1. Discussing the question of closure in hypertext fiction, George Landow wrote in Hypertext 2.0 (1997) that "[u]nlike texts in manuscript or print, those in hypertext apparently can continue indefinitely, perhaps infinitely" (191). Since the publication of Hypertext 2.0, experimental hypermedia has brought to the fore a new variety of open-ended literary works that both challenge and extend the typically link/node and word-based hypertexts to which Landow refers, such as Michael Joyce's seminal hypertext story, Afternoon. (For online examples of word-based hypertext, see Joyce's Twelve Blue and Judy Malloy's LOve One.) A general survey of activities over the last four years reveals prominent changes not only in the way literary materials are composed but also in the tools of their distribution. Networked personal computers, coupled with advanced Web design software, provide a relatively cheap and easy mode of production and distribution not widely available before. As distribution machines, networked computers clearly change the relationship between author and reading public, most obviously in terms of the speed and range of distribution. Furthermore, new programming interfaces offer a whole host of gadgets--including animation, streaming video, vector motion, cascading styles, layering, and interactive behaviors--that together comprise some of the latest compositional tools of today's screen-based writing. When put to use on the digital page, these devices alter the time of literary performance in ways significantly different from print-based, or even first-generation hypertextual, writings. Duration (scene progression, sequencing, real-time motion) is now built into the metalanguage of literary composition as a device, along with more conventional devices like line, paragraph, prosody, character, and plot. Moreover, the primary locale for this new performance, the World Wide Web, provides a zone of perpetual currency, or fleeting stability, or both (depending upon one's perspective), which challenges conventional notions of the "past" and "present" of literary activity, in terms of the creative process as well as the distribution of a finished literary product. Three questions thus arise that will be treated throughout this essay: First, what are some of the ways in which computer technologies are currently used to create and distribute a time-based, hypermedia writing (with time-based defined for this study as hypermedia works whose "play" on the screen, either in whole or in part, is encoded into the work and computer-driven)? Second, how can time-based literary works of this kind be read in relation to traditional reading practices? Third, given the ephemeral nature of Web-based hypermedia, how might literary criticism in general accommodate this evolving art form?
   
   
   Hypermedia and the Evolution of High-Tech Text
   
   
2. Experiments with visual form in the language arts, of which hypermedia is the latest example, have a long and recently well-charted history (see, for example, Dick Higgins' Pattern Poetry: Guide to an Unknown Literature). In The Visible Word, Johanna Drucker surveys in particular the typographical experiments of the early Dada and Futurist writers, such as those of F. T. Marinetti, whose deployment of marked typography in poetry, she argues, owes much to a contemporary explosion of typographical play in commercial advertising. According to Drucker, by the end of the nineteenth century,
   

   [t]he advertising realm had burgeoned in response to the demands of the mass market strategies of industrial production. The result was an increased interest in graphics as an interface between producer and consumer to maximize the potential of a medium which had undergone, by contrast, comparatively little change in the three centuries since its invention. (94)

3. Dada and Futurist artists exploited this phenomenon in two ways:
   

   [They] were aware of the place the particular visual properties of type, layout and graphic design had in the social realm of public language and saw that they had become sufficiently codified and organized so that they could be manipulated. They were also aware that the distinct separation of the two typographic domains, the public/commercial and the literary, made the appropriation of these publicity techniques to literary works an activity which was subversive to the visual codes on which the authority of the literary text had been established. (102-103)

   The early mission of these artists involved, in part, then, the conscious manipulation of typographic norms already dominant in the commercial realm. As Richard Lanham has argued, electronic text in general represents "the perfect fulfillment of the Italian Futurists' desire to abolish the book in favor of a more dynamic medium" (x-xi). More specifically, current experiments in Web-based writing, following the Futurist lead, enact a similar appropriation of commercial scripting tools and techniques. As with the Futurists, this appropriation marks the continuing effort to destabilize the book, if not "abolish" it, as the central organizational medium for literary practice.
   


