Benchmark DAC Proposal 2
From WRTwiki
Benchmark Fiction:
A Framework for Comparative New Media Studies
Christy Dena
University of Melbourne
School of Creative Arts
Parkville, Australia
Jeremy Douglass
UC Santa Barbara
2607 South Hall, UCSB
Santa Barbara, CA 93106-3170 USA
Mark Marino
UC Riverside
8325 Fordham Rd.
LA, CA 90045 USA
Abstract
How do we compare eliterature forms? The Benchmark Fiction Project began as a proposal to adapt the same short story into many elit forms (hypertext, IF, wikifiction, etc.) with the goal of creating a set of elit editions for critical and pedagogical purposes. This quixotic process of ?benchmarking? highlights issues of adaption and remediation, reveals form/content assumptions, and facilates understanding of emerging elit types for both the designers and experiencers of the works created. Work on the first story to be benchmarked, ?The Lady or the Tiger? (1882) by Frank R. Stockton, inspired a lightweight software framework for displaying source texts through interchangeable display modules. The resultant ?benchfics? provide lessons in adaptation in a variety of digital forms. An overview of the project, theoretical background, and practical design decisions are described here.
Categories and Subject Descriptors
H.5.4 [Information Interfaces and Presentation]: Hypertext / Hypermedia ? architectures.
General Terms
Design, Experimentation, Theory
Keywords
benchmark fiction, benchfic, eliterature, adaptation, crossmedia, chatbot fiction, hypertext fiction, interactive fiction, wikifiction, media specific analysis, remediation, transmedial worlds, transmedial narrative
1. INTRODUCTION
How are the multitudinous forms of eliterature similar, and how are they different? How do we critically compare these forms?
A benchmark fiction, or ?benchfic,? is an elit adaptation for the purposes of comparing the underlying media forms. The term ?benchmark? here is playfully repurposed from the fields of computer science and strategic management in order to emphasize the focus on utility and standards. While ?benchmark? originated as a surveying term for a point of reference, in contemporary computer science, ?benchmarking? has come to mean the execution of a software test in order to ascertain the relative performance of underlying hardware.
The production and comparison of benchfic is analogous to this process of testing hardware via software. With benchfic, the soft ?content? of a story might be ?run? on the hard ?form? of two different systems of elit implementation in order to examine differences in those two specific forms. Thus, Frank R. Stockton?s short story ?The Lady, or the Tiger? might be adapted as a Storyspace hypertext fiction as well as an Inform interactive fiction. Alternately, the content of a work of elit (e.g. ?Afternoon: a story? by Michael Joyce?) might be separated from that form (Storyspace) and then re-implemented on some other form (e.g. HTML). A benchfic compares, therefore, on three levels: between elit types, between software implementations, and between media channels.
The central method of producing benchfic ? the separation of ?form? from ?content? ? is highly problematic. It is difficult or impossible to pinpoint where form ends and content begins in a given work, particularly artistic work designed to be experienced as a unified whole. Yet, with benchfic as with other processes of adaptation, translation and remediation, the problem of determining which elements to hold constant and which to vary is in large part the value of the undertaking.
In the process of producing benchfic, one?s concept of ?form? is formalized. These formalizations may have no greater power than to make explicit (and thus reveal) their own limitations. There is no one correct approach ? rather, the Benchmark Fiction Project proposes the ongoing aggregation of a multitude of parallel elit adaptations, each with their own claim to what in translation studies is termed ?equivalence.? Rather than creating simple assumptions about how elit forms operate, benchmark fiction creates an opportunity to critically examine the assumptions and arguments we already make.
2. BENCHFIC IN THEORY
2.1 Benchmarking Elit
The comparison of adaptations is common both in criticism and in the classroom ? for example, considering the text of a Arthur C. Clarke?s ?2001: A Space Odyssey? side-by-side with the Kubrick film.
