Hypertext Paths and the World-Wide Web: Experiences
`with Walden’s Paths
`
`Richard Furuta, Frank M. Shipman III, Catherine C. Marshall, Donald Brenner, and Hao-wei Hsieh
`Center for the Study of Digital Libraries and Department of Computer Science
`Texas A&M University
`College Station, TX 77843–3112, USA
`E-mail: ffuruta, shipman, marshall, dbrenner, haoweig@csdl.tamu.edu
`
`ABSTRACT
`Walden’s Paths applies the concept of hypertextual paths to
`the World-Wide Web. Walden’s Paths is being developed
`for use in the K–12 school environment. The heterogene-
`ity of the Web coupled with the desirability of supporting
`the teacher-student relationship make this an interesting and
`challenging project. We describe the Walden’s Paths imple-
`mentation, discuss the elements that affected its design and
`architecture, and report on our experiences with the system
`in use.
`
`KEYWORDS: Walden’s Paths, hypertext tours and paths,
`meta-structure, educational applications
`
`1 INTRODUCTION
`Hypertext paths, also called tours, provide a meta-structuring
`mechanism that allows the affiliation of elements from many
`different hypertexts. The association of paths with hyper-
`text is a concept as old as hypertext itself. Bush’s Memex
`[2] incorporated paths for two purposes: first, to provide a
`personal means for remembering and organizing found in-
`formation, and second, to provide a means to communicate
`that information to friends and associates.
`
`Paths appear in implemented systems beginning around
`1988 in Trigg’s Guided Tours [14], integrated into Note-
`Cards [4], and in Zellweger’s Scripted Paths [15, 16, 17],
`a stand-alone environment. Guided Tours, centered around
`a graphical representation, provided the means to asso-
`ciate sets of cards with each stop along the tour. Scripted
`Paths allowed stops in documents or applications supported
`
` This work is supported by the DARPA CAETI program. Approved
`for Public Release, Distribution Unlimited. This material is based in part
`on work supported by the Texas Advanced Research Program under Grant
`Number 99903-230. The project’s World-Wide Web pages are located at
`http://www.csdl.tamu.edu/walden/.
`
`Permission to make digital/hard copies of all or part of this material for
`personal or classroom use is granted without fee provided that the copies
`are not made or distributed for profit or commercial advantage, the copy-
`right notice, the title of the publication and its date appear, and notice is
`given that copyright is by permission of the ACM, Inc. To copy otherwise,
`to republish, to post on servers or to redistribute to lists, requires specific
`permission and/or fee.
`Hypertext 97, Southampton UK
`c1997 ACM 0-89791-866-5...$3.50
`
`by “compliant” programs specified with a programming-
`language-like specification language.
`
`Both Guided Tours and Scripted Paths were specialized
`mechanisms built for their own customized environments.
`The published accounts suggest that they were intended for
`use by a sophisticated audience of hypertext authors and
`readers. The documents or nodes included in a Guided
`Tour or a Scripted Path were considered to be under local
`control—created in the same environment and acknowledg-
`ing the same sensibilities as the tour or path meta-structure
`document. Experiences with Guided Tours noted, however,
`that a tour author needed to contribute significant annotative
`material to pitch an existing hypertext to a new audience of
`peers, and that the Tours themselves tended to have a linear
`spine, a minimal subset of tour stops that were assumed to
`be necessary for explication and intelligibility [6]. These use
`characteristics foreshadow in a limited way what we see to-
`day in a more heterogeneous environment of documents, au-
`thors, and readers.
`
`The World-Wide Web provides an entirely new context for
`hypertext application. Although the technology of the Web
`is largely derivative from earlier applications, the environ-
`ment that it has created is novel. Unlike its predecessors,
`the Web is highly heterogeneous, both in readers but also in
`information provided. Readers range from highly-educated
`academicians to elementary school students. Ages range
`from retirees to pre-schoolers. People make material avail-
`able on the Web for all conceivable reasons—to communi-
`cate, educate, persuade, promote, defraud, and delude, to
`name just a few of the possibilities.
`
`Path-like mechanisms that support personal-space organiza-
`tion are commonplace in Web browsers. The popular Web
`browsers all retain bookmark lists, sometimes hierarchical in
`structure. Brown and Shillner have described a new browser
`called Deckscape [1] that represents path-like groups of Web
`pages as decks, with particular implementation support for
`collecting and browsing decks.
`
`Less common, however, are Web applications that use paths
`to support interpersonal communication. Sharing of book-
`mark files is provided in some systems, but paths have addi-
`
`167
`
`1 of 10
`
`