4. Miekal And's "after Emmett" provides an interesting example of such aforementioned Web-based typographic manipulation. In this animated elegy for visual poet Emmett Williams, assembled sequences of rapidly changing font characters invoke the popular use of the animated gif in the heyday of Web advertising. Likewise, the poem playfully acknowledges both Emmett Williams' (and Miekal And's) debt to the typographic experiments of the Dada and Futurist poets. In the same way that a Futurist collage might both mimic and frustrate the conventions of typography, the animated performance of And's poem aggravates and potentially undermines the "readability" of the message. Many of the major tools now used in this brand of literary gaming--animated gifs, java and javascript, HTML editing packages like Macromedia's Dreamweaver, Microsoft FrontPage, multimedia authoring and imaging software like Director, Flash, Freehand, and Adobe Illustrator--exist primarily in the service of (and to perpetuate) commercial, particularly advertising and promotional, interests. A burgeoning aesthetic among Web artists recognizes these current "publicity techniques" as exploitable via the same reasoning which undergirds the Futurist appropriation of typography.
   
   
5. More important to this study are the specific artistic uses to which Web design technologies can be put for the purposes of literary production. As a compositional device in time-based hypermedia writing, the animated gif was the first tool to change the way text can be read on the computer screen. Several Web writers have used gif animation to morph alphabetic characters or whole words and/or generate small-scale movements or displacements on the virtual page. And's "after Emmett," linked above, as well as Annie Abraham's "understanding," are two good examples. With more recent tools, such as Dynamic HTML (Javascript) and those providing "streaming" capabilities (most notably Shockwave, Flash, and Quicktime), practitioners of time-based writing can populate the browser window with a practically endless array of movements, morphs, and other dynamic behaviors. Jim Andrew's "Enigma n," for instance, choreographs simple animation and primitive user interaction in what amounts to a compelling "demo" of late 1990's Dynamic HTML scripting. Reshuffling letters in the word "meaning," Andrews uses Dynamic HTML to aggravate the static letter-objects ("m-e-a-n-i-n-g") encountered when one accesses the page. Initially the user has three options: "Prod," "Stir," and "Tame." When selected, each option yields a different manipulation of the letters; in each case, though, the letters are nudged on a click into asynchronous circular and elliptical orbits. After all three options have been selected in any order, the option to "Spell" appears. If clicked, "Spell" brings the letters back to rest at the center of the screen. But it's a false ending, since cycling through any of the initial options eventually introduces further options: "0/1," which freezes the action at any given point; "Colour" and "Discombobulate," which render alterations in color schemes and font sizes respectively; "Speed," which yields a pop-up menu to the left of the screen with a series of selections either slowing or quickening the revolutions of letter-objects; and finally, "About" and "Runaway," the latter of which links out to one installment of the peripherally engaging poem "ADVXES," by Ted Warnell. Thus the options for play and replay in this poem make it difficult to determine its spatial and temporal boundaries. As with nodal hypertext, the time of performance depends strictly upon the user's selections and habits of navigation when faced with options to manipulate the surface of the text.
   
   
6. Other works, utilizing vector animation software (such as Macromedia Flash), integrate sound as well into the performance. Alicia Bonadonna's "Lucinda" foregoes text entirely; instead, interlaced voices "recite" a short poetic narrative centering on the "not quite pretty" Lucinda, while, in the foreground, a kaleidoscopic array of objects dances across the screen. The performance loops after a sequence of about 30 seconds, suggesting that this piece is temporally open-ended. However, nothing changes in a repetition of the sequence, suggesting a work of finite duration that simply repeats, like a track on a CD or a poem read more than once.
   