By contrast to this readerly form of comparison, the more writerly comparison of creative adaptation is not a common practice. This is especially unfortunate as it coincides with the areas such as new media which are particularly characterized by astounding variation, formal innovation, and multiplicity. Here first-time authoring and first-time reading in new forms is the norm, and claims about the best / latest / smartest / freest medium are common. Yet, in precisely the area of culture where authors and readers both need some sense of the comparative capacities of digital art forms, actual examples of broad adaptation sets (such ?The Hitchhiker?s Guide to the Galaxy? by Douglas Adams) are significantly less common.
That analysis of adaptations is enlightening seems self-evident from centuries of critical tradition. Yet how practical is the production of elit adaptations as a critical practice ? and is the real potential for pragmatic lessons learned? The Benchmark Fiction Project directly addresses the debates over the transmedial nature of narrative and the ludic qualities of a work. On the former, we consider there to be both transmedial and media-specific qualities of the selected text to be benchmarked, and seek to explore their difference. In this regard, our research hypothesis is akin to narratologist David Herman?s:
Although narratives in different media exploit a common stock of narrative design principles, they exploit them in different, media-specific ways, or, rather, in a certain range of ways determined by the properties of each medium [7].
We extend this hypothesis to include ludic design principles. The aims, therefore, are to investigate the properties of the medium that ?determine? the narrative and ludic features, and the nature of the effect. This investigation is both process and product oriented. First, through the process of attempting adaptation into a target elit form, we hope to identify different writing practices that emerge while trying to render similar effects in different new media. Second, we hope to generate example products which allow critical communities more opportunities to make apples-to-apples comparisons of new media fiction implementations ? similar matter can be pulled up in two different applications and read closely side by side.
Ultimately, these experiments may prove more interpretive than empirical, in that we will not constrain ourselves to attempting pedantic remediations of the text into each form. In order to accentuate the strengths and limitations of each form, the story will be adapted. The results then will reflect particular readings of the source text as well as interpretations of the nature of the forms themselves. The individual reader or interactor will apply their own heuristics to the benchfic in its various editions. Furthermore, the process creating a benchfic does not foreclose the possibility of creating supplemental or even counter-adaptations. There can always be more benchfic.
2.2 Benchfic and Cross-media Storytelling
When a set of benchfic are created, what is their critical relationship to each other? The entire work ? the sum of all the texts ? is a rendition of ?cross-media storytelling.? In 2001, media theorist Henry Jenkins observed the occurrence of increasing media channels and the introduction of stories that are delivered over multiple channels [8]. Through an analysis of ?The Matrix? franchise, Jenkins posited the notion of ?transmedia storytelling?:
In the ideal form of transmedia storytelling, each medium does what it does best?so that a story might be introduced in a film, expanded through television, novels, and comics, and its world might be explored and experienced through game play. Each franchise entry needs to be self-contained enough to enable autonomous consumption. That is, you don?t need to have seen the film to enjoy the game and vice-versa.
More recently, Jill Walker has explored the notion of ?distributed narrative? [18], while Lizbeth Klastrup and Susana Tosca have investigated the qualities of a work that can persist over multiple channels and time through the notion of ?transmedial worlds? [11]. All of these approaches require the ?text? to be viewed in a multi-text, and multi-channel environment, what Christy Dena conceives as ?polymorphic works: texts in many forms? [3]. Polymorphic works are an extension of literary theorist Itamar Even-Zohar?s ?polysystem studies? (PST). [4] During the early 1970s Even-Zohar adopted systems studies and specifically PST in order to assist with the complex socio-semiotic phenomena surrounding translations, and has been developing the approach for the last three decards. Semiotician Jay Lempke has also recognized the critical approaches needed to capture the peculiar instances of multi-text environments:
Not only do we not have adequate models of semiotic effects and inter-discursivity for each of these media individually, but many of the discursive and ideological effects of interest in inter-media franchises depend on inter-relations among presentations in coordinated, multiple semiotic media. [12]
Benchmarking enters this discourse as a creative and critical approach that recognizes the phenomenon of cross-media storytelling and assists in the development of new media by providing a concentrated collection of texts with various inter-text relations.
