I will be coming back the next week. The carnival bb is over. bb Stylistic Summary of CROONERS Version 7, pg. 6 "Be" Verb Statistics 6 of CROONERS's sentences contain(s) at least one "be" verb (18.8% of all sentences). Overall, the text has 7 "be" verbs. This text does not rely heavily on "be" verbs. Analyzing CROONERS for Version 7, pg. 5 prepositions p-prepositions patiently for me to come. I casually look over my shoulder and take one last PPP PP PPPP look at the red neon sign that has been turned off. As I slowly turn the pp pp corner I sadly wonder if I will be coming back the next week. The carnival is over. Preposition Statistics CROONERS contains 54 prepositions (including infinitive "to's") in 29 sentences. The text has 1 preposition for every 8.4 words, and an average 1.7 prepositions per sentence. Your text does not have too many prepositions, but you may still find ways to eliminate a few. See if you can.

Figure 3. Portions of "Be" and "Preposition" Analyses.
p. 52


Paragraph 1


Figure 4. Sentence-length Graph.

first sentence outline of their essays checked these sentences in editing groups, making sure that each one provided adequate transitions. To insure cohesion in an essay, the "transition" section highlights transition words and phrases and pronouns.

One nice feature of WANDAH is its "comment" program. In class, students trade disks and read each other's texts, inserting comments as they read, another form of peer editing. When printed or looked at on the screen, these essays include the peer review, which is underlined within the text. Likewise, the instructor can also comment on students' essays with this facility, but teachers using "comment" point out that students writing the comments often learn as much or more as the students receiving the comments.

In Utah school districts, saying anything negative about WANDAH is almost heresy; but even WANDAH has its faults. The main one is also its best-selling point; one of the rare software programs with not only word processing but also

p. 53

prewriting and rewriting tools, it also runs on its own operating system, making it impossible to "upload" files to a mainframe or to transfer them to another disk. In short, a student must use WANDAH to the exclusion of any other software. Most of the time, WANDAH's less powerful and flexible word processor is adequate; there are times, however, in research writing or journalism classes when writers may need more space or more formatting options. More sophisticated writers balk at the slowness of the word processor, but it must be remembered that advanced writers are not the audience for WANDAH.

Before teachers and administrators join the rush to invest in any software that offers text analysis, they should consider the value of computer style checkers, one of the debates in the world of computers and writing. Some teachers see style checkers as harmful while other use them along with instructor guidance, pointing out their limitations. Good scores on style statistics do not necessarily equal good writing although Reid and Findlay (1986) found correlations between WW analyses and essay quality.

For those who opt for text-analysis software, their primary concern should be matching software to program needs. Administrators look first at cost. No computer-writing program is inexpensive; however, hooking "dumb" terminals to an existing mainframe may be cheaper in the short run although demand may exceed mainframe capacity and create massive frustration for both students and teachers eventually. On the other hand, a mainframe offers electronic mail, an important pedagogical tool in writing classes. Micros, however, offer networking possibilities,

p. 54

and as noted earlier, WANDAH's "comment" encourages collaboration. In the best of all possible academic worlds, money would not be a concern, and writing programs would match software and hardware by purpose.

Programs that encourage a skills-based approach to writing may choose software that identifies superficial problems while a process-based approach writing program may prefer software that attempts to get at changes in macrostructure as well as microstructure. To be considered also is the level of student abilities. For example, WANDAH is being tested in middle schools; WW would not work for that audience but moves up the college ladder well. Though, simplicity may mean that a program is not readily modifiable to adjust for local needs such as inserting varying standards to measure essays against. Finally, are the computers used in a centralized or decentralized manner? And, do classes meet in a computer lab or is computer use limited to outside class?

The four schools described here successfully married software to program goals. Like any marriage though, the partnership is not without faults. Those faults can be minimized by cool evaluation of both curricular needs and software.

p. 55


The author would like to express her appreciation to Mark Scherting for his assistance in visiting the computer-writing sites and analyzing the programs. I would also like to thank Barbara West of Weber State College, Pam Gardner and Deanne Thomas of Utah Technical College, and Patricia Stoddart of Logan High School for their gracious hospitality during our site visits.


MacDonald, N (1986). News. WUG Newsletter, 1(2), 20.

Reid, S. & Findlay, D. (1986). WRITER'S WORKBENCH analysis of holistically scored essays. Computers and Composition, 3(2), 6-32.

Stoddart, P. (1985). Computers as tools for writing in Logan High School. Writing Lab Newsletter, 9, 5-6.

Joyce Kinkead teaches at Utah State University, Logan, Utah.