   
7. Fundamental to the task of defining (and "reading" in any conventional sense) Web-based hypermedia installations, such as "after Emmett," "Enigma n," "understanding," and "Lucinda," is the challenge of assigning priority to any one of their compositional elements: Are they poems, stories, or kinetic sculptures? Are they pieces of art or works of literature? Are they temporally finite or open-ended? Clearly, the integration, or intermediation, of text, spoken word, animation, image, and sound in current Web-based experiments marks one of the more compelling challenges for contemporary literary practice and criticism. In this integration, we find the legacies of the international visual-concrete poetry movement (see UBUWEB for more on the history of visual and concrete poetry), as well as the cross-genre performance events popularized by the Fluxus Group (c.1960). We also see traces of the real and virtual kineticism of the Kinetic Art movement of the Fifties and Sixties, plus more recent experiments (by performance artist Eduardo Kac in particular) in holographic and video poetries. Because current computer-based practice owes allegiance to these early informants, as well as to film, video, and broadcast techniques, the task of constructing a poetics for what is obviously a cross-genre experiment becomes especially difficult. At this point, I want to use the term "writing" to describe these efforts for reasons that I hope will become clear in the latter portions of this essay.
   
   
   Reading Time



8. In "Time: The Final Frontier," hypertext author Robert Kendall adjusts the traditional recipe for hypertextual production: "Time is the element that must be added to the raw configuration of nodes and links to produce a textual realization--a finished structure." Kendall's focus is hypertext fiction, in particular a reader's linear engagement (through time) with an electronic fiction that is otherwise radially propagating (in space), structurally polylinear, and conceivably endless. Since all reading is temporal, Kendall suggests, the decisions a reader makes in time--most notably where and when to begin, how much and how long to read (and reread visited nodes), and when to close out the hypertext window and call the experience finished--must be figured into the overarching equation by which "textual realization" or structural completion is defined. Kendall argues more broadly in "The Final Frontier" that our descriptions of hypertext tend to favor spatial, topological metaphors. As Kendall notes, we "endow hypertext with a virtual physicality via metaphors of pages, paths, and webs. We speak of visiting Web sites. We like to simulate topographies for our hypertexts via maps" ("Frontier"). It seems clear to me that our habit of describing electronic literary texts in terms of a delimited and mappable space marks, among other things, a nostalgia for the static page and two-dimensional planar representation. However, description of just about any current Web installation in terms strictly related to spatial distribution would overlook the time-related factors that also contribute to the literary experience. At this point, literary criticism must modify its descriptions to account for these factors, some of which I would like to address here.
   
   
9. Taken broadly, digital writing, with the help of high-speed computing, has distilled the device from the murky waters of theory, offering at least one prominent example of what Lanham called "the extraordinary convergence of twentieth-century thinking with the digital means that now give it expression" (51). Just as hypertext has made use of the hyperlink to formalize intertextuality, hypermedia uses the timeline and other motion devices to compose temporality in the literary installation. This shift from space to time is tantamount to an aesthetic shift from mapping, and radial structures, to happening, morphosis, and temporal experiences. As Kurt Brereton suggests, the poem has shifted from a flat, constructed surface to "a virtual field unfolding in time." More generally, the kinetic hypermedia installation looks a little more like theater, film, or video than writing per se--and again this points to the intermedial nature of the work. One example is the collaborative project "~~Water~~Water~~Water~~" (by Christy Sheffield Sanford and Reiner Strasser), which integrates Javascript and Flash in a multi-layered performance of visual textures, sounds, images, and cascading pop-up windows. The attempt to literalize "flow" in this piece, handled consciously and skillfully by its makers, invites the reader/viewer to "give way(s)," as Michael Joyce has phrased it, to a text "caressed into motion or repose" by its user (187). With motion built directly into the architecture of the piece, the time of reading comes alive as an agent of literary realization. For artists working in this medium, the timeline asserts itself as a literary device functional in ways similar to such conventional devices as rhyme, plot, character, and poetic line.
   