2.3 Benchfic and Media Specific Analysis
As adaptations, a benchfic must vary while retaining some equivalence with the source text. A set of benchfic are adapted in parallel from a single source, and thus, by extension, must both vary from and have some equivalence with each other. What is the relationship of these specific differences to this general commonality?
In ?Writing Machines,? N. Katherine Hayles describes Media Specific Analysis (MSA) as
a mode of critical interrogation alert to the ways in which the medium constructs the work and the work constructs the medium... MSA attends both to the specificity of form...and to citations and imitations of one medium in another. MSA moves from the language of text to a more precise vocabulary of screen and page, digital program and analogue interface, code and link, mutable image and durable mark, computer and book. [6]
Characteristically, Hayles emphasis here is on the specifics of materiality, rather than transcendent terms such as ?work? or ?text?.
Our analysis makes use of this more precise vocabulary, even while it reserves a register for the work which is transcendental. We propose that not only does the physical medium define the final product, but that the specific software has an influence over defining the parameters of all works that are produced using it. A simple example might be the use of Storyspace, the hypertext authoring and publishing software developed by Jay Bolter, John Smith, and Michael Joyce, and published by Eastgate Systems. Joyce recounts wanting to develop a system that would allow him to tell a story in a way that the codex book could not (personal interview). Later hypertexts, such as Shelley Jackson?s ?Patchwork Girl? (1995) explore the limits of that system, but remain expressions of the capacities and constraints present in the software package. Macromedia Flash presents another example. Flash movies vary widely in their characteristics, but they all exist as expressions of what Flash permits.
This view of elit authoring as existing within a certain scope of available formal experimentation recalls Lev Manovich?s claim that ?the greatest avant-garde film is software such as Final Cut Pro or After Effects? [13]. Such a piece of software, in Manovich?s terms, ?contains the possibilities to combining together thousands of separate tracks into a single movie, as well as setting various relationships between all these different tracks ? and thus it develops the avant-garde idea of film as an abstract visual score to its logical end.? We take this notion further, to explore the ways in which pieces of authoring software come to shape their fields, or perhaps, to use Deleuze and Guattari?s model, their plateaus of art objects.
At some point it becomes difficult to separate the software from the material systems that run them, and we do not want to gloss over the differences in user experiences based on the machines they use to view these systems. Media Specific Analysis warns against believing in an easy equivalence that floats above material difference. For our purposes, where we differ is in our greater emphasis on the idea of representation and structure in digital text (e.g. the HTML H1 tag) as opposed to the screen instance (e.g. 24pt Arial Bold). This balances the specificity of presentation with the specificity of the code.
2.4 Benchfic and Remediation
In "Remediation," discuss the Jay David Bolter and Richard Grusin discuss the way media oscillate between transparent immediacy and opaque hypermediacy. While mediation is ever-present, immediacy is a rhetoric that naturalizes the media frame and effaces it from view, while hypermediacy emphasizes the technical qualities of the medium and holds them forward for greater attention. The stances of natural being vs. skillful overcoming become an agonistic game, which is in particular played out as the New Media child tries to kill the Old Media father, and become him. In particular, their definition of a medium as "that which remediates" [2] reveals that individual media forms are understood through difference as they enter the mediasphere. In the case of elit media forms, wikifiction is for example that which is not hypertext fiction in specific ways - that these differences exist are certain, but what they are remains for us to elucidate.
It is worth mentioning that the call for Benchmark Fiction and the act of benchmarking is a move towards more opaque hypermediacy. Benchfic are not creations that are to be accepted on their own terms, with attention only to whatever natural inner logic that might entail. Rather, benchmarking is a comparative process, and when an act of adaptation evokes what in translation studies would be termed a 'target' text in terms of some 'source,' the rhetoric of derivation and constructedness de-naturalizes the target in the eyes of the viewer. The resulting benchfic becomes opaque and significant in formal feature to the viewer, rather than a transparent conduit of contents.
Benchmarking is in part a proposed practice of construction adaptations, and in part a proposed critical methodology of attending to them for the purposes of media comparison. Taken together, this construction and attention can be understood in Bolter and Gruisin?s terms as a rhetoric of hypermediacy.
