Journal of Business and Technical Communication...3. to provide feedback to writers or promote communication with collab-orators (e.g., by making annotations while reading that are

http://jbt.sagepub.com

Communication Journal of Business and Technical

2001; 15; 333 Journal of Business and Technical CommunicationJoanna L. Wolfe and Christine M. Neuwirth

From the Margins to the Center: The Future of Annotation

http://jbt.sagepub.com/cgi/content/abstract/15/3/333 The online version of this article can be found at:

Published by:

http://www.sagepublications.com

can be found at:Journal of Business and Technical Communication Additional services and information for

http://jbt.sagepub.com/cgi/alerts Email Alerts:

http://jbt.sagepub.com/subscriptions Subscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

© 2001 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. at PENNSYLVANIA STATE UNIV on February 7, 2008 http://jbt.sagepub.comDownloaded from

http://jbt.sagepub.com/cgi/alerts

http://jbt.sagepub.com/subscriptions

http://www.sagepub.com/journalsReprints.nav

http://www.sagepub.com/journalsPermissions.nav


JBTC / July 2001 Wolfe, Neuwirth / FUTURE OF ANNOTATION

This article describes the importance of annotation to reading and writing practices andreviews new technologies that complicate the ways annotation can be used to supportand enhance traditional reading, writing, and collaboration processes. Important direc-tions for future research are discussed, with emphasis on studying how professionalsread and annotate, how readers might use annotations that have been produced by oth-ers, and how the interface of an annotation program affects collaboration and communi-cation on revision. In each area, the authors emphasize issues and methods that will beproductive for enhancing theories of workplace and classroom communication as well asimplications for the optimal design of annotation technologies.

From the Margins to the CenterThe Future of Annotation

JOANNA L. WOLFEUniversity of Texas at AustinCHRISTINE M. NEUWIRTHCarnegie Mellon University

Information and information technologies are supported bycommunities of practice and institutions and, in turn, influencethose communities and institutions (Brown and Duguid). As a

case in point, annotations were central to knowledge sharing in medi-eval literary cultures but became more peripheral with the advent ofprint technologies. Medieval readers persistently used the interlinearspaces and margins of manuscripts to discuss, critique, and learnfrom the annotations left behind by earlier readers. In fact, annotationwas such a central part of medieval reading practices that annotationswere routinely transcribed along with the primary text, and often theannotations eclipsed the primary text as multiple readers added newlayers of commentary and responded to one another’s interpretationsand notes (see Figure 1). This marginal commentary not onlyincluded exegesis but also drawings, corrections, comments to orabout other readers, and discussions about the nature and art ofreading.

In contrast to the richness of medieval annotation practices, printannotation practices are relatively impoverished. Medieval literarycultures were able to support the exchange and discussion of annota-tions because multiple readers typically all had access to the same

333

Journal of Business and Technical Communication, Vol. 15 No. 3 July 2001 333-371© 2001 Sage Publications, Inc.



material copy of a text, which then served as a public resource forsharing information. In cultures with print technologies, however,most readers purchase individual copies of a text, so any annotationsthey make in the margins or on the text itself generally remain privateand cut off from other readers. Indeed, annotating copies of a text thathas been loaned is considered defacement. Thus, the only formalmechanism that print technologies provide for sharing annotations isthrough formal publication channels, such as the exegetical footnotesused in scholarly editions of texts. These scholarly annotations arephysically separated from the text to which they respond and aremuch more restricted in form and function than the handwrittenannotations readers typically produce while reading a text (Marshall,“Toward”). As a result, readers in cultures with print technologies, ascompared with medieval readers, have limited opportunities for dia-logue and learning through observation of others’ interactions with atext.

334 JBTC / July 2001

Figure 1. Bible with commentary from the Glossa Ordinaria, ca. 850-1499. M389.SOURCE: From the Philip Bliss collection of medieval manuscript fragments. Courtesyof Department of Special Collections, Stanford University Libraries.NOTE: In this medieval annotation system, we see several layers of commentary aslater generations of readers add and respond to annotations made by theirpredecessors.



Although print technologies contributed to the decline of annota-tions as critical forums, digital technologies promise to revive thesepractices and create new ones. In the past few years, a number of soft-ware and hardware applications have emerged that capture and dis-tribute the annotations made by readers of digital texts. (Figure 2, forexample, depicts an interface for carrying out digital conversations

Wolfe, Neuwirth / FUTURE OF ANNOTATION 335

Figure 2. Screenshot of the PREP Editor.SOURCE: Data for screen from Wojahn. Used with permission.NOTE: This aligned annotation interface supports marginal commentary through adesign reminiscent of medieval annotation systems.



about a text, similar to the marginal glosses in Figure 1.) These appli-cations include widely used commercial products, such as MicrosoftWord and Adobe Acrobat; collaborative writing products, such asHoughton-Mifflin’s CommonSpace; tools facilitating annotation anddialogue on the World Wide Web, such as CoNote (Huttenlocher), Crit-ical Tools, Microsoft Office 2000, and ThirdVoice;1 applications support-ing informal interactions, such as Voicefax (Frohlich and Daly-Jones);new research technologies for displaying annotations, such as FluidDocuments (Chang et al.); and handheld devices, such as XLibris(Schilit et al.), which use electronic pens to capture the materialityassociated with reading physical documents.2 As more and morereaders take advantage of new annotation tools, they will be leavingtangible traces of their mental activities as they interpret, analyze, andcritique texts. Many of these annotations will be housed in public (orsemipublic) databases where they can be accessed and shared by sub-sequent readers of a text.

What contributions, if any, will these new technologies for captur-ing, displaying, and distributing annotations make to our goals asreaders and writers and to workplace communication practices? Willdevices for capturing annotations facilitate reading processes, or willthey be more time-consuming and disruptive than their paper-basedequivalents? Will databases that store and distribute the annotationsof previous readers provide others with useful information, or willreaders exposed to the annotations of others find them distracting?New technologies for distributing and displaying annotations, alongwith familiar computer functions, such as searching, filtering, andcustomizing appearance, may increase the utility of the annotation asa communicative forum. Or they may not.

In this article, we review the literature on reading, writing, andannotating as well as technologies supporting these activities with aneye toward identifying needed research that can help shape commu-nities of practice and promote useful new technologies. Our review isorganized around the roles that annotation plays in contemporaryprint-based publication practices. Today, a sharp distinction existsbetween private annotations that readers make for their own pur-poses and public annotations that are intended to be shared withother readers. Currently, four main functions of annotation can beidentified:

1. to facilitate reading and later writing tasks (e.g., by making self-directed annotations while reading);




2. to eavesdrop on the insights of other readers (e.g., by examining anno-tations made by previous readers of a text);

3. to provide feedback to writers or promote communication with collab-orators (e.g., by making annotations while reading that are directed toother authors); and

4. to call attention to topics and important passages (e.g., by makingannotations while authoring that are directed to the readers).

For each function, we review studies of people’s practices, thedevelopment of relevant technologies, and potential applications forresearch on annotation from changes in the training of technical com-municators to recommendations for the design of new reading andwriting software. We believe that scholars should keep these applica-tions in mind as they design and frame their studies, so they canaddress their research beyond the professional writing community tocomputer design communities. The References section of this articleprovides a good overview of the conferences and publications wheresuch work would find a receptive audience. By addressing computerdesign communities, professional writing researchers will be able toform connections and alliances that can involve them in the design ofnew technologies. We have much to offer computer designers, both inimagining new applications and uses of technology and in testing theusability of new software and hardware.

FACILITATING READING ANDLATER WRITING TASKS

Most research on reading has been conducted in academic settingsand has focused on reading to learn. In this section, we call for moreresearch on reading in professional settings with a special emphasison the ways in which professionals annotate. This research has twomain applications: to help designers of computer software and hard-ware build better reading tools and to help instructors devise inter-ventions that may enable students to master the reading and writingstrategies in the discourse communities they hope to join.

What We Know about Annotating While Reading

People often annotate documents in the course of reading them. Ina diary study of how 15 people from a variety of professions read inthe course of their working lives, Annette Adler and her colleagues




found that annotation occurred in conjunction with reading morethan 25% of the time. Moreover, note taking on a separate documentoccurred in conjunction with reading an additional 22% of the time.Most workplace annotation is a solitary activity although workersoften refer to their annotations in later conversations with colleagues(Blakeslee; Whittaker, Frohlich, and Daly-Jones). And people oftenmake and review annotations while listening (e.g., in a meeting)although little is known about this activity (but see Whittaker,Hyland, and Wiley for related practices). Handwritten annotationscan take a variety of forms; the most common of these are highlights,underlines, marginal commentary, arrows, question marks, andasterisks.3 Often, these markings are cryptic and telegraphic and fol-low a coding scheme that is only apparent to the annotator (Marshallet al.).

However, this research on the ways in which professionals anno-tate does not sufficiently address the question of why they make theseannotations. What social and cognitive processes do these annota-tions support in workplace settings? Although empirical researchinvolving students suggests that annotations improve comprehen-sion, facilitate rereading and review of documents, and help writersbridge reading and writing practices, very little of this research hasbeen conducted in workplace settings.

Most of the experimental research has focused on students who arereading to comprehend and retain information (see Anderson andArmbruster; Caverly and Orlando). Readers who annotate text seg-ments increase the amount of time and effort spent on those segments.This increase results in improved recall and performance on test itemsrelated to those segments. Why those segments are better recalled isusually explained by the principle of “levels of processing” (Craikand Lockhart): The durability of memory is a function of the depth ofprocessing, where greater depth means a greater degree of semanticprocessing. When the depth of processing of the material is appropri-ate to the task, performance on the task improves.

Unfortunately, how often readers in professional settings makesuch annotations for comprehension and recall or whether workers insome environments would benefit from an increase in such practicesis unclear. Moreover, whether different annotation strategies havedifferent effects is also unclear. Studies examining the easiest form ofannotation—simply underlining—indicate that underlining is equiv-alent in effectiveness to other techniques for improving comprehen-sion and recall, such as rereading, answering periodic study ques-




tions (sometimes inserted by the author into the text itself), orsummarizing the text, but may be less effective than more time-consuming strategies, such as note taking or outlining (Anderson andArmbruster).