   
10. It makes sense, furthermore, to discuss these new works in the context of a "textual" practice. Creative hypermedia installations combine at least two (for now) modes of sense experience: sight and sound. What gets seen, of course, can be divided, as it often is, into several functioning modes: text, static image, moving image, video, and perhaps some amalgam of two or more. Such weaving of sensory input, however, operates squarely in the realm of textual practice. "Text," as Roland Barthes writes in The Pleasure of the Text, "means Tissue…worked out in a perpetual weaving" (64). Elsewhere, Barthes calls attention to the "stereographic plurality" of the text and its effect on readerly perception: "what [the reader] perceives is multiple, irreducible, coming from a disconnected, heterogeneous variety of substances and perspectives" (Image – Music – Text 159). Etymologically, the term "text" suggests a broader potential for intermedial work than its popular use to mean "written material" might imply. The word "text" dates back to the Indo-Euro "tek-" ("to make"), which is the source of the English "technical." In Latin, the past participle "textus" was used as a noun meaning "woven material," and only metaphorically to refer to a literary composition--that is, letters woven together (Ayo 526). Thus, the word "text" in its primary usage is a fitting term for today's post-textual (that is, beyond the printed page) but very much texted (woven/built/made) hypermedia or high-tech efforts. To speak of hypermedia "writing" in these broader terms, therefore, seems appropriate to me as well.
    
    
11. In any event, while a hypermedia installation such as "~~Water~~" makes use of traditional text (words and letters), the work also asks to be read as a literary document via selective strategies spanning at least two domains: that of conventional poetry or prose, and that of concrete or visual poetry. In either case, the installation orchestrates what I see as a conflation of two different kinds of duration: the duration invoked by a linear, interactive reading (left to right and/or top to bottom), plus the real-time duration of streaming content spanning a finite number of real seconds or minutes. In these modes, a hypermedia installation asks to be both watched--as something akin to theater, film or video--and simultaneously read. Hypermedia thus realizes the computer's potential as "a device of intrinsic dramaticality" (Lanham 6)
    
    
12. As these examples have shown, literary hypermedia invites a different response from the reader accustomed to static, black and white pages. But the illuminated page is not without its predecessors. Where text approaches image we are asked to consider the two in relation, as we have over the centuries with a wide body of subgenres, including Medieval illuminated manuscripts, the pattern poetry of George Herbert, William Blake's plates and etchings, Apollinaire's Calligrammes, and more recently the typographical experiments alluded to above. Now, where text approaches kinesis either as image or alongside image, we are asked to consider the two in relation in motion, as well as in relation to motion. The fun starts when motion begins to distort, modify, or revise the given relation. In practice, hypermedia writing invites different readings at different moments--not to mention potentially whole new readings on different days, in different performance environments (on different systems or browsers, for example), with different speeds of download, different entrance and exit points, and different navigation choices. Of course, a conventional page reading offers similarly variant experiences, since all readings (like all writings) can be viewed as contextualized and temporal. Likewise, first-wave hypertext has made the "multiple reading" its central selling point. However, hypermedia writing takes it one step further by reifying (then manipulating) an otherwise uncontrolled variable, namely, the duration of the literary performance.
    
    
13. Thus the preliminary conclusion for literary criticism is clear. In addition to the two famous critical questions "What does it mean?" and "How does it work?", we must now ask, "What happens as it works?". That is to say, we can now look more closely at how the sub-languages of computer coding provide opportunities for visual-textual pacing (real-time rhythms) that writers until now have not been able to control. To keep up with this new writerly device of time-based programming, to register what happens in relation to what exists, and then to attempt a reading or critique of this relation, the hypermedia reader must "watch" with a particularly broad focus in mind. With an unfixed or unfixable content, the reader cannot describe the boundaries of line, stanza, paragraph, or chapter, let alone the boundaries of text, image, page and book, on which traditional readings have depended for much of their authority and accuracy. Hypermedia writing invites the reader to witness the unfolding or streaming of a screen content. The reader, then, tunes into this unfolding with the eye and ear of an observer. In an odd twist, hypermedia texts in which the timeline serves as a compositional device actually remove an element of interactivity and user participation by which nodal hypertext has distinguished itself. In nodal hypertext, the user in most cases determines screen changes by clicking hyperlinked elements. In time-based hypermedia, the text might move (between or among frames) without any user interaction (mouse clicks) at all. The implied conflict here between user activity and user passivity, while outside the bounds of this essay, nonetheless calls for further study.
    