`

`tional communicative potential that is not being tapped.
`
`In this paper we discuss Walden’s Paths, a Web-based path
`implementation. Walden’s Paths are intended to be com-
`municated from path author to path readers. The basic
`metaphor is that of a network-based meta-document: a doc-
`ument whose elements are themselves documents. In gen-
`eral the author of the path is not the author of the support-
`ing documents, a characteristic that raises interesting rhetor-
`ical and societal issues. Moreover because our current appli-
`cation environment is in the primary and secondary schools
`(i.e., in support of curriculum for younger students, most of
`whom are between 5 and 18 years old), we found additional
`needs for contextualization perhaps not as important in the
`peer-to-peer environments present in earlier systems.
`
`The next section of this paper describes the characteristics of
`the target student population. Sections 3 and 4 describe the
`Walden’s Paths implementation and the architectural consid-
`erations behind it. We then discuss our experiences with
`Walden’s Paths in use in sections 5 and 6. These experiences
`have lead to further extensions and developments, summa-
`rized in section 7. Finally we conclude with more global ob-
`servations in section 8.
`
`2 STUDENTS AND THE WEB
`How do students experience the World-Wide Web as it is
`today? We had the opportunity to observe four classes of
`sixth graders from a local middle school who were using
`Netscape’s browser to explore the Web [8]. The classes took
`turns in a campus engineering computer lab at Texas A&M
`University for these sessions; each student who participated
`in a session sat at his or her own PC, so each was free to ex-
`plore at will. And explore they did.
`
`At the start of each session, all the Netscape browsers in the
`room displayed the same starting page, a page that had been
`explicitly designed by the instructors for this instructional
`setting. The page included “starting out” links to different
`kinds of on-line places: Museums and public institutions
`(e.g., the Franklin Institute Science Museum, the Smithso-
`nian Institute, and the White House), collections of pictures
`(e.g., the entomological collection at University of Illinois),
`peer-related Web activities (e.g., Chico Jr. High goes on-
`line), and entertainment-oriented sites (e.g., the Cartoon Fac-
`tory Animation Art Gallery).
`
`Instruction at the beginning of the sessions covered the
`“how-tos”—how to navigate and how to search; the instruc-
`
` The Walden’s Paths name was inspired by Henry David Thoreau’s clas-
`sic book Walden, first published in 1854. In Walden Thoreau describes his
`life for the two years and two months in 1845 and 1846 he spent living in
`the woods a mile from the nearest neighbor in a self-built cabin at Walden
`Pond, located outside of Concord, Massachusetts. Although Thoreau sought
`the solitude of life at Walden Pond, he quickly discovered that his daily ac-
`tivities created paths in the woods, and he suspected that others adopted his
`paths, keeping them from fading back into the woods. It is this relationship
`between early trailblazers and later explorers that motivated our choice of
`the Walden’s Paths name.
`
`tors encouraged active exploration beyond the initial set of
`links. The computers in this lab imposed some restrictions
`on what the students could access and run (for example, they
`could not download games or play audio files although some
`tried); but in general the browsers provided much unfettered
`access to popular Internet resources.
`
`We observed a variety of phenomena that suggest some de-
`sign elements and considerations of a path facility that will
`meet the needs of elementary and secondary students.
`In
`general, the students’ experiences were positive. Many were
`pleased with what they found, and the apparent ease with
`which they could navigation among pages and sites. Of
`course, a few were not engaged, and were more or less pas-
`sive, clicking a few times, and deciding they were getting
`nowhere; a few “got stuck” on pages, and could not return
`to a familiar orienting place. Five phenomena stood out: the
`difficulties associated with getting started and reaching nav-
`igational dead ends; the sociability of Web use and the col-
`laborative nature of exploration; the success of simple navi-
`gational modes over more sophisticated ones; the confusion
`in interpreting the signs and signals of network performance;
`and finally, the compelling quality of participation, rather
`than just interaction.
`
`Starting out and getting stuck. Some students hesitated when
`they were confronted with the unfamiliar software and range
`of navigational choices; they found it difficult to set out on
`an unstructured exploration (undertaken for no apparent rea-
`son other than to explore). Students occasionally “got stuck”
`as well. That is, they reached an apparently nonsensical page
`which offered them no easy way back to the special page set
`up for the class. In most cases, this happened when students
`pressed the Home button, which brought them to the home
`page for the engineering lab that normally took place in the
`room, an unfamiliar landmark which did nothing to reorient
`them. A stuck student would usually remain stuck unless
`there was some sort of intervention.
`
`Active collaboration. We were struck with the sociability of
`Web use; students showed each other what they’d found and
`how they got there. Many of the students chose to work to-
`gether, which included showing each other paths to interest-
`ing Web sites (one particularly popular site—trophy pages
`for many of our young cybernauts—was MTV’s Beavis and
`Butthead collection). When we listened to the students talk-
`ing to each other, we heard things like, “How did you do
`that?” and “How did you get there?” Students helped each
`other navigate in a very literal sense: they would help an-
`other student retrace their own steps to get to a particular
`page on a Web site. Most did not use the shortcut of telling
`the other student the URL of the page of interest, although
`during the sessions that took place late in the day, it became
`clear that such a strategy was growing in popularity.
`
`Navigational strategies. Most students—perhaps just be-
`cause they were neophytes—appeared to prefer the standard
`
`168
`
`2 of 10
`
`