3. BENCHFIC ANCESTRY
Has their been anything like benchfic before? As the emphasis of remediation on modes of reception reminds us, any set of works including an original and several adaptations might be considered benchfic when considered together for the purpose of comparing media ? regardless of the conditions or intentions that set was created.
While benchfic may be a mode of understanding, benchmarking can be described as an act and a creative process. What other practices or traditions have used creative variation for the purposes of evaluation? Two initial examples of such praxis arise from ancestors to eliterature ? specifically, the fields of graphic design and software engineering.
3.1 Lorem Ipsum
The conventional ?lorem ipsum? passage is a design tradition of using an invariant placeholder passage during mockups, layouts, and considerations of typography. Graphic designers pour this passage into sample layouts, creating the appearance of prose without introducing actual legibility to distract from evaluation of the formal design.
Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
In use since the 1500s, Lorem Ipsum has seen its share of varied designs. However, rather than interacting with those forms by constraining or being constrained by them, the purpose of Lorem Ipsum was to fill without filling, and to appear readable without being read. Contemporary widespread illiteracy in Latin only aids in the desired effect of a design whose appearance of legibility can be evaluated without the eyes being inadvertently drawn from the surface of the page into the meaning of the text. Interestingly, this led to a situation in which print culture became the steward of a text that was commonly known to be a nonsensical mish-mash of pseudo-Latin, illegible in the most general sense. In fact, this common knowledge was incorrect. Richard McClintock identified Lorem Ipsum as an excerpt from Cicero?s 45 BC treaty on ethics, ?de Finibus Bonorum et Malorum.?
Lorem ipsum is Latin, slightly jumbled, the remnants of a passage from Cicero's de Finibus 1.10. 32, which begins ?Neque porro quisquam est qui dolorem ipsum quia dolor sit amet, consectetur, adipisci velit....? (There is no one who loves pain itself, who seeks after it and wants to have it, simply because it is pain... ). What I find remarkable is that this text has been the industry's standard dummy text ever since some printer in the 1500s took a galley of type and scrambled it to make a type specimen book; it has survived not only four centuries of letter-by-letter resetting but even the leap into electronic typesetting, essentially unchanged except for an occasional ing or y thrown in. It's ironic that when the then-understood Latin was scrambled, it became as incomprehensible as Greek; the phrase 'it's Greek to me' and 'greeking' have common semantic roots! [13]
Throughout its perilous voyage down through the ages, Lorem Ipsum has highlighted the media specific qualities of a given design through a particular kind of invisibility or transparency. It achieved this through a combination of naturalized convention and customary illegibility. It is not structured content (with headings, footnotes, etc.) that interacts with the constraints of a particular design, but rather unstructured content which fills it. Rather than being read in difference contexts, it is that which is not read at all. As such it represents one limit case on what benchfic can be ? for acts of reading and varieties of reader-text interaction are being benchmarked, then, without reading, the formula of evaluating common content in disparate forms can have no meaning.
3.2 Hello, World!
In computer programming there exists a tradition of writing a minimal program, which simply outputs text to some display. By convention, this text is always ?Hello, world!,? and thus a vast catalog of Hello World programs (as they are called) have made their way into textbooks, manuals, and documentation for every imaginable computer language.
Each example is in fact two texts. The first, the output or result, is held invariant (?Hello, world!?), in order to highlight the various syntactical means whereby the various programming languages achieve the end of similar behavior using the second, the source code or cause. At present, numerous catalogs collect these program source codes side-by-side for mutual consideration ?the Wikipedia article ?Hello World Program? currently lists 153 and counting:
While small test programs existed since the development of programmable computers, the tradition of using the phrase "Hello world!" as the test message was influenced by an example program in the book The C Programming Language, by Brian Kernighan and Dennis Ritchie, published in 1978. The example program from that book prints "hello, world" (i.e., no capital letters, no exclamation sign; those entered the tradition later). The book had inherited the program from a 1974 Bell Laboratories internal memