For almost ten years--first at a resource-poor university in Arkansas and now at the resource-rich University of Michigan--I have either had sole responsibility for a computer-equipped classroom/writing lab (Arkansas) or I have been a principal figure in making decisions about equipping such a facility for writing instruction (Michigan). Based on these experiences--and on the experience of working with primary and secondary school teachers through the National Writing Project and in several other college/K-12 collaborations--I can say that several factors make choosing software today just as difficult as in the days when selecting software for Computer-Assisted Instruction (CAI) simply meant choosing which drill-and-practice set to inflict on hapless students (or choosing, even better, not to inflict any of them on students).

Today, a wide range of products, some even designed by composition faculty, makes choosing software for instructional use much harder. Because quite a bit of software that supports progressive writing instruction exists, we still are left with the same number of potential products to evaluate and select as we had before, when all the CAI software was based solely on traditional, formalist notions of writing instruction. Further complicating the selection is the availability of a number of programs that, although not designed specifically for the writing classroom, in some way enable activities that writing teachers encourage: Word-processing programs enable revision, making process-based teaching easier to accomplish; spell checkers, grammar checkers, and thesauri promise to help students perfect the products of their writing processes; invention software (often called idea processors) claims to make students think more critically and therefore write more effectively; communications software allows students to write to each other or to share texts for purposes of collaboration or peer review.

In addition, as software has become capable of more and more functions, it has also become more and more expensive, and running it requires more and more of the computer's available Random Access Memory (RAM), meaning either that the leftover RAM is insufficient for students writing longer documents or that even more money is needed for upgrading computers. These issues are particularly acute in primary and secondary schools because money to spend on computers is much harder to come by there than it is at the college level. Purchases at the K-12 levels are much more permanent: Once bought, a computer or an item of software will be used for far longer than at any other level of education (hence the large number of school systems--even well-to-do ones--still using Apple IIs, Commodore 64s, etc.). In the face of all these difficulties, solving the software purchasing puzzle is paramount to the success of a computer-equipped classroom/lab, for without software that will support the teaching endeavors in those facilities, we might as well run them as fancy do-it-yourself typing pools.

In the spirit of not giving up in the face of a challenge, I offer the following criteria for selecting software for a computer-equipped classroom/writing lab.

Why do teachers want the computer's help? Every set of hints about buying computer hardware or software begins with some variation of this question, but for writing teachers, the question is especially important. With a few notable exceptions, software is designed primarily for commercial use, so everything from what it does to how easy it is to learn is based on assumptions that simply do not apply in the classroom. Thus, if we go into the selection process without a firm notion of what we want the software to do, we are not likely to end up with useful software.

In the beginning, those who developed and marketed CAI software merely tried to let the computer take over the drudge work (teaching grammar, for example) as if the computer were a book, usable in one area or on one topic at a time. Fortunately, such ideas have not prevailed, and today, we use the computer as a tool that supports a much wider pedagogical range. Thus, the computer's place in the classroom is both more central and more subordinate than it was a decade ago: We use computers to enable a range of teaching and learning activities, but we no longer think that the computer itself can teach anything.

In selecting software, then, we need to consider whether we want students to be able to collaborate, communicate, evaluate their own and each others' writing, develop an effective and efficient writing process, learn how to make their written products as correct and as clean as possible, or accomplish a host of other activities or tasks that some, most, or all writing teachers would want students to do. The first step in buying the right software is to make a fairly specific assessment of the needs and the practices of the teachers who are likely to teach in the classroom or lab, and then make sure that the software supports--or can be made to support--as many of those needs as possible.

Consider a program's versatility. The best purchase may not be the one piece of software that perfectly fits a specific need but rather the application that can be used for the widest range of purposes, as long as that one crucial need is met well. For example, a really powerful word-processing program can, with a little ingenuity and a few macrocommands, accomplish many of the most useful tasks of the so-called revision software, making the word-processing package a better value. Revision software often does not come with an adequate word-processing function. In addition, versatility in terms of input and output is even more important. The text files from invention applications must be readable by a word-processing program whose files must in turn be fed into a revision application, be uploaded into some sort of e-mail application, or be usable by any number of other types of programs. For this reason, looking at a more expensive but more fully integrated suite of programs may make more sense than selecting separate programs that perform the same functions but may not be as fully compatible. Value in this area means buying programs that can accomplish more than one task and fill more than one need, applications that can allow teachers' natural inventiveness and resourcefulness full play.

Focus on how quickly students can learn to use the program and how easy it is to use. For those who teach on the quarter system, for instance, ease of use is more important than for those on semesters. Most software, word-processing progams in particular, is designed for business use, where people can be professionally trained and then spend an extended time applying that training. In a business setting, then, a program that is powerful but difficult to learn and use may not cause the problems it will in a classroom, where students must begin writing essays immediately and have a maximum of 16 weeks of using an application.

Ease of use--transparency, in computer terminology--means that teachers can devote less time to teaching computer skills and more time to teaching writing, and it means that students spend less time being frustrated by the technology and more time exploring its capabilities and benefits. In general, if students cannot master the basics of an application upon first using it and become familiar with most of its functions within a week, then it is probably too difficult to use in a writing class. Exceptions to this rule should occur only when the program delivers especially valuable, unique capabilities.

Finally, consider whether you can write your own application. Before 1987, when Apple released HYPERCARD, this option was neither attractive nor realistic for most teachers. But today's hypertext-based authoring systems--IBM's TOOLBOOK provides much the same flexibility and ease of use as HYPERCARD does--allow users who know very little about computer programming to develop applications that fit a variety of needs and that, unlike commercial applications, are actually designed by people who know something about writing instruction. Developing powerful or highly sophisticated applications takes time and a good deal of expertise, but simpler programs with useful, specific functions are well within reach of the novice whose institution can supply moderate financial and technical support. Or, if even this prospect seems daunting, teachers may seek out programs that other teachers have developed. These applications are constantly being demonstrated or discussed in conference presentations, and the developers are usually more than willing to share the fruits of their labors for free or at very little cost.

William Condon is Associate Director for Instruction in the English Composition Board at the University of Michigan.