`

`lighted when they were given the opportunity to participate,
`rather than just navigate. Of course, this was most evident in
`the active collaborative partnerships that formed, but it also
`manifested itself in other ways. One student was particularly
`pleased that she had located an on-line pen-pal service. With
`some guidance, she submitted her name and U.S. Mail ad-
`dress, and requested a pen-pal in China. She was careful to
`explain that she needed to send and receive paper mail, since
`she had no permanent email address. Her enthusiasm was
`considerable, and she was anxious to share her experience
`with her classmates.
`
`Anticipating these phenomena, and addressing them through
`the design of the path mechanism, and the design of indi-
`vidual paths, will help primary and secondary students use
`the Web effectively. Certain principles emerge: first, make
`it easy to get started and to stay oriented; acknowledge the
`social and participatory nature of exploration; keep naviga-
`tional options accessible; and finally, develop caching and
`path validation schemes that hide some of the complexity
`(and frustration) of navigating the Web. Next, we describe
`the Walden’s Paths mechanism, which takes into account
`these principles.
`
`3 MECHANISM
`While the Internet provides access to a wealth of new mul-
`timedia materials, especially through the extensive, encyclo-
`pedic materials available on the World-Wide Web, the stu-
`dents need to be provided a focus based on the curricular ob-
`jectives.
`
`Walden’s Paths is designed to enable teachers to make use of
`materials available via the Internet by creating directed paths
`over the World-Wide Web [13]. Besides providing an or-
`dering of pages, Walden’s Paths allows teachers to provide
`additional context for the page through annotation. By pro-
`viding text or other annotations in addition to the content of
`the page, the teacher may provide a rhetorical structure to
`the path as a whole, create transitions to fill in informational
`gaps between pages, and create emphasis to particular as-
`pects of the materials.
`
`Figure 1 shows a page from a path on endangered species in
`Texas. At the top of the page is a header including control
`buttons for moving back and forth in the path, numbers in-
`dicating the current position and length of the path, and the
`annotation/introductory text provided by the teacher. The in-
`troductory text on the pages of this path bridge together in-
`formation on specific plant and animal species with informa-
`tion about wildlife preservation and ecological policies.
`
`Students following a path may follow encountered links to
`freely examine the information space. A student’s need for
`detailed explanation, alternative discussion, or details on re-
`lated topics are satisfied by his or her own desires to explore,
`to construct knowledge, to find information. Such explo-
`rations don’t cause the student to lose the organizing context
`
`169
`
`Figure 1: Page from path on endangered species in
`Texas
`
`hypertext navigation that the Web offers rather than using
`search engines, or the shortcut of typing in a URL. Students
`were shown how to search at the outset of the class, but only
`a few tried the search engines; still fewer tried them with
`search terms of their own, preferring to use instead the re-
`liable keywords demonstrated by the instructors. This said,
`the classes near the end of the day were affected by the dis-
`coveries of the students in the earlier sessions. During a
`morning session, one student discovered MTV’s Beavis and
`Butthead site. Soon other students were following his path
`there, and apparently by the end of the day, students had fig-
`ured out how to pass the URL to one another to navigate
`there directly.
`
`Interpreting net delays. Network delays (and occasional un-
`predictability in response) caused the students some confu-
`sion. Because most of them did not have much (if any)
`experience with networked computers, they were unable to
`make sense of this sort of behavior. Characteristics of net-
`work and Web use—slow response, missing pages, and ac-
`cess denied messages—that are merely annoying to experi-
`enced computer-users becomes baffling to students (and oth-
`ers) with a less complete understanding of distributed sys-
`tems.
`
`Participation, not just interaction. The students seemed de-
`
`3 of 10
`
`