Annotations made while reading have also been found to facilitaterereading, providing readers with efficient retrieval cues when theyreview the subject matter (Kiewra) and helping them to locate specificinformation from the text. In the workplace, this type of review oftenoccurs just before meetings (Marshall et al.). The importance of thisreview function is emphasized by usability studies indicating thatmany people dislike reading online because the lack of annotationfunctions makes it difficult for readers to locate information that theyhave already read (Wright). Little research has been done on annota-tions made while reading to do (e.g., reading a reference manual tofind how to position a picture in a word-processing program)although a small study contrasting five textbooks and two program-ming manuals indicates that annotations on manuals were fewer innumber and of different types than those on textbooks (Nielsen).Nonetheless, many people do not like updates to reference manualsbecause they lose the annotations they have made on the previousmaterial, suggesting that these annotations are adding value in theskimming and rereading of the information (Walker).

Not infrequently, professionals read in order to write their owntexts. In academic settings, writing-from-sources tasks have beenfound to be cognitively complex activities that pose considerable dif-ficulties for novice writers (Geisler; Kaufer and Geisler; Penrose andGeisler). Theories of writing processes typically identify the follow-ing activities in writing from sources: acquiring knowledge, viewingknowledge from different perspectives to gain new insights, structur-ing knowledge according to those perspectives, selecting and possi-bly creating knowledge to meet goals for discourse, and rearrangingthese selections so that readers with different perspectives will findthem equally coherent. Annotation often plays a key role in facilitat-ing these activities. Joanna Wolfe (“Pedagogical”), for instance, foundthat people reading letters to the editor in preparation for writingtheir own responses to a particular issue annotated the source materi-als for a variety of purposes. Some of the reasons that they providedfor making specific annotations were to monitor comprehension andmark key points for later review, mark passages intended for citationin their own texts, and express critiques and evaluations that theyused to construct novel claims from the materials. Catherine Marshall




(“Toward”) similarly speculated that many of the annotations ontexts used in upper-level university classes functioned to mark quota-tions or to record interpretations that the annotator intended to use inan original paper.

The contexts for reading appear to shape the type and quantity ofannotations readers make. In a study of a reading group of profes-sional researchers, Marshall and her colleagues found that discussionleaders read more carefully and made more annotations than othermembers of the group. Wolfe (“Pedagogical”) likewise found thatinstructors reading letters to the editor in preparation for writing theirown response texts made very different types of annotations thaninstructors reading the letters in preparation for discussing them in acomposition class. When preparing their own texts, instructorsincluded more annotations that evaluated the source materials andfewer didactic annotations that labeled organizational features andrhetorical strategies.

Technologies for CreatingAnnotations While Reading

Despite the increasing distribution and many possible benefits ofreading documents online, most readers—in professions rangingfrom researchers to technical editors to students—prefer to printpaper versions of online documents before reading them (Dayton;Haas, Writing; Ovsiannikov, Arbib, and McNeil). Readers give manyreasons for this preference: Paper is more legible than computer dis-plays, paper is portable, paper allows readers to move back and fortheasily between multiple documents or between different parts of thesame document, paper documents can be easily annotated, and paperallows readers to get an overall impression of a document (seeDillon). However, new, extremely high-resolution screens are underdevelopment, and portable devices with pen-based inputs are now onthe market, though in early stages. These new appliances have thepotential to overcome many of the drawbacks of working with digitaltext.

Although reviewing annotated documents on a computer screenmay come with some costs, digital annotation technologies offer func-tions that are unavailable to readers making annotations on paper.XLibris (Schilit et al.), a pen tablet display device designed to mimicthe feel of paper documents (see Figure 3), allows readers to make




marks anywhere on a page with a variety of ink and highlighter pens.XLibris scans readers’ free-form annotations and uses key words asso-ciated with these markings to run implicit queries in a full-text data-base. The results of these queries are compiled in a further reading listthat is presented at the end of the document. In an experiment com-paring the effectiveness of queries derived from annotations withqueries based on readers’ relevance judgments, the queries derivedfrom annotations produced better results, suggesting that such pro-


Figure 3. A reader using XLibris to annotate a document.SOURCE: From the FX Palo Alto Laboratory, Inc. Used with permission. CopyrightFXPAL, 2001.NOTE: XLibris takes the ease and convenience of paper-based annotation practices andenhances them with digital search, storage, and display capabilities.



grams can enhance traditional reading practices (Golovchinsky,Price, and Schilit).

XLibris also provides advanced annotation support by collectingall of the annotations along with relevant clippings from the primarytext in a separate reader’s notebook. Readers can then use search fea-tures to locate specific annotations in the context of the primary text,or they can review the clippings in the reader’s notebook to remindthemselves of the text’s most salient or thought-provoking points.Professional researchers interacting with XLibris used the reader’snotebook in both expected and unexpected ways. One individualused the reader’s notebook to find a reference annotated in the text;another marked passages he intended to discuss with a colleague andthen used the reader’s notebook to coordinate a brief discussion; athird reader attempted to use the reader’s notebook to build his owntable of contents for the document and to label questions to ask in ameeting but ended up abandoning the scheme (Marshall et al.).Despite its user-friendly features, one subject complained about thedevice’s inability to fold a page so the text and a reference at the end ofthe document could be seen simultaneously. Another subject com-plained about the small size of the margins, a problem that could beeasily addressed.

Because writing from sources is such a cognitively complex task,tools for supporting this process are particularly welcome. One veryearly effort, Notes, was a hypertext application developed to investi-gate how writers acquire and structure new knowledge (Neuwirthet al., “Notes”). Notes was intended to augment writing processes byoffering an easier method (than marginal annotations, paper notecards, and other computer-based alternatives) for writers to (1) recordtheir own ideas, such as reactions, inferences, and plausibility assess-ments, while reading (allowing a writer to select a region of text,choose the menu item, and jot down an idea); (2) recover the contextfor those ideas (allowing a writer to go from any note to the relevantregion in a full-text online source); and (3) view ideas from multipleperspectives (allowing a writer to make multiple arrangements ofnotes and compare those arrangements). The Notes program differedfrom existing hypertext systems in that the user interface was opti-mized for taking notes from source texts. Although used and studiedextensively in the field and evaluated for usability, the program wasnot directly evaluated against more traditional paper-based strate-gies for taking notes.




Future Research Addressed to Design Communities

Reading devices such as XLibris have several implications forresearchers in professional communication. First, the designers ofthese devices need to know more about reading in the workplace,especially the physical ways that readers interact with texts. Studiesshould be conducted on how readers proceed through documents,when and why they interrupt their reading, the types and variationsof annotation and other forms of note taking that accompany reading,and the physical locations in which reading occurs. A range of meth-odologies could be used. Researchers could videotape readers as theyinteract with documents and observe the multiple activities thatoccur in tandem with reading. The documents read by the readers,along with any notes or other writing they produced while reading,could be collected and analyzed with the help of interviews or ques-tionnaires. Reading logs, such as those used by Adler and colleagues,would be useful for tracing reading through various workplacesettings.

Communication researchers might stress the potential designimplications of their work by imagining new combinations of discreteactivities that professionals make in the course of reading. Forinstance, if managers are found to move frequently from memos tocalendars, researchers might suggest an interface that brings up a cal-endar whenever a date is underlined in a reading appliance. In study-ing a professional reading group, Marshall and colleagues found thatreaders often moved back and forth to the list of works cited, suggest-ing the need for an annotation feature that would bring up the com-plete reference whenever the reader noted an in-text citation. Detailedobservations of reading activities taking place in naturalistic settingscan thus provide data that can be used to discover new possibilitiesfor reading appliances.

Comparison studies of readers making and using self-directedannotations in paper and electronic environments are also needed.These comparisons would have important implications for the designof computer interfaces. Some initial research in this area suggests thatmaking annotations in computer versus paper environments has lit-tle effect on undergraduates’ recall of informational texts (Rice; VanOostendorp). This research, however, has limited relevance for read-ing in professional settings because professionals and students oftenread for substantially different purposes. Designers also need com-




parisons of readers reviewing their own annotations in paper andcomputer environments. Although many annotation technologiesclaim to support annotation review, few studies have systematicallytested the effectiveness of such support.

Finally, comparison studies of experts and novices using these newtechnologies are also needed. Often, new technologies are only testedin expert settings where insufficient attention is paid to the costs andbenefits these new technologies might pose to beginners in particularknowledge domains. For instance, although XLibris’s advancedsearching and querying features offer exciting possibilities for profes-sional researchers, these features might prove to be overwhelming toreaders new to a particular knowledge domain. Domain experts andnovices might require different presentations or different weightingcriteria for searches and references from features such as the reader’snotebook or further reading lists.

Differences between Expert andNovice Annotation Strategies

Studies based on read-aloud protocols show differences in thereading strategies of experienced and novice readers. For instance,Graham Smart found that bank executives performing read-aloudprotocols made use of a variety of critical strategies. The bank execu-tives began by previewing a paper and then proceeded in an active,question-driven manner, taking a distinctly critical stance and contin-uously challenging the line of argument. Davida Charney, in a studycomparing the reading strategies of graduate students and estab-lished scientists, found that the scientists read in a less linear fashionand were more likely than were the graduate students to activelyemploy strategies that would help them resist the text and its claims.Christina Haas and Linda Flower, in turn, found that graduate stu-dents employed more rhetorical reading strategies than undergradu-ate students. Although more recent research implies that at least someof this difference is due to their lack of familiarity with the topic,undergraduates still seem to use fewer evaluation strategies than domore experienced readers (Haswell et al.). The suggestion that rhetor-ical reading strategies grow with increased experience is supportedby Haas’s (“Learning”) longitudinal study of one biology studentwhose conception of the function of texts and the role of authors grewin complexity as she began to have laboratory experience and to workin a more realistic workplace setting.