    
14. Time-based hypermedia, particularly on the Web, asks the reader to be in attendance for an event made radically public by its existence in a network. Focused awareness of proximal engagement in the duration of a publicly distributed work marks at least one major aspect of an effective reading strategy, as well as one major difference between reading Web-based hypermedia and reading pages in a book. In other words, the Web offers space for a public performances in the same way cafés, salons, and college lecture halls have provided space for literary readings. The obvious difference, however, is that Web works come to us locally, funneling the time of performance into our own personal (PC) spaces. Obvious also is the greater ease with which writers can seek exposure via the free-roaming channels of the Internet. Distribution via the Web, however, complicates the time of this performance in ways I will address in the next section.
    
    
    E-lastic Publishing and Critical Glimpses
    
    
15. It has become a cliché that the Web is a constantly changing and somewhat unsteady environment. Users often struggle with servers that are down, removed and relocated sites, and other modifications, some intentional and some not, to the fabric of the Internet. The Web can be understood, then, as a zone of perpetual currency or fleeting stability in which new instantiations constantly displace the old. More importantly, as the technology changes, and with it the skills of today's hypermedia writers, it has become common to overwrite old work with work revised or adapted using the new tools, rendering obsolete any notion of co-existing editions. As Lanham argues, "[t]o volatilize text is to abolish the fixed 'edition' of the great work and so the authority of the great work itself" (xi). We now have, replacing the great work, the text that exists "in potentia, as what it can become, in the genetic structures it can build. It is volatile both in how it is projected onto an electronic screen and in how it works in the world." (Lanham 19).
    
    
16. One effect of the Web--and thus one way in which hypermedia "works in the world"--is that it seems to have blurred distinctions between two conventionally separate literary activities, namely writing and publishing. Whereas in the print world composition is typically considered a pre-publication activity, with nearly instantaneous Web postings the writing process becomes more closely aligned with the processes of publishing. Web-based writers might post initial versions of work to discussion lists and then incorporate reactions and suggestions into future instantiations. Obviously this is similar to the practice of distributing a draft of a story or poem to colleagues and friends, but the difference lies in the marked absence of a publishing date fixing the end of the writing process and, by extension, the beginning of the reading process. As argued above, the continuous overwriting of new versions enters the fabric of digital production in a way that it can't in the print-based paradigm. In Web-based literature, therefore, the once-discrete acts of composition and publication resituate themselves in concerted acts of witnessing, and the witnessed act of literary performance assumes the contours of its time (i.e., its given version at a given time) as a book would the shape and texture of its binding cover or chosen typeface. We can talk of different experiences with a given installation over time, but to talk of it in terms of "pre" (composition) and "post" (publication) activities makes little sense in a distribution environment constantly eluding such distinctions.
    
    
17. Therefore, the reader of Web-based hypermedia is asked to consider the time of a literary work in still another sense: When (as well as where) should we look to find the work? First of all, the surfaces of hypermedia writing are by nature fluid and ephemeral, giving way(s) to other surfaces as either (a) the scripted timeline progresses, or (b) the user makes a selection. Moreover, a perpetually revised and updated Web corpus makes it difficult for a critic to frame this body of work for close study. The art of Web-based textuality thus emerges as an art of the ephemeral, much closer to performance (like live music, theater, or even the low-tech and, to some, anti-technological "spoken word" reading) than to traditional literary publications. As George Landow has pointed out, echoing Lanham above, the infusion of the literary corpus with such a wide and diverse range of hypermediated literary productions destabilizes not only the authority of the "great work" but also the "mastery" of the literary critic. As Landow claims, "[t]he critic has to give up not only the idea of mastery but also that of a single text at all as the mastery and mastered object disappear" ("What's a Critic To Do?" 35). In reading today's hypermedia literature, he continues, the critic "becomes more like the scientist, who admits that his or her conclusions take the form, inevitably, of mere samples" (35).
    