`

`Figure 2: Off-path pages include button to return to
`path
`
`provided by the path as Walden’s Paths augments off-path
`pages with controls that allow immediate return to the path.
`Figure 2 shows a page available to the student by browsing
`off from the endangered species page in Figure 1.
`
`Keeping the focus of the project on providing teachers an
`easy way to collect materials for use in the classroom has
`motivated a KISS (“keep it simple, stupid”) strategy. While
`earlier work on guided paths allowed the creation of generic
`directed graphs, Walden’s Paths currently uses a strictly lin-
`ear representation—a path consists of pages 1 through n in a
`strict sequence.
`
`To support teachers authoring paths we have created a Path
`Authoring Tool. It supports the process of locating sites with
`relevant information, selecting individual pages from those
`sites, and ordering and annotating those pages to form the
`path. Figure 3 shows a prototype of this tool.
`
`Prior to the existence of the Path Authoring Tool, several of
`the paths were authored by collecting and organizing Web-
`based materials in VIKI [7], a spatial hypertext system de-
`signed to support information analysis. From these experi-
`ences functionality was added to VIKI so information could
`be arranged into spatial structures (e.g., lists or outlines) and
`exported as Walden’s Paths path files [12].
`
`170
`
`Figure 3: Prototype of the path authoring tool
`
`4 ARCHITECTURE
`Figure 4 shows the architecture for Walden’s Paths.
`Walden’s Paths consists of two main components: the Path
`Authoring Tool and the Path Server. As described above,
`the Path Authoring Tool enables the teacher to create, mod-
`ify, validate, and reuse paths. When a path is saved by the
`teacher it is stored with the Path Server. The Path Server is
`the implementation means through which students navigate
`the paths.
`
`In order to avoid requiring students or teachers to use spe-
`cific software, the Path Server works as an intermediary be-
`tween the students’ Web browser (e.g., Netscape) and the
`http servers providing the original material being used in
`the paths. The Path Server is a Common Gateway Interface
`(CGI) that recognizes requests for path information. When it
`receives a request, the Path Server determines the path and
`page number of the request, or if the request is to go off
`of the path, and retrieves the needed information to dynami-
`
`4 of 10
`
`