Experienced and inexperienced readers also seem to differ in theways they make and use annotations when reading to write. In astudy of six college students writing from sources, Mary Kennedyfound that students characterized as “fluent” readers made morenotes and annotations than did students characterized as “not-so-fluent” readers (447). Moreover, the fluent readers more frequentlyused their notes for more varied reasons when preparing to write anobjective essay based on the source materials. Although NancySpivey, in an earlier study, had found no differences in the annota-tions of expert and novice readers, this discrepancy is probably due tothe research task. While Spivey had students write a synthesis ofsource materials, Kennedy instructed her subjects to write an objec-tive essay. The need to interpret and evaluate material they had readprobably caused Kennedy’s fluent readers to make greater use ofannotations as they wrote. Although few researchers have specifi-cally compared differences between the reading and annotation strat-egies of experienced professionals in workplace settings and those ofundergraduates or beginning professionals, we would expect to finddifferences similar to those Kennedy found.

Future Research Addressed to WritingInstructors and Professionals

Given the number of cognitive processes that annotation can sup-port, explicit instruction in annotation strategies may benefit studentsin writing courses. Such instruction may help students comprehendand recall information as well as integrate and evaluate informationfrom multiple sources. Although no studies that we know of haveexplored how instruction in annotation strategies could help studentsin business and technical communication classes, researchers havefound such instruction valuable in other areas. Lorraine Higgins, forinstance, describes how intervention in composition students’ note-taking and annotation strategies improved their abilities to constructarguments from source texts. Michele Simpson and Sherrie Nist useda seven-part annotation strategy to teach students in a learning-skillsclass how to study effectively for exams. This strategy consisted ofteaching students to annotate by writing brief summaries in the textmargins using their own words, enumerating multiple ideas, notingexamples of concepts, marking key information on graphs and charts,jotting down possible test questions, noting confusing ideas, andselectively underlining key words or phrases. Students using this




multipart annotation strategy to study informative texts not only per-formed better on comprehension tests than students who used otherstrategies, such as previewing to generate study questions, rereading,or outlining, but they also spent substantially less time studying.

Interventions to improve annotation strategies for technical andbusiness writing students might be based on studies of the readingpractices of professionals in various fields. Researchers could easilygather data on annotation by collecting copies of texts that variousindividuals in an organization have read closely. To help interpretsome of the more idiosyncratic markings, subjects might be providedwith stimulated-recall questionnaires asking them to identify theirreasons for making each annotation (Doheny-Farina and Odell). Sucha questionnaire was used by Wolfe (“Pedagogical”) and proved to bean effective device for interpreting annotations. In addition, compari-sons of annotation practices might be done across professions as wellas between novice and expert readers in a single field. Studies of theeffectiveness of particular classroom interventions are also needed toexplore which ones would help students become better readers andwriters in the discourse communities they hope to join.

Researchers studying annotation in workplace environments mayfind that professionals in some situations do little annotation. In thesecases, researchers might want to step in to see if explicitly trainingprofessionals in annotation strategies might improve communicationwithin a particular workplace. Rachel Spilka advocates such inter-vention in workplace practices, stating that workplace practitionersare often unaware of ineffective communication strategies until theyreceive researcher feedback. Thus, communication researchers maybe able to contribute to as well as learn from the communities theystudy.

EAVESDROPPING ON INSIGHTSOF OTHER READERS

Technologies that easily allow readers to share private handwrit-ten annotations made by others do not currently exist, but they even-tually will become viable resources for professional writers and writ-ing instructors. Annotations made in tools such as XLibris arecurrently private, but these tools could be modified to allow sharing(with the annotators’ permission). Ilia Ovsiannikov, Michael Arbib,and Thomas McNeil provide a good overview of current technologies




that allow readers to share typed annotations and describe technicalfeatures desirable in databases for storing personal annotations ondocuments. Business and technical communicators should begin toanticipate the changes to reading and writing practices that will resultfrom technologies that blur distinctions between private and publicannotations. Research on sharing annotations is still exploratory, sowe advance a few hypotheses about the potential costs and benefits ofeavesdropping on personal annotations and suggest studies thatmight test these hypotheses.

What We Know about Sharing Annotations

Most of the extant studies examining how annotations made byone reader can benefit later readers of the same text have tended tofocus narrowly on how annotations affect study time and recall ofmaterial or performance on multiple-choice tests. These studies gen-erally conclude that reading texts that have been highlighted byexpert readers increases study time and improves recall on theemphasized items (Cashen and Leicht; Crouse and Idstein;Schumacher and Nash) but is not as effective as having readers maketheir own highlights (Fowler and Barker; Rickards and August;Schnell and Rocchio). Readers also appear to benefit more fromprehighlighted material if they believe that the annotator was anexpert (Fowler and Barker; Marshall, “Toward”).

However, anecdotal evidence suggests several reasons why anindividual might find others’ annotations valuable. In his influentialarticle presenting the earliest description of a hypertext-like docu-ment system, Vannevar Bush describes how readers’ annotationsmight be used to form trails in a hypertext system that later readerscould follow as paths connecting documents to one another. Like-wise, Van Dam, another early hypertext innovator, justifies hisemphasis on supporting the annotation of electronic documents bysaying,

The reason I encouraged such annotations was that I remembered thatwhen I was in college . . . I would always grab the dirtiest copy of a bookfrom the library, rather than the cleanest one, because the dirtiest oneshad the most marginalia, which I found helpful. (891-92)

Van Dam presumably used these marginal notes as aids in interpret-ing texts. Marshall (“Annotation”; “Toward”) similarly found that




some students liked to buy used literature and philosophy books withextensive marginal commentary. These students found such interpre-tive markings useful—so much so, in fact, that one reader had evenhighlighted the annotations left behind by a previous owner of thetext. David Nichols and colleagues report that users of a digitallibrary of Renaissance manuscripts reacted positively to tools thatallowed readers to share annotations. A parallel practice to annota-tion eavesdropping can be found in the compiled publication of anno-tations by a famous person (e.g., several volumes of marginalia bySamuel Coleridge and Edgar Allen Poe have been published).

Eavesdropping might also come into play when writers examineannotations made by their readers. For example, architects often scancopies of plans that their colleagues have marked so that they canquickly locate changes to the original design. Doctors similarly lookat annotations fellow practitioners have made on medical charts tofind noteworthy sections of a patient’s history (Luff, Heath, andGreatbatch).

Eavesdropping on others’ annotations thus seems to help subse-quent readers locate and comprehend information, find new refer-ences worth pursuing, interpret texts, and locate changes to docu-ments. Readers in the workplace might also eavesdrop on annotationsto glean information about how different readers interpret a text, locatepotentially troublesome passages in need of elaboration or clarifica-tion, or home in on controversial sections. However, shared annota-tions might also have unanticipated drawbacks.

Shared Annotations Might ProvideInformation about Audience

One exciting possibility is that eavesdropping on annotationsmight provide a source of audience information. Researchers havelong recognized the importance of audience in business and technicalcommunication. Documents in these settings frequently have multi-ple audiences (Dautermann; Doheny-Farina), and often writers havelittle firsthand knowledge about their primary readers (Huettman).Beginning writers in these professions are often unaware of the cru-cial role that audience plays in shaping their writing (Paradis, Dobrin,and Miller). However, even experienced writers may have difficultypredicting the emotional and psychological reactions of readers(Suchan and Dulek).




Annotations might help writers build more complex and accuratemodels of their readers because they provide insight into readers’processes and values. Technical communicators have used think-aloud protocols as a source of audience data to improve their docu-ments. But this technique is labor intensive and, in some contexts,may distort the data.4 Karen Schriver has shown that reviewing areader’s think-aloud protocol can improve a writer’s ability to antici-pate readers’ needs in general. Schriver used transcripts of readersthinking aloud as they tried to comprehend problematic texts to teachrevision to juniors and seniors from 10 professional writing classes.Her students became more adept at identifying text areas that wouldcause problems for readers than did students who were instructedusing a variety of audience analysis heuristics and peer review meth-ods. David Roberts and Patricia Sullivan similarly report success withhaving students complete think-aloud protocols of classmates’writing.

If reviewing readers’ think-aloud protocols helps writers to antici-pate readers’ needs, then writers, both professionals and students,should benefit as well from eavesdropping on annotations made dur-ing reading for similar purposes. Although annotations are not as richin information as think-aloud protocols, they can reflect readers’problems, questions, and evaluations of a text. To test whether eaves-dropping on expert annotations might influence the reading andwriting strategies of beginning writers, Wolfe (“Effects”) asked stu-dents enrolled in lower-level composition courses to write essaysbased on letters to the editor. Some of the letters contained no annota-tions, others contained underlining, and others contained evaluativecommentary in the form of positive and negative assessments inter-spersed throughout the primary text. Students receiving annotatedmaterials were told that a composition instructor planning to use theletters as discussion material had made the annotations. Wolfe foundthat students who received evaluative annotations were less likelymerely to summarize (rather than argue for or against) the texts thanwere students who received the same material without annotations.Furthermore, Wolfe also found that the perceived stance of the anno-tator appeared to influence students’ writing. Students who per-ceived the annotator as adopting the same position they pursued intheir essays may have put less effort into persuading their readersthan did those who perceived the annotator as adopting an unde-cided position. Thus, when students seemed to believe that the anno-




tator was already persuaded to their point of view, they tended towrite essays containing large amounts of irrelevant or inappropriatematerial (such as comments about the sex lives of the letter writers)and relatively little argument. Possibly, these students weredemotivated by annotations that seemed to agree with their point ofview.

Although Wolfe found mixed benefits to the annotations she used,her research does suggest that some annotations might give begin-ning professional communicators information about their readers.Researchers might begin by exploring whether reading memos,reports, or manuals that have been annotated helps beginning writersto anticipate the needs and expectations of their readers. JamesParadis, David Dobrin, and Richard Miller describe the case of oneyoung engineer who, to the dismay of his supervisor, documentedone long project in a report of nearly 75 pages. Clearly, the beginningengineer had no idea what kind of effort it took to read such a long,ragged, and confusing document. Such extreme neglect of readers’needs might be forestalled if writers routinely worked from sourcedocuments containing appropriate annotations. For instance, view-ing the evaluations, questions, and comments of a reader expressingirritation with a document might help young writers empathize withthe needs and expectations of their readers.