    
18. Generalizing this destabilizing effect to the reader at large, Espen J. Aarseth remarks that, for readers accustomed to linear texts, the non-linear text "cannot be read, only glimpsed and guessed at" (65). A poetics of hypermedia writing, particularly that performed on the Web, is a poetics, therefore, of collected "samples" and "glimpses." To read the new hypermedia literature, we cannot rely on the conventional critical standpoint grounded in distance from, and hence "mastery" over, a clearly delineated text. In fact, as literary documents become less fixed, more dynamic, and more immersive, they begin to look less like "poems" and "stories" and more like virtual or "artificial reality" environments, such as those described by Myron Kreuger more than a decade ago. In such environments, with participants processing different feedback, the very notion of "interpretation" or "analysis" becomes suspect:

    An artificial reality can take steps to individualize responses and to thwart analysis. If each person has a different experience, each will experience less pressure to arrive at the "right" interpretation. Since each person moves about the space differently, each will receive different feedback, even if the controlling program is exactly the same. (Krueger 84)

19. Krueger's description of physical navigation, while specific to three-dimensional artifical spaces, applies to two-dimensional screen immersion as well. The habits of "'right' interpretation" common to conventional literary exegesis make little sense in the domain of hypermedia. Or, as Lanham phrases it, "[a]n ever-varying chameleon text forever eludes definitive explanation" (7). Granted, literary theory, particularly as it manifests itself in post-structuralism, has never argued strictly for "right" interpretation. As implied above, even page-bound readings take place in time and are thus dynamic and open to all interpretation, not just "right" interpretation. Reversing Aarseth's statement above, in other words, we could say that linear texts as well "cannot be read, only glimpsed and guessed at." It might be more appropriate to conclude, then, that readerly responses to artificial reality environments or screen-based hypermedia do not "thwart analysis" so much as reinforce what we've learned recently about analysis in general--that it is time- and context-dependent and bound inextricably to the experiences and assumptions of the analyst, depending on the "different feedback" received at the time of reading or experiencing the text. The difference, of course, between the "static" and the "dynamic" text is that the latter cannot be pinned down, literally and figuratively, since its materials (moving pixels) resist efforts to do so. Extending Landow's insightful claim, critics must therefore operate scientifically (thus experientially) in relation to an ephemeral hypermedia literature, engaging it as witnesses to a series of performances, rather than as custodians of a discrete literary product. The job of the literary critic thus becomes more like that of the theater critic, whose sense of the "whole" is mitigated by the mutability of a time-based performative genre.
    
    
20. Running counter to my argument for Web ephemerality, however, is the fact that digital storage is increasingly cheaper, more stable, and easier to obtain (free disk space at My Space being just one of many examples). Many hypermedia installations do, in fact, reside unchanged on the Web for several years, as evidenced by most of the examples linked in this essay. Their stability over time has made possible their performance here. Still, even the vicissitudes of browser upgrades can sometimes render pre-existing work literally unreadable. According to private testimonies I've received from Web artist colleagues, for example, the introduction a few months ago of Netscape 6.0 rendered obsolete several Dynamic HTML features on which many of their Web installations depended. (So far, no word yet on whether version 6.1 has fixed the problem.) Likewise, there are no guarantees that global platform standards, plug-ins, and media extensions will remain static or, if changed, friendly to older versions. To work in a digital medium, particularly over the World Wide Web, is clearly to embrace the dual sense of disk immortality and design ephemerality. The work can conceivably "last and last" as a chunk of kilobytes on a host server, but the work also outmodes itself perpetually as internal and external changes force modifications to what stands for the literary document at any given time.
    