`

`Figure 4: Diagram of the Walden’s Paths architecture
`
`cally generate the page.
`
`Because schools frequently have limited connectivity and
`bandwidth, the Path Server can also cache pages of the path
`prior to their use in the classroom or cache pages as they are
`browsed by the class. Since a single server is used by all the
`students, the cache is shared rather than being per individ-
`ual like the browser caches. Of course this assumes that the
`school has a local area network with one machine running
`the Path Server (currently this machine must be a Sun work-
`station.)
`
`The Walden’s Paths architectural design was influenced by
`both practical and pedagogical considerations. Our project
`team is relatively small in number—particularly when com-
`pared with the giant programming teams occupied in pro-
`ducing the currently-popular Web browsers (e.g., Netscape).
`We had no desire to attempt to track the rapid changes tak-
`ing place in browser technology, no resources to attempt
`to support the multiple hardware platforms that are com-
`monplace in school environments, and no interest in fight-
`ing the frequently political battles necessary to get schools
`to adopt standard software configurations in their client ma-
`chines. Consequently we decided to implement our system
`as a protocol filter—tapping into and modifying the stan-
`dardized protocol stream that is exchanged between browser
`and server in a way that is transparent to both. As such we
`are buffered from the effects of updates to the browser, the
`part of the Web environment that is currently most subject to
`rapid change.
`
`In related work, the Footsteps project [10] also provides a
`Web-based path mechanism and independently has devel-
`oped some of the same implementation and architectural
`components as Walden’s Paths, although they have not yet
`addressed issues of authoring, annotation, and genre. We ex-
`pect that our experiences in use of Walden’s Paths will be of
`interest in that project.
`
`5 EXPERIENCE WITH USE OF WALDEN’S PATHS
`The Path Server has been available to the public since March
`1996, accessible via the World-Wide Web and supporting an
`
`increasingly-large collection of paths written by authors both
`within the project but also authors recruited from the teach-
`ing community. The first path contributed from outside of
`the project was written in April 1996 by a teacher who saw
`one of our early presentations. Additional paths have been
`contributed by teachers recruited from elementary and sec-
`ondary school technology coordinators from schools around
`Texas; in particular six teachers wrote nine paths during a
`three-day period in June 1996 as part of an extended work-
`shop. Since the Authoring Tool was not available at the time
`of this workshop, we defined an ASCII-based input repre-
`sentation and asked the teachers to use generally-available
`text editors to create paths in the representation. The teach-
`ers were uniformly successful in doing this, and primarily
`relied on cut-and-paste operations that mimic functions now
`provided automatically by the Path Authoring tool. Walden’s
`Paths currently is being considered for deployment in several
`K–12 schools.
`
`The teachers who have written paths in the system have been
`enthusiastic in their evaluations. Comments indicate that
`they view paths as an effective way to draw together mate-
`rials from disparate sources. Indeed one teacher from the ex-
`tended workshop was enthusiastic enough that on his return
`to his home school, he recruited the assistance of a graduate
`student and implemented his own subset version of Walden
`Paths. The primary requests by the teachers for enhance-
`ment have been for mechanisms to increase the level of in-
`teraction students have with paths to help maintain and mea-
`sure student engagement with the material. Two approaches
`that we are now considering include providing opportuni-
`ties for associating student-generated annotations with paths
`and extensions to our path specification to permit conditional
`branches based on student responses to questions.
`
`One surprise for us has been the rapidity with which our
`technology “disappears”—becomes an element of the land-
`scape rather than a focus of attention. We had our first hint
`that this was occurring shortly after we placed our public
`version on the Web. This initial release carried notifications
`that the software and paths were prototypes and should be
`viewed as examples of the capabilities of the system rather
`than as finished products. One of the strongest comments
`we received at the feedback email address concerned not the
`technology but instead the content of one of the pages in-
`cluded on one of the paths. The commentator clearly had
`become engaged enough with the path material that distinc-
`tion between the contents (i.e., the original material) and the
`container (i.e., the path system) had disappeared.
`
`We saw similar effects with the teacher-authored paths. The
`extended workshop concluded with project presentations by
`the participants. After a brief, almost cursory, acknowledg-
`ment of the technology, most of the teachers focused on the
`content contained within the path. Once again, the material
`of importance was the contents rather than the container.
`
`171
`
`5 of 10
`
`