Shared annotations might also help experienced writers addressthe needs of their audiences. Oftentimes, writers in business settingshave little contact with the people to whom they write (Suchan andDulek). When preparing to write on a sensitive issue, experiencedwriters might therefore find it useful to eavesdrop on annotationsmade by particular readers to learn about the biases and values thatshape their responses. This practice might similarly benefit writers ofcomputer documentation. Barbara Mirel describes how writers ofsoftware manuals tend to present the needs of their readers narrowly,focusing on the tasks a user needs to perform rather than the largergoals a user might have for undertaking that task. Examining thenotes that computer users make in their manuals might provide docu-mentation specialists with needed information about the goals andactivities of their readers. In some cases, annotations from readers whohave elaborated about a procedure or commented on an example—information that is now often found separately in answers to fre-quently asked questions (FAQs), with no connection to the originaldocumentation—might be directly published in the document.




A number of methodologies could be employed here. Wolfe’sstudy comparing the performance of student writers receiving cleandocuments with that of student writers in experimental groupsreceiving annotated documents could be repeated using participantsand materials from a professional communication context (“Effects”).Researchers should also investigate whether some kinds of annota-tions are more effective than others are. Wolfe, for instance, found thatunderlining had no impact on students’ writing, but strongly positiveor negative commentary did influence students. Think-aloud proto-cols might provide useful information about how writers use annota-tions to construct their audiences. In addition, studies might addresslong-term effects: Do annotations only influence writing on the task athand, or can exposure to annotated materials over time graduallylead writers to internalize their readers’ needs?

Shared Annotations Might Help CommunitiesLocate Points of Consensus or Controversy

Marshall suggests that annotations might be used to locate pointsof consensus—sections of a text that multiple readers seem to havemarked as important or noteworthy (“Toward”). To explore whethera consensus might occur in readers’ annotations, Marshall examinedsix used copies of a theoretical text used in an upper-division com-puter science course. Although the overall level of agreement was nothigh, readers’ consensus was consistently higher than was predictedin consensus values based on chance. Hypertext developers mightuse this kind of information about community consensus to build tai-lored summaries of documents for different readerships. These sum-maries would differ from author-produced abstracts in that theywould reflect only the information that particular discourse commu-nities agreed was valuable. For instance, articles published in techni-cal journals often contain information about writing practices evenwhen that is not the article’s main focus. Readers in technical commu-nication would probably only be interested in the sections about writ-ing; thus, a computer program that scanned markings made by vari-ous communication researchers and pulled out text sections thatcontained consensus might help subsequent researchers quickly andeasily locate information in texts they might otherwise overlook.

Marshall’s exploratory study was very limited in the number ofsubjects included and the types of materials employed (“Toward”).




Therefore, replicating this study in different contexts to see to whatextent members of different communities agree on information that isworth annotating would be useful. Careful analysis of this agreementmight allow professional communicators to suggest patterns of anno-tation that computer designers might use. Moreover, researchersmight want to look not just for text segments that annotators agreedwere important but also for those segments that seemed to elicitcontroversy.

Shared Annotations Might HaveUnanticipated Drawbacks

Reading annotations made by other readers almost certainlyincurs cognitive costs as well as benefits. Although some readersactively seek out annotated copies of texts, others just as activelydespise them. Annotations can serve communal purposes, but theycan also clutter a text, interfere with comprehension, and invoke con-tempt for the intellectual abilities of previous readers. In an examina-tion of reviews commissioned by the publisher of The Presence of Oth-ers (Lunsford and Ruszkiewicz), a composition reader containing sixessays annotated by both of the editors and a student volunteer, Wolfe(“Pedagogical”) found many reviewers expressing reservationsabout the annotated essays. Although most of the reviewers liked theannotations, others complained that students found the annotationsdistracting or intimidating, and some reviewers feared that the anno-tations would stymie independent thought because students mightview them as correct responses to the readings. In an exploratorystudy of students’ written responses to the academic essays in ThePresence of Others, Wolfe found no positive effects for the annotatedessays and even uncovered evidence that might suggest that theannotations interfered with student comprehension.

Although Wolfe attributes these unpromising results to the didac-tic nature of the published annotations, which often self-consciouslycalled attention to the annotators’ own acts of reading, this studynonetheless underscores the potential drawbacks of sharing annota-tions. Nichols and colleagues similarly found that although users oftheir digital library liked its annotation capabilities, they were gener-ally concerned that the commentary might quickly become unman-ageable. Thus, research is needed that examines the situations inwhich annotations might hinder reading and writing processes.Again, studies of readers thinking aloud as they try to understand




various annotated texts would provide important information aboutthe costs and benefits of these materials.

PROVIDE FEEDBACK TO WRITERS OR PROMOTECOMMUNICATION WITH COLLABORATORS

Annotations directed to others most commonly serve as a meansfor readers to provide feedback on a work in progress but can alsosupport other communicative functions, such as to communicatethoughts about a finished document to later readers. A large numberof technologies have been developed specifically with this form ofannotation in mind. These technologies include annotation featuresin Microsoft Office 2000, Adobe Acrobat, Lotus Notes, CommonSpace,PREP Editor, MATE, and others.

In this section, we discuss situations in which annotations aredirected to others, how annotation interfaces affect collaborativeauthoring and editing, how voice annotations compare with writtenones, and how conversations surrounding documents unfold inannotation media. We focus on the subtleties of tool design because,despite the large number of applications designed to support other-directed annotation, few of these systems have been empiricallytested. Because these tools have great potential to influence the dailylives of professional communicators, research leading to specific rec-ommendations for the use and development of annotation software isbadly needed.

Situations in Which Annotations MadeWhile Reading Are Directed to Others

The most common uses of other-directed annotation occur duringthe document development process. These situations includecoauthoring, document review, professional editing, and instruction.Unlike annotations made for oneself while reading, other-directedannotations are likely to be incorporated into or otherwise influencethe final copy of the text. The audience for these annotations is gener-ally small and relatively well defined. However, that is not always thecase because annotations made during the process of documentreview can simultaneously be addressed to an author and to unspeci-fied potential readers, such as the author’s supervisor. For instance,Paradis, Dobrin, and Miller, in their study of research and develop-




ment workers at Exxon, found that marginal comments made duringdocument review allowed subsequent reviewers to catch up withwhat was going on in other departments. Collections of marginalcomments served as a repository for cross-fertilization of ideasthroughout the organization.

Other-directed annotations can also be used to advertise informa-tion about a communally shared document. In the London Under-ground, train controllers cover their timetables with cellophanesheets and mark temporary changes with a felt pen to share schedulechanges and other important information with others. Doctors markmedical charts to alert future practitioners to particular sections of apatient’s medical history (Luff, Heath, and Greatbatch). Governmentemployees often annotate documents with information about deci-sions or outcomes. For many of these government documents, thesemarginal notations become an intrinsic part of the official decision oraction (Houser and Hart). In these situations, annotations becomepart of the shared knowledge base of that group.

Most research on other-directed annotation has been in the area ofclosely coupled interaction between coauthors or between authorsand editors. Less has been done in situations where many readersprovide feedback, such as in large-scale requests for comments(RFCs). J. J. Cadiz, Anoop Gupta, and Jonathan Grudin have begunwork in this area, providing a descriptive study of a large develop-ment team—roughly 450 people producing about 9,000 annotationson about 1,250 documents over 10 months—using a Web-based anno-tation system. As in other RFCs, comments were not so much aboutthe documents as about the software specifications described in thedocuments.

How Annotation Interfaces AffectQuantity and Quality of Commenting

Many annotation technologies have been developed to facilitateexchanges between collaborators who are not colocated (see ACMSIGCHI and ACM SIGGROUP for current information). This effort tosupport long-distance collaboration reflects the fact that organiza-tions increasingly rely on distributed teams to provide faster, moreflexible, and, ultimately, more effective responses to a dynamic, tur-bulent global environment (DeSanctis and Jackson). Editors and writ-ers in many organizations, however, have begun to use annotationtechnologies even in situations where collaborators are in the same




place. Although most technical editors working on-site with authorsstill prefer to handwrite comments, they are under increasing pres-sure from their management to use online editing programs becauseof the benefits these programs offer to the organization (Dayton;Farkas and Poltrock). Some of these organizational benefits includegreater speed in preparing documents, better version control, betterarchiving, increased productivity, and improved systems integration(Farkas and Poltrock).

Handwritten comments, however, may have some benefits that aresubtle but nevertheless important. Because handwritten commentsare difficult to read, they are often interpreted in face-to-face meetings(Blakeslee). Such face-to-face communication can support moreinteractivity (e.g., clarifying questions, quick follow-ups) andexpressivity (e.g., eye contact, tone of voice), properties known to beimportant when communicating about highly equivocal problemsthat are sometimes at issue in a document (Kraut et al.). Because expe-rienced technical editors rate the ability to maintain positive relation-ships with the authors extremely high on their list of crucial skills(Duffy), to find that editors resist technologies reducing face-to-facemeetings would not be surprising. More research on the differentcommunities of practice and their relationship to technology isneeded in this area.

Different applications support different models of editing andreview processes. The most common models include the commentmodel, the edit-trace model, and the traditional markup model. Thesemodels, however, are not mutually exclusive, and some tools (e.g.,CommonSpace, PREP Editor) support both edit trace and comments onthe edits (Neuwirth et al., “Flexible”). Of these models, the edit-tracemodel is thought by some to encourage heavier editing and lessregard for the author’s original text than the other models (Dayton;Farkas and Poltrock) although no studies have confirmed these spec-ulations. Experienced technical editors also seem to have a preferencefor mixed models, either markup and comment or edit trace and com-ment (Duffy). Given that making a change is sometimes easier thancommunicating the problem (e.g., when the tone of a sentence is prob-lematic, changing it may be easier than describing the problem), toolsthat support mixed models are likely to be more successful. Edit-tracemodels have been criticized for being less economical and harder tointerpret (Farkas and Poltrock), but some research software hasattempted to address such issues (Neuwirth et al., “Flexible”), andsoftware is likely to improve. The editing model, however, that a pro-




gram employs may have substantial impact on how editors perceivetheir tasks and how authors, in turn, perceive editors. Research in thisarea therefore could greatly aid development.