    
21. Voluntary modification on the Web is one such "internal" change and can be traced, in part, to upgrade hysteria and, more generally, the "tear-it-down, build-it-up" ethos of a Western mindset now gifted with the tools to do in cyberspace what we have done so well in physical space. The Web, after all, is sold to us as a "clean" environment, albeit virtual, whose streets and gutters require no sweeping because they're repaved, if not utterly replaced, on a near-daily basis. I'd like to suggest again that the fetish of freshness (note the use of "refresh" in Internet Explorer to label the browser's self-updating capability) likewise represents one of the "publicity techniques" that today's Web artists might want to take on, directly and indirectly, as points of departure. Sites like jodi.org have paved the way for this kind of perpetual Web remediation, turning the culture of hysteria back onto itself in a subtly self-conscious parody of the major Web pathologies (such as endless software upgrades and updates, browser wars, disappearing Web sites, and ubiquitous banks of meaningless computer code). A visit to the "jodi" page might bring up a mock "404" message telling you that the site is unavailable or, perhaps, momentarily out of the office. Other acts of Web remediation (by which I mean the distribution of alternative content via the latest tools and gimmicks of pop Web culture) might involve a similar process of software appropriation and redirection.
    
    
22. "Versioning" (voluntary modification or overwriting the old with the new, often via new software tools) is thus another way in which Web works change over time. Critical investigation of the literary corpus on the Web might therefore focus less on when a particular work came on the scene (for example, its date of publication, since, as I argue above, such calendar dates, even when available, mean little on the Web), and more on what version it serves to demonstrate. In other words, particular works could be time-stamped for generational distinctions or software versions. In such a schema, Miekal And's work excerpted above, for example, would be called "gif generation" hypermedia, whereas Alicia Bonadonna's piece might be labeled "Flash generation." Given what are often radical differences between software versions, a "Flash 3" designation might mean something entirely different to the reader than a "Flash 5." Obviously the terms used to mark such distinctions require further consideration and refinement beyond these somewhat pedestrian placeholders. Also, the extent to which such "generational stamps" might underwrite more serious concerns involving software availability and cost, as well as the attendant class, economic, and educational divides, deserves further study beyond the scope of this essay. But a clearly defined set of designations would, I think, prove beneficial to a comprehensive hypermedia poetics. The introduction of such designations would, I hope, be approached collaboratively and cautiously, in an effort to avoid potential over-classification and misidentification. In short, should such a scheme prove beneficial to literary study, the works and their chosen media should determine the terms by which the discussion of works proceeds.
    
    
    Conclusion
    
    
23. Reading "time" in literary hypermedia requires that we read in time, that we get to the work in time, and that we use our time wisely in navigating its surfaces. These are more than just metaphors, since erasure and modification on the Web often mitigate, sometimes eliminate, opportunities for studying and understanding this new and vigorous body of work. A new poetics of hypermedia writing therefore begins with a clear sense of critical presence in relation to the work at hand. The historical freshness of the practice--that is, that the literary experiments we now witness mirror the rapid development of a technology (personal network computing) itself in its infancy--makes developing such a poetics a difficult challenge indeed. To many of us, writers and readers alike, the task of engaging this work seems daunting if for no other reason than the sharp, upward slope of the learning curve now before us. It seems helpful to the venture, therefore, that we situate current experiments in the context of a larger literary tradition, particularly that tradition of literary experimentation for which the play between and among media has been central to the creative act. Hypermedia installations are textual assemblies in the most fundamental sense. Their textures are likewise both familiar and strange, innovative and recursive. Understanding the time of literary hypermedia, in terms of both the coded time of the document and the time of its performance on an otherwise spatially distributed World Wide Web, is key to understanding what makes it simultaneously strange and familiar to today's readers.

Works Cited