`

`We find the message of these effects to be both reassuring but
`also cautionary. The effects are reassuring because they are
`precisely what we wish to attain—the provision of new lay-
`ers of meta-structure that enhance rather than diminish ac-
`cess to the content that they place into new contexts. But
`these reactions also remind us of the changing nature of own-
`ership and attribution for networked information; effects that
`have implications for the models we implement. We return
`to this point subsequently.
`
`6 META-DOCUMENTS, DOCUMENTS, AND WALDEN’S
`PATHS
`Paths are an interesting form of document (i.e., a meta-
`document): one in which the constituent parts are them-
`selves documents, not the more primitive elements that make
`up most documents. The rhetorical purpose of a path is
`likely to be distinct from that of its components. Compo-
`nents frequently are part of a tightly interlinked network de-
`voted to elucidation of a particular point of view about a se-
`lected set of topics. On the other hand the path is a unifying
`structure, directed towards contextualizing multiple points of
`view about a topic into an organizing framework.
`
`Because of Walden’s Paths’ use in conveying curriculum
`from teacher to student, there is a strong need to be able to
`make explicit the path’s the rhetorical structure; since the
`path is a meta-document, the student often needs help in un-
`derstanding why an element has been included. This charac-
`teristic differs from a bookmark facility, in which the more
`homogeneous nature of the environment allows implicit ex-
`pression of structure. Furthermore because the rhetorical
`goals of the constituent documents are independent and dis-
`tinct from that of the path, there is often a desire to adapt the
`underlying document to bring it into closer agreement—this
`is a question of interest not only from a technical perspective
`but also from a societal one. Finally, from a practical point
`of view the dynamic nature of the Web guarantees that the
`constituent documents will be evolving, mutating, and disap-
`pearing. Successful implementation of paths must maintain
`the integrity of the meta-document in the face of this change.
`
`Paths that have been authored in Walden’s Paths to date
`appear to have been created with one of three overlapping
`rhetorical structures: extended bookmark lists, extended
`tours, and curriculum-based expositions.
`
`Extended bookmark lists appear to be generated as an or-
`ganizing mechanism—a translation of a “hot list” into a
`sequentially-traversable form. Figure 5 shows an entry taken
`from one such path (indeed in this path, the content is a “vir-
`tual” hot list—the dynamically-constructed results of a Ya-
`hoo search). In extended bookmark lists, annotations tend
`to be brief or absent. If present they often provide a short
`abstract of the page’s contents—almost as a memory-jogger
`to aid the author in remembering what is important about
`the page. Frequently the entries on the extended book-
`mark list are intended to serve as starting points for fur-
`
`172
`
`Figure 5: Path as extended bookmark list
`
`ther exploration—multiple stops within a particular neigh-
`borhood are infrequent.
`
`Extended tours resemble the narrative that one might get
`from an audio guide while walking through a museum. Here
`the role of the annotation is to contextualize the individ-
`ual stops within the larger framework of the collection as a
`whole. Figure 6 show an entry taken from this kind of path.
`In these paths multiple stops within a neighborhood are com-
`mon as the extended tour seeks to point out the key locations
`explicitly in order to guarantee that they have all been seen.
`
`The curriculum-based exposition both contextualizes the
`material it organizes but also provides an interpretation. The
`structures seen here have been more complex than those of
`the other forms. Annotations are used for curricular goals—
`for example questions may be posed to attempt to retain stu-
`dent attention (Figure 7). Introductory and interpretive ma-
`terial is often added as preface and delimiter for path seg-
`ments. Figure 8 shows one such example that is entirely
`annotation—this stop along the path includes no reference to
`WWW page. Path authors often wish to modify distracting
`backgrounds or to “rework” inappropriately-focused source
`material. Figure 9 shows an example in which the size of the
`original page’s textual material has been minimized and fig-
`ure 10 an example where the material included is illustration
`but not the original surrounding text.
`
`The intellectual contribution of the Walden’s Path author is
`the path and the annotation. However, curricular goals some-
`times require modification or interpretation of others’ intel-
`lectual contributions. Such adaptations are commonplace
`and accepted in print form (e.g., in textbooks). However,
`pages contained on a path are images of source material, not
`
`6 of 10
`
`