Even within particular models of the editing and review process,the visual layout of the interface has been found to affect reviewers’abilities to detect problems. In a study comparing three different com-menting interfaces, Patricia Wojahn, Christine Neuwirth, andBarbara Bullock found that MBA students asked to review drafts oftwo different business cases responded to more problems in a textwhen using an interface that either inserted annotations directly intothe text (an interlinear interface) or aligned annotations horizontallyin the margin (an aligned interface, see Figure 2) than they did whenusing a footnote interface, such as Microsoft Word. Similarly, VanOostendorp found that a footnote interface reduced the compactnessof readers’ annotations. These results strongly suggest that a footnoteinterface substantially increases the difficulty of recognizing prob-lems and producing feedback.

Annotation interfaces also appear to affect the amount of commu-nication between reviewers and other collaborators. Wojahn,Neuwirth, and Bullock found that participants using a footnote inter-face were less likely to respond to annotations made by a previousreviewer than were those using interlinear or aligned interfaces. Verylimited, but nonetheless intriguing, evidence that an aligned-annota-tion interface might increase communication between collaborators isoffered by Neuwirth and Wojahn’s observational study of students inan argumentation course. The researchers found that an annotationprogram using an aligned interface that supported comments oncomments seemed to have encouraged students to ask questionsabout the teacher’s comments on their writing rather than simply dis-missing comments they did not understand. Moreover, instructorsusing the program frequently responded to comments made by peerreviewers, a practice that is relatively rare in traditional classroomcommenting situations.

Pen-based annotation interfaces using digital ink appear to offergreat potential for online editing. Editors frequently resist online edit-ing because keyboard input is slower than handwritten marks, andthe two-dimensional space of the computer screen interferes withtheir abilities to navigate within documents (Dayton). Pen-basedinput may alleviate these obstacles to online editing. MATE(Hardock, Kurtenbach, and Buxton) and PenEdit (see Farkas and




Poltrock) are two prototypes of digital-ink programs. Pen-based sys-tems have yet to be empirically evaluated.

Studies on annotation interfaces suggest two main directions forfuture research. First, many more studies are needed comparing dif-ferent types of interfaces. In addition to the layouts mentioned, theincreasingly popular sticky-note interface, in which reviewers inserttextbox icons on top of the text much like a reader might use a paperPost-it note to mark points on a paper document, has yet to be system-atically evaluated (theory, however, indicates it will be inferior to sys-tems that use interlinear or aligned interfaces because sticky notesoften obscure the underlying text). Second, most of the research onannotation interfaces has been conducted from the viewpoint of theperson making the annotations. However, these designs are verylikely to affect how authors view editors and other collaborators.More research is therefore needed from the point of view of the personreceiving the comments.

We imagine future studies on annotation interfaces triangulatingmethods such as discourse analysis (e.g., length of annotations,amount of mitigation) and task performance (e.g., number and typeof problems identified, time spent on task) with surveys and inter-views querying both authors and reviewers for their preferences andperceptions regarding feedback produced in different environments.Such research should be conducted in both workplace and classroomcontexts because the relationship between student and instructor isnecessarily different from that of their counterparts in businesssettings.

Spoken versus Written Annotations

Up to this point, we have discussed annotations that are written.New media, however, allow voice annotations as well written ones.Because writing is more effortful to produce than speech, causingwritten comments to be somewhat terse, researchers have begun toexplore the possibility of using voice as a medium to communicateannotations. The research below focuses on identifying situations inwhich voice annotations might be preferable to written ones and viceversa.

Voice communication (face-to-face meetings, telephone conversa-tion, etc.), of course, is widely used in collaborative writing situationstoday. Technical writers tend to be very savvy about when they




choose to use spoken feedback and when they prefer to use writtencommentary to discuss concerns and problems about texts. SamDragga found that technical editors preferred to express directives,questions, and suggestions in written formats, whereas complimentstended to be expressed through oral commentary. Spoken compli-ments were viewed by these writers as genuine and supportive; how-ever, when compliments were written, they were perceived as conde-scending or patronizing. Dragga also found that spoken feedbackseemed to draw attention to the collaborative nature of the writingproject and to minimize the defensive and possessive behaviors writ-ers often display regarding their writing. In a related comparison offace-to-face and written collaboration, Lynn McGee found that tech-nical writers’ preferences for different media changed throughout thedocument process. At the beginning of the document process, writerspreferred to use cross-functional team meetings and other forms offace-to-face communication but preferred textual forms of communi-cation, such as e-mail and document review, toward the end of theprocess. We might therefore hypothesize that in situations when col-laborators do not have the luxury of face-to-face communication,voice annotations might be preferred in the early stages of the docu-ment process with increasing reliance on written annotations towardthe end of the process.

In studies comparing electronic spoken and written annotations,the commenting media influenced the type of feedback produced.Robert Kraut and colleagues found that MBA students using voicemedia produced more comments on high-level concerns and used amore personal style of commenting than did students using a key-board. In a follow-up study, Christine Neuwirth and colleagues(“Distributed”) looked at annotations produced by reviewers ofmanuscripts written by members of a university computer sciencedepartment. Although comments about high-level substantiveissues, such as quality of the argument, were the most common for allreviewers, reviewers using voice mode produced more annotationsabout purpose and audience, whereas reviewers using keyboardmode tended to produce more comments about substance. In addi-tion, comments from reviewers using the voice mode were morepolite and provided more reasons for their concerns.

Neuwirth and colleagues (“Distributed”) also looked at how thesecomments were perceived by the authors of the documents underreview. The findings suggest that reviewers who produce annotationsin voice mode may be perceived as more trustworthy and somewhat




more likeable than reviewers who produce their annotations via key-board. This last result is consistent with studies in writing classroomsshowing positive responses to instructor-produced voice feedback(Pearce and Ackley; Van Horn-Christopher) and has strong implica-tions for interpersonal relationships between authors and reviewersor editors in business settings. Except for low-level comments, writersfound the voice annotations just as usable as the written annotations.

Unfortunately, voice annotation has its trade-offs. A speciallydesigned computer interface (as in Neuwirth et al., “Distributed”) isneeded to overcome production problems common with audiotapehardware (Pearce and Ackley) and problems with reviewing longaudiotaped segments (Byrne). Neuwirth and colleagues (“Distrib-uted”) found that although writers liked the use of voice commentaryon high-level problems, they did not like voice commentary for low-level stylistic feedback. However, reviewers and authors may differ intheir preferences. For instance, in a study comparing e-mail and voicemail usage, Maha El-Shinnawy and M. Lynn Markus found thatworkers in an aerospace and defense systems organization often pre-ferred one medium for producing messages but a different mediumfor receiving messages.5 These findings indicate the need to evaluatecarefully the benefits and drawbacks of using spoken and writtenmedia to communicate different types of feedback.

Annotation as Conversation

Annotation programs not only allow collaborators to comment ondocuments, but they also allow extended conversations to take placein the context of a common text. To aid distributed teams, PREP Editor(and its commercial version, CommonSpace) was designed to supportconversations over distances by allowing writers to annotate not onlythe original document in the margin but also the comments them-selves (Neuwirth et al., “Issues”). Figure 2 depicts an example of thePREP Editor in which reviewers respond to one another’s annota-tions. Each user’s annotation can form a conversational turn, with theconversation unfolding over time as the document is passed back andforth (Francik et al.). Web technologies such as ThirdVoice andStickyChats (Churchill et al.) have been developed with this conversa-tional use of annotations specifically in mind. Readers select text in adocument, and comments associated with it appear in the docu-ment’s margin, where they can be in turn annotated by subsequentreaders. Thus, unlike e-mail or Usenet discussions, conversations car-




ried out via technologies such as ThirdVoice or StickyChats do not havethe burden of quoting large blocks of text to remind readers of thelarger context of the message.

Very little research has been done on distributed conversationsusing such technologies. However, Neuwirth and Wojahn found thatthe conversational space provided by an annotation program encour-aged students to ask questions about the instructor’s commentary ontheir writing and also seems to have encouraged the instructor torespond to comments made by peer reviewers. This in-context con-versation thus appears to make the document review process moredialogic. Cadiz, Gupta, and Grudin have also begun some work in theuse of annotations to coordinate conversation in large groups sharingan interest in common documents. People liked using the annotationsto resolve minor issues outside of large group meetings, thus preserv-ing face-to-face meeting time for more important topics. However,people preferred e-mail over annotations when a quick response wasrequired. The researchers also found that people did not repeat a com-ment if someone else had already made the same point. On one hand,that saves time for document reviewers, but it also results in anabsence of information about consensus. Wojahn, Neuwirth, andBullock found that previously posted annotations distracted review-ers away from problems in the text that had no annotations. Boththese findings suggest the importance of having reviewers operateindependently of one another in their reviews. The PREP Editor sup-ports merging the comments of two independent reviewers later. Inlarge-scale, loosely coupled collaborations, conversations via annota-tion threads are relatively rare. Cadiz, Gupta, and Grudin found thatof the 6,263 threads in their database, only 479 threads consisted ofthree or more annotations. How these annotation threads might com-pare with conversation threads made in other communication envi-ronments such as Usenet-type discussion forums or e-mail is not clearalthough preliminary evidence indicates that conversations are rarein these environments as well (see Whittaker et al., “Dynamics”).

Research is needed to compare the costs and benefits of dialoguevia annotation technologies with that conducted via other computer-mediated settings. Discourse analysis methods might be used toexamine conversations conducted via annotation technologies (cf.Cadiz, Gupta, and Grudin) and e-mail listservs or Usenet groups (cf.Whittaker et al., “Dynamics”), comparing the number and length ofmessages, depth of conversational threads, quantity of repeated text,




number of agreements and disagreements, use of mitigation, andother factors. These methods might be combined with surveys orinterviews with participants soliciting their impressions of the con-versations and the technological environments. In addition, studiesmight examine whether the communication environment affectscomprehension or recall of the conversations. Finally, studies mightexamine whether voice annotations enhance computer-mediatedconversations.

CALL ATTENTION TO TOPICS ANDIMPORTANT PASSAGES

Most of the research we have discussed so far focuses on annota-tion technologies that support processes such as reading or collabora-tion. However, annotation programs have also been used to presentfinished texts. In these cases, the annotation is specifically intended tobe part of the final, public document. In journals and academic docu-ments, the most common form of published annotation is the foot-note. However, technical documents often make use of marginalannotations or call-outs. Excellent examples of annotations in techni-cal writing can be found throughout Karen Schriver’s Dynamics inDocument Design.

A few studies of the effects of annotation call-outs have been donein the context of reading to learn (Duchastel and Chen). PhillipeDuchastel identifies two functions for call-outs: (1) identifying andcharacterizing specific textual contents and their relationship to thewhole (e.g., What will be treated in this section? How does it fit withthe rest? What does this term mean?) and (2) summarizing (e.g., Whatis essential? What is secondary?). Call-outs might be helpful becauseas readers process texts, they are assessing the relationship and valueof information encountered in terms of their evolving representationsof the topic, and call-outs can help with that assessment. Research onhow readers actually use such call-outs, especially in the workplace,is much needed.

In print documents, call-outs appear statically in the margins, anddigital technologies offer opportunities for notes that are dynamicallyauthored. Digital technologies also provide new options for varyinghow annotation call-outs appear. Fluid Documents (see Figure 4) is ageneral computer architecture that uses lightweight interactive ani-mation to incorporate annotations in their context. Small visual cues




in the primary text, such as underscoring, announce the presence ofsupporting material. When the reader moves the mouse over a cue,the annotation expands gradually while the surrounding primarytext changes position to accommodate the annotation. Althoughinteractive animation architectures such as Fluid Documents have thepotential to support all of the annotation functions we have discussedso far, we believe that this technology has the most potential as amedium for displaying supplemental information (such as defini-tions, related topics, or elaborations) on published documents. FluidDocuments might also reduce some of the cognitive difficulties associ-ated with reading hypertexts by providing contextual informationabout links before a reader follows them (Zellweger, Chang, andMackinlay). However, whether the potential benefits of this architec-ture would counterbalance the distraction caused by the animationitself is far from clear, although an initial observational study suggeststhat viewers do not find them distracting (Zellweger et al.).

Research is needed on whether using interactive animation to dis-play annotations affects comprehension of the primary text. Simple


When in the Course of human Events, it becomes necessary for onePeople to dissolve the Political Bands which have connected themwith another, and to assume among the Powers of the Earth, theseparate and equal Station to which the Laws of Nature and ofNature's God entitle them, a decent Respect to the Opinions ofMankind requires that they should declare the causes which impelthem to the Separation.We hold these Truths to be self-evident, that all Men are created

As unequal in many ways ashumans may be, no one humanor class of humans is superior toanother human or class ofhumans.

equal, that they are endowed by their Creator with certainunalienable Rights, that among these are Life, Liberty and the Pursuit of Happiness -- That to secure these Rights, Governementsare instituted among Men, deriving their just Powers from theConsent of the Governed, that whenever any Form of Government

Figure 4. Fluid Documents in interlinear compression mode. After the readermoves the mouse over the underscored text “all Men are created,” themain text lines gradually move apart and compress their interline spacingto make room for the annotation, shown in bold, to grow.

SOURCE: Copyright Xerox Corporation, 1998. Courtesy of the Fluid Document Projectat Xerox Palo Alto Research Center (see http://www.parc.xerox.com/istl/projects/fluid/).NOTE: This dynamic design is one of several innovative designs for the display ofannotations on electronic documents.



tests of recall and comprehension might be given to individuals askedto read texts with static and dynamic annotations. Think-aloud proto-cols might also provide important information about how readersprocess and respond to these annotations. Finally, readers should besurveyed for their perceptions and reactions to dynamically anno-tated text. Print technologies, by their very nature, deal withprecomposed static annotations. Fluid Documents, while introducingdynamic display, still assumes that the annotations have beenpreauthored. In contrast, systems are being developed in which theannotations are computed dynamically. Jamey Graham developed asystem, called Reader’s Helper, that dynamically annotates phrases in adocument that match a reader’s indicated topics of interest, with thegoal of helping readers locate relevant sections of a document morequickly. The system, however, has not been empirically evaluated.SuperBook, however, which highlights the table of contents and textbased on a user-entered search query, was extensively studied andfound more effective than print documents on a range of measures(Landauer et al.).6

CONCLUSION

We have sketched a broad agenda for future work on annotations.Much more information is needed about how workers in various pro-fessions read and annotate texts. Observational studies of workers asthey read, cued-recall questionnaires prompting for information onannotations, and expert-novice comparisons would all be appropri-ate research methods here. We also need studies that assess the effec-tiveness of teaching students (or professionals) various annotationstrategies that might help them participate in different discourse com-munities. In addition, research on the benefits or drawbacks of shar-ing annotations made by previous readers will help professional com-municators prepare for new technologies that blur distinctionsbetween public and private annotations. Particularly useful would bethink-aloud protocols that examine how writers use annotations byothers and comparison studies of texts produced by writers workingfrom annotated and clean materials.

More studies are needed on the effects of annotation interfaces onmaking, receiving, and using feedback for revision. Interviews andquestionnaires assessing how writers perceive feedback delivered viadifferent media, textual analyses of the types of changes editors make




and on the communicative effectiveness of these changes, and think-aloud protocols of writers using annotations to revise would all beappropriate methods of analysis. Finally, research testing subjects’comprehension and recall as they read information via differentannotation-display interfaces would help document designers makedecisions about the presentation of their work.

As new reading technologies make the act of reading and annotat-ing electronic documents increasingly comfortable, we will encoun-ter new questions about how to make and use annotations to facilitateother reading and writing processes. The study of annotation practicesand uses will play an important role in these future developments.

NOTES

1. ThirdVoice has generated considerable controversy because the annotations arestored in a central server and are visible to anyone looking at a Web page using theThirdVoice software.

2. Systems allowing annotation on nontextual media are also under development,and little empirical work has been done to evaluate the use of these systems. For exam-ple, the Filochat system synchronizes handwritten note taking in a meeting with audio-tape recordings of the meeting (Whittaker, Hyland, and Wiley). David Bargeron andcolleagues describe a system for annotating Web-based streaming video that has beenused to annotate videos of classroom lectures with questions and answers. Anotherannotation system allowing for annotations on multimedia texts in educational set-tings is reported by Brian Smith and Brian Reiser.

3. In an early study of annotations, Nielsen proposes an interesting taxonomy.4. Nicole Ummelen and Rob Neutelings, for instance, found that subjects thinking

aloud while reading were more likely to judge a text’s content positively, whereas sub-jects reading silently tended to judge the text negatively.

5. This finding may change with new advances in technology. A new system,Jotmail (see Whittaker et al., “Jotmail”), appears to mitigate some of the problems in lis-tening to voice mail.

6. Margin Notes (Rhodes) is another system that processes a text and annotates it onthe fly. The system is able, for example, to annotate a document with citations to otherarticles that are similar to the contents of a paragraph.

REFERENCES

ACM SIGCHI. Special Interest Group on Computer-Human Interaction. Home page.10 Feb. 2001. http://www.acm.org/sigchi.

ACM SIGGROUP. Special Interest Group on Supporting Group Work. Home page. 10Feb. 2001. http://www.acm.org/siggroup/.




Adler, Annette, Anuj Gujar, Beverly L. Harrison, Kenton O’Hara, and Abigail Sellen.“A Diary Study of Work-Related Reading: Design Implications for Digital ReadingDevices.” Computer-Human Interaction 1998. New York: Association for ComputingMachinery, 1998. 241-48.

Adobe Acrobat. Computer program. Vers. 4.0. 10 Feb. 2001. http://www.adobe.com/products/acrobat/.

Anderson, Thomas H., and Bonnie B. Armbruster. “Studying.” Handbook of ReadingResearch. Ed. David P. Pearson. New York: Longman, 1984. 657-79.

Bargeron, David, Annop Gupta, Jonathan Grudin, and Elizabeth Sanocki. “Annota-tions for Streaming Video on the Web.” 10 Feb. 2001. http://www.research.microsoft.com/research/coet/mras/www8/paper.doc.

Blakeslee, Ann M. Interacting with Audiences: Social Influences on the Production of Scien-tific Writing. Mahwah, NJ: Lawrence Erlbaum, 2000.

Brown, John Seely, and Paul Duguid. The Social Life of Information. Boston: HarvardBusiness School Press, 2000.

Bush, Vannevar. “As We May Think.” Atlantic Monthly July 1945: 101-08.Byrne, Ros. “Written Feedback on Student Assignments: Another Look.” Business Com-

munication Quarterly 60.2 (1997): 100-08.Cadiz, J. J., Anoop Gupta, and Jonathan Grudin. “Using Web Annotations for Asyn-

chronous Collaboration around Documents.” Proceedings of the ACM 2000 Confer-ence on Computer Supported Cooperative Work. New York: Association for ComputingMachinery, 2000. 309-18.

Cashen, Valjean M., and Kenneth L. Leicht. “Role of the Isolation Effect in a FormalEducational Setting.” Journal of Educational Psychology 61 (1970): 484-86.

Caverly, David C., and Vincent P. Orlando. “Textbook Study Strategies.” Teaching Read-ing and Study Strategies at the College Level. Ed. Rona F. Flippo and David C. Caverly.Newark, DE: International Reading Association, 1991. 86-165.

Chang, Bay-Wei, Jock D. Mackinlay, Polle T. Zellweger, and Takeo Igarashi. “ANegotia-tion Architecture for Fluid Documents.” Proceedings of the 11th Annual ACM Sympo-sium on User Interface Software and Technology. New York: Association for ComputerMachinery, 1998. 123-32.

Charney, Davida. “How Evolutionists Read ‘The Spandrels of San Marco.’ ” Under-standing Scientific Prose. Ed. Jack Selzer. Madison: University of Wisconsin Press,1993. 203-30.

Churchill, Elizabeth, Jonathan Trever, Sarah Bly, and Les Nelson. “StickyChats: RemoteConversations over Digital Documents.” Proceeding on the ACM 2000 Conference onComputer Supported Cooperative Work. New York: Association for Computer Machin-ery, 2000. 350.

Coleridge, Samuel. Marginalia. Ed. George Whalley and H. J. Jackson. Princeton, NJ:Princeton University Press, 1980. Vol. 12 of The Collected Works of Samuel TaylorColeridge. 16 vols. to date. 1969- .

CommonSpace. Computer program. Vers. 3.5. Boston: Houghton Mifflin, 2000. 10 Feb.2001. http://www.sixthfloor.com/CS1.html.

Craik, Fergus I. M., and Robert S. Lockhart. “Levels of Processing: A Framework forMemory Research.” Journal of Verbal Learning and Verbal Behavior 11 (1972): 671-84.

Critical Tools. Computer program. Austin: University of Texas, 2000. 10 Feb. 2001. http://www.cwrl.utexas.edu/~criticaltools/.

Crouse, James H., and Peter Idstein. “Effects of Encoding Cues on Prose Learning.”Journal of Educational Psychology 63 (1972): 309-13.




Dautermann, Jennie. “Negotiating Meaning in a Hospital Discourse Community.”Writing in the Workplace: New Research Perspectives. Ed. Rachel Spilka. Carbondale:Southern Illinois University Press, 1993. 98-110.

Dayton, David. “Technical Editing Online: The Quest for Transparent Technology.”Journal of Technical Writing and Communication 28 (1998): 3-38.

DeSanctis, Geraldine, and Brad M. Jackson. “Coordination of Information TechnologyManagement: Team-Based Structure and Computer-Based Communication Sys-tems.” Journal of Management Information Systems 10 (1994): 85-100.

Dillon, Andrew. Designing Usable Electronic Text: Ergonomic Aspects of Human Informa-tion Usage. London: Taylor & Francis, 1994.

Doheny-Farina, Stephen. “Writing in an Emerging Organization: An EthnographicStudy.” Written Communication 3 (1986): 158-85.

Doheny-Farina, Stephen, and Lee Odell. “Ethnographic Research on Writing: Assump-tions and Methodology.” Writing in Nonacademic Settings. Ed. Lee Odell and DixieGoswami. New York: Guilford, 1985. 503-36.

Dragga, Sam. “Responding to Technical Writing.” The Technical Writing Teacher 18(1991): 202-21.

Duchastel, Phillipe C. “Marginalia.” The Technology of Text: Principles for Structuring,Designing and Displaying Text. Vol. 2. Ed. David H. Jonassen. Englewood Cliffs, NJ:Educational Technology Publications, 1985. 210-36.

Duchastel, Phillipe C., and Y. Chen. “The Use of Marginal Notes to Assist Learning.”Educational Technology 20 (1980): 41-45.

Duffy, Thomas. “Designing Tools to Aid Technical Editors: A Needs Analysis.” Techni-cal Communication 42 (1995): 262-77.

El-Shinnawy, Maha, and M. Lynne Markus. “Acceptance of Communication Media inOrganizations: Richness or Features?” IEEE Transactions on Professional Communica-tion 41 (1998): 242-53.

Farkas, David K., and Steven E. Poltrock. “Online Editing, Mark-Up Models, and theWorkplace Lives of Editors and Writers.” IEEE Transactions on Professional Communi-cation 38 (1995): 110-17.

Fluid Documents. Computer architecture. Palo Alto, CA: Xerox Palo Alto Research Cen-ter. 10 Feb. 2001. http://www.parc.xerox.com/istl/projects/fluid/.

Fowler, Robert L., and Anne S. Barker. “Effectiveness of Highlighting for Retention ofText Material.” Journal of Applied Psychology 59 (1974): 358-64.

Francik, Ellen, Susan Ehrlich Rudman, Donna Cooper, and Stephen Levine. “PuttingInnovation to Work: Adoption Strategies for Multimedia Communication Systems.”Communications of the ACM 34 (1991): 52-63.

Frohlich, David, and Owen Daly-Jones. “Voicefax: A Shared Workspace for VoicemailPartners.” Companion Proceedings of Conference on Computer Human Interaction. NewYork: Association for Computing Machinery, 1995. 308-09.

Geisler, Cheryl. Academic Literacy and the Nature of Expertise: Reading, Writing, andKnowing in Academic Philosophy. Hillsdale, NJ: Lawrence Erlbaum, 1994.

Golovchinsky, Gene, Morgan N. Price, and Bill N. Schilit. “From Reading to Retrieval:Freeform Ink Annotations as Queries.” Proceedings of SIGIR ’99. New York: Associa-tion for Computing Machinery, 1999. 19-25.

Graham, Jamey. “The Reader’s Helper: A Personalized Document Reading Environ-ment.” CHI ’99. New York: Association for Computing Machinery, 1999. 481-88.

Haas, Christina. “Learning to Read Biology: One Student’s Rhetorical Development inCollege.” Written Communication 11 (1994): 43-84.




———. Writing Technology: Studies on the Materiality of Literacy. Mahwah, NJ: LawrenceErlbaum, 1996.

Haas, Christina, and Linda Flower. “Rhetorical Reading Strategies and the Construc-tion of Meaning.” College Composition and Communication 39 (1988): 167-84.

Hardock, Gary, Gordon Kurtenbach, and William Buxton. “A Marking Based Interfacefor Collaborative Writing.” Proceedings of the Sixth Annual ACM Symposium on UserInterface Software and Technology. New York: Association for Computing Machinery,1993. 259-66.

Haswell, Richard H., Terri L. Briggs, Jennifer A. Fay, Norman K. Gillen, Rob Harrill,Andrew M. Shupala, and Sylvia S. Trevino. “Context and Rhetorical Reading Strat-egies: Haas and Flower (1988) Revisited.” Written Communication 16 (1999): 3-28.

Higgins, Lorraine. “Reading to Argue: Helping Students Transform Source Texts.”Hearing Ourselves Think: Cognitive Research in the College Writing Classroom. Ed.Ann M. Penrose and Barbara M. Sitko. New York: Oxford University Press, 1993.70-101.

Houser, Walter R., and Jason Hart. “Software Helps Tidy Up FOIAResponses.” Govern-ment Computer News 16.33 (1997): 29-31.

Huettman, Elizabeth. “Writing for Multiple Audiences: An Examination of AudienceConcerns in a Hospitality Consulting Firm.” Journal of Business Communication 33(1996): 257-74.

Huttenlocher, Daniel. “CoNote: A System for Supporting Collaboration with SharedDocuments.” 10 Feb. 2001. http://www3.cs.cornell.edu/dph/docs/annotation/annotations.html.

Kaufer, David S., and Cheryl Geisler. “Novelty in Academic Writing.” Written Commu-nication 6 (1989): 286-311.

Kennedy, Mary Lynch. “The Composing Process of College Students Writing fromSources.” Written Communication 2 (1985): 434-56.

Kiewra, Kenneth A. “A Review of Note-Taking: The Encoding-Storage Paradigm andBeyond.” Educational Psychology Review 1 (1989): 147-72.

Kraut, Robert, Jolene Galegher, Robert Fish, and Barbara Chalfonte. “Task Require-ments and Media Choice in Collaborative Writing.” Human-Computer Interaction 7(1992): 375-407.

Landauer, Thomas, Dennis Egan, Joel Remde, Michael Lesk, Carol Lochbaum, andDaniel Ketchum. “Enhancing the Usability of Text through Computer Delivery andFormative Evaluation: The SuperBook Project.” Hypertext: A Psychological Perspec-tive. Ed. Andrew Dillon, Cliff McKnight, and John Richardson. New York:Horwood, 1993. 72-136.

Lotus Notes. Computer program. Release 5.0. Lotus Development Corporation. 10 Feb.2001. http://www.lotus.com/home.nsf/welcome/notes.

Luff, Paul, Christian Heath, and David Greatbatch. “Tasks-in-Interaction: Paper andScreen Based Documentation in Collaborative Activity.” Computer-Supported Collab-orative Work 1992 Proceedings. New York: Association for Computing Machinery,1992. 163-70.

Lunsford, Andrea A., and John J. Ruszkiewicz. The Presence of Others. 3rd ed. New York:St. Martin’s, 1999.

Marshall, Catherine C. “Annotation: From Paper Books to the Digital Library.” DigitalLibraries 1997. New York: Association for Computing Machinery, 1997. 131-40.

———. “Toward an Ecology of Hypertext Annotation.” Hypertext98. New York: Asso-ciation for Computing Machinery Press, 1998. 40-49.




Marshall, Catherine C., Morgan N. Price, Gene Golovchinsky, and Bill N. Schilit. “Intro-ducing a Digital Library Reading Appliance into a Reading Group.” Digital Libraries99. New York: Association for Computing Machinery, 1999. 77-84.

McGee, Lynn. “Communication Channels Used by Technical Writers throughout theDocumentation Process.” IEEE Transactions on Professional Communication 43 (2000):35-50.

Microsoft Office 2000. Computer program. Microsoft Corporation. 10 Feb. 2001. http://www.microsoft.com/office/.

Microsoft Word 2000. Computer program. Microsoft Corporation. 6 Apr. 2001. http://www.microsoft.com/office/word.

Mirel, Barbara. “Beyond the Monkey House: Audience Analyses in ComputerizedWorkplaces.” Writing in the Workplace: New Research Perspectives. Ed. Rachel Spilka.Carbondale: Southern Illinois University Press, 1993. 21-40.

Neuwirth, Christine M., Ravinder Chandhok, Davida Charney, Patricia Wojahn, andLoel Kim. “Distributed Collaborative Writing: A Comparison of Spoken andWritten Modalities for Reviewing and Revising Documents.” Human Factors inComputing Systems. New York: Association for Computing Machinery, 1994. 51-57.

Neuwirth, Christine M., Ravinder Chandhok, David S. Kaufer, Paul Erion, JamesMorris, and Dale Miller. “Flexible Diff-Ing in a Collaborative Writing Environment.”Conference Proceedings on Computer Supported Cooperative Work (CSCW ’92). NewYork: Association for Computing Machinery, 1992. 147-54.

Neuwirth, Christine M., David S. Kaufer, Ravinder Chandhok, and James H. Morris.“Issues in the Design of Computer Support for Co-Authoring and Commenting.”Conference on Computer Supported Cooperative Work (CSCW ’90). New York: Associa-tion for Computing Machinery, 1990. 193-95.

Neuwirth, Christine M., David Kaufer, Rick Chimera, and Terilyn Gillespie. “TheNotes Program: A Hypertext Application for Reading and Writing.” Hypertext ’87Proceedings. New York: Association for Computing Machinery, 1987. 121-41.

Neuwirth, Christine M., and Patricia G. Wojahn. “Learning to Write: Computer Sup-port for a Cooperative Process.” CSCL: Theory and Practice of an Emerging Paradigm.Ed. Timothy Koschmann. Mahway, NJ: Lawrence Erlbaum, 1996. 147-70.

Nichols, David M., Duncan Pemberton, Salah Dalhoumi, Omar Larouk, Clair Belisle,and Michael B. Twidale. “DEBORA: Developing an Interface to Support Collabora-tion in a Digital Library.” Proceedings of the Fourth European Conference on Research andAdvanced Technology for Digital Libraries (ECDL2000). Lisbon, Port.: Springer-Verlag,2000. 239-48.

Nielsen, Jakob. “Online Documentation and Reader Annotation.” Proceedings of the 1stInternational Conference on Work with Display Units. Amsterdam: Elsevier, 1986. 526-29.

Ovsiannikov, Ilia A., Michael A. Arbib, and Thomas H. McNeil. “Annotation Technol-ogy.” International Journal of Human-Computer Studies 50 (1999): 329-62.

Paradis, James, David Dobrin, and Richard Miller. “Writing at Exxon ITD: Notes on theWriting Environment of an R&D Organization.” Writing in Nonacademic Settings.Ed. Lee Odell and Dixie Goswami. New York: Guilford, 1985. 281-308.

Pearce, C. Glenn, and R. Jon Ackley. “Audiotaped Feedback in Business Writing: AnExploratory Study.” Business Communication Quarterly 58.3 (1995): 31-35.

Penrose, Ann M., and Cheryl Geisler. “Reading and Writing without Authority.” Col-lege Composition and Communication 45 (1994): 505-20.

Poe, Edgar Allen. The Centenary Poe: Tales, Poems, Criticism, Marginalia, and Eureka. Ed.Montagu Slater. London: Bodley Head, 1949.




PREP Editor. Computer program. Carnegie Mellon University, 1990-93. 10 Feb. 2001.http://eserver.org/software/prep.

Rhodes, Bradley J. “Margin Notes: Building a Contextually Aware Associative Mem-ory.” Proceedings of the 2000 International Conference on Intelligent User Interfaces. NewYork: Association for Computing Machinery, 2000. 219-24.

Rice, Gary E. “Examining Constructs in Reading Comprehension Using Two Presenta-tion Modes: Paper vs. Computer.” Journal of Educational Computing Research 11(1994): 153-78.

Rickards, John P., and Gerald J. August. “Generative Underlining Strategies in ProseRecall.” Journal of Educational Psychology 67 (1975): 860-65.

Roberts, David D., and Patricia A. Sullivan. “Beyond the Static Audience Construct:Reading Protocols in the Technical Writing Class.” Journal of Technical Writing andCommunication 14 (1984): 143-53.

Schilit, Bill N., Morgan N. Price, Gene Golovchinsky, Kei Tanaka, and Catherine C.Marshall. “As We May Read: The Reading Appliance Revolution.” Computer Jan.1999: 65-73.

Schnell, Thomas R., and Daniel J. Rocchio. “A Comparison of Underlining Strategiesfor Improving Reading Comprehension and Retention.” Reading Horizons 18 (1978):106-09.

Schriver, Karen A. Dynamics in Document Design: Creating Text for Readers. New York:Wiley Computer Publishing, 1997.

Schumacher, Gary M., and Jane Gradwohl Nash. “Conceptualizing and MeasuringKnowledge Change Due to Writing.” Research in the Teaching of English 25 (1991): 67-96.

Simpson, Michele L., and Sherrie L. Nist. “Textbook Annotation: An Effective and Effi-cient Study Strategy for College Students.” Journal of Reading 34 (1990): 122-29.

Smart, Graham. “Genre as Community Invention: A Central Bank’s Response to ItsExecutives’ Expectations As Readers.” Writing in the Workplace: New Research Per-spectives. Ed. Rachel Spilka. Carbondale: Southern Illinois University Press, 1993.124-40.

Smith, Brian K., and Brian J. Reiser. “What Should a Wildebeest Say? Interactive NatureFilms for High School Classrooms.” Proceedings of the Fifth ACM International Confer-ence on Multimedia. New York: Association for Computing Machinery, 1997. 193-201.

Spilka, Rachel. “Influencing Workplace Practice: A Challenge for Professional WritingSpecialists in Academia.” Writing in the Workplace: New Research Perspectives. Ed.Rachel Spilka. Carbondale: Southern Illinois University Press, 1993. 207-19.

Spivey, Nancy Nelson. “Transforming Texts: Constructive Processes in Reading andWriting.” Written Communication 7 (1990): 256-87.

Suchan, James, and Ron Dulek. “Toward a Better Understanding of Reader Analysis.”Journal of Business Communication 25 (1988): 29-46.

Third Voice. Computer program. Third Voice, Inc. 10 Feb. 2001. http://www.thirdvoice.com.

Ummelen, Nicole, and Rob Neutelings. “Measuring Reading Behavior in Policy Docu-ments: A Comparison of Two Instruments.” IEEE Transactions on Professional Com-munication 43 (2000): 292-301.

Van Dam, Andries. “Hypertext ’87 Keynote Address.” Communications of the ACM 31(1988): 887-95.

Van Horn-Christopher, Doris A. “Voice-Graded Business Communication Documents.”Business Communication Quarterly 58.3 (1995): 35-38.




Van Oostendorp, Herre. “Studying and Annotating Electronic Text.” Hypertext and Cog-nition. Ed. Jean-Francois Rouet, Jarmo J. Levonen, Andrew Dillon, and Rand J. Spiro.Mahwah, NJ: Lawrence Erlbaum, 1996. 137-48.

Walker, Janet H. “Document Examiner: Delivery Interface for Hypertext Documents.”Proceedings of the ACM Conference on Hypertext. New York: Association for Com-puting Machinery, 1987. 307-23.

Whittaker, Steve, Richard Davis, Julia Hirschberg, and Urs Muller. “Jotmail: AVoicemail Interface That Enables You to See What Was Said.” Proceedings of the CHI2000 Conference on Human Factors in Computing Systems. The Hague, Neth.: Associa-tion for Computing Machinery, 2000. 89-96.

Whittaker, Steve, David Frohlich, and Owen Daly-Jones. “Informal Workplace Com-munication: What Is It Like and How Might We Support It?” Proceedings of theCHI’94 Conference Companion on Human Factors in Computing Systems. New York:Association for Computing Machinery, 1994. 131-37.

Whittaker, Steve, Patrick Hyland, and Myrtle Wiley. “Filochat: Handwritten Notes Pro-vide Access to Recorded Conversations.” Conference Proceedings of Human Factors inComputing Systems: “Celebrating Interdependence.” New York: Association for Com-puting Machinery, 1994. 271-76.

Whittaker, Steve, Loren Terveen, Will Hill, and Lynn Cherny. “The Dynamics of MassInteraction.” Proceedings of the ACM 1998 Conference on Computer Supported Coopera-tive Work. New York: Association for Computing Machinery, 1998. 257-64.

Wojahn, Patricia. “Communication in Peer Reviews: Effects of Computer Interfaces,Problem Equivocality, and Initial Annotations.” Diss. Carnegie Mellon University,1999.

Wojahn, Patricia G., Christine M. Neuwirth, and Barbara Bullock. “Effects of Interfacesfor Annotation on Communication in a Collaborative Task.” CHI ’98, Conference Pro-ceedings on Human Factors in Computing Systems. New York: Association for Com-puting Machinery, 1998. 456-63.

Wolfe, Joanna. “The Effects of Annotations on Student Readers and Writers.” DL ’00:Proceedings of the Fifth ACM Conference on Digital Libraries. New York: Association forComputing Machinery, 2000. 19-26.

———. “Pedagogical Uses of Annotations and Annotation Technologies.” Diss. Uni-versity of Texas at Austin, 2001.

Wright, Patricia. “The Need for Theories of NOT Reading: Some PsychologicalAspects of the Human-Computer Interface.” Working Models of Human Perception.Ed. Ben A. G. Elsendoorn and Herman Bouma. San Diego: Academic Press, 1989.319-40.

XLibris. Computer program. Palo Alto, CA: Xerox FX Palo Alto Laboratory. 10 Feb.2001. http://www.fxpal.xerox.com/xlibris/.

Zellweger, Polle T., Bay-Wei Chang, and Jack D. Mackinlay. “Fluid Links for Informedand Incremental Link Transitions.” Hypertext ’98, Proceedings of the Ninth ACM Con-ference on Hypertext and Hypermedia: Links, Objects, Time and Space-Structure inHypermedia Systems. New York: Association for Computer Machinery, 1998. 50-57.

Zellweger, Polle T., Susan H. Regli, Jack D. Mackinlay, and Bay-Wei Chang. “TheImpact of Fluid Documents on Reading and Browsing: An Observational Study.”Proceedings of the CHI 2000 Conference on Computer-Human Interaction. New York:Association for Computer Machinery, 2000. 249-56.




Joanna L. Wolfe is an assistant instructor and PhD candidate in computers and Englishstudies at the University of Texas at Austin. Her work on computer-mediated communi-cation has recently appeared in Written Communication and Computers and Com-position. She may be reached by e-mail at [email protected].

Christine M. Neuwirth is an associate professor of English and human-computer inter-action at Carnegie Mellon University. Her research interests lie primarily in the areas ofcomputer-mediated communication, especially support for the cognitive and social pro-cesses of reading and writing. She may be reached by e-mail at [email protected].