`

`Figure 6: Path as extended tour
`
`Figure 8: “Preface” material
`
`Figure 7: Path annotation in support of student en-
`gagement
`
`Figure 9: Minimized content
`
`173
`
`7 of 10
`
`

`

`ful lifetime of paths. Taken as a whole, material on the Web
`provides a dynamically changing collection of material—
`both content but also structure are undergoing continuous
`change. The environment on which paths are built rep-
`resents a snapshot of the Web, frozen at the time of a
`path’s creation. Without attention, the path-defined meta-
`structure rapidly loses synchrony with the external world’s
`dynamically-defined Web structure.
`
`We take two approaches towards limiting these entropy-like
`effects: detecting change and controlling the time of change.
`Detection of change depends on discovery of modifications
`to the underlying substrate (e.g., changes to addresses and
`to referenced pages) and initiation of corrective action. At
`present these activities are carried out manually, but we are
`in the process of implementing path validaters that verify the
`correspondence between the current environment and the en-
`vironment that existed when the path was written. Control of
`change centers around “freezing” the state of the path and its
`contents at well-defined points in time. Towards this end, we
`have implemented and use a “path compiler.” This allows us
`to give demonstrations without concern about changes to the
`substrate or to the network environment. Once again, how-
`ever, a potential conflict exists between path compiler and
`content author since some Web pages are inherently dynamic
`(e.g., real-time or customized information) and compilation
`violates that author-desired characteristic.
`
`7 EXTENSION AND CURRENT DEVELOPMENTS
`The Walden’s Paths system continues to evolve, driven by
`our user’s requirements and needs. This section describes
`some of the mechanisms that are currently being prototyped,
`relating them to the issues discussed in the previous section.
`
`7.1 Attribution and separation of content from annota-
`tion
`In the previous section we discussed attribution of content
`from the viewpoint of the content author. As accurate at-
`tribution is a keystone of academic life, it is not surprising
`that it is also a goal of the teachers who are prospective path
`authors. Similarly, clear separation of annotation and origi-
`nal content also is desired by teachers—while the content au-
`thor may desire this because it avoids confusion with other’s
`words, teachers also benefit because it more clearly identifies
`their comments to their students.
`
`Our ability to unambiguously separate annotation from con-
`tent inherently is compromised by our system’s architecture.
`The architecture’s advantage of using the client-supplied
`user interface also requires that the annotation share display
`space with original content. While we can use graphical con-
`ventions to reduce confusion (and we continue to refine our
`designs to accomplish this), the possibility remains that some
`referenced page could be designed in a way to cause confu-
`sion.
`
`Figure 10: Partial content
`
`reformulations cast into uniform format, and hence disagree-
`ments are possible about appropriate and fair use of the ma-
`terial.
`
`Discussions centered around transclusion [9, 11] suggest that
`acknowledgment and renumeration of original authors is an
`appropriate management model, but our experiences in the
`Digital Libraries domain [3] suggest that such acceptance
`will not be sufficient while transition is taking place. Re-
`cently botanists with whom we are working in the digital li-
`brary domain encountered a sticky problem of reuse of some
`of their botanical images. The original source of the reused
`images was appropriately attributed. However the surround-
`ing text incorrectly identified the species represented in the
`picture. The botanists were unhappy with this situation, even
`with accessibility of the original identifications and “renu-
`meration” in the form of academic visibility, as they felt that
`the graphic juxtaposition of image and text was sufficient to
`cause confusion in the reader’s mind as to who made the in-
`correct identification. In a sense, the form is as much a part
`of the intellectual contribution as is the content.
`
`We encounter a combination, consequently, of competing
`and conflicting user requirements in developing specifica-
`tions for Walden’s Paths prototypes. Curriculum-based ex-
`position benefits from extensive and flexible mechanisms to
`alter the appearance and content of path pages. Content au-
`thors, on the other hand, are often most comfortable with dis-
`play mechanisms that fix both the content but also its context
`and appearance (consider, for example, the use of Adobe’s
`Acrobat language for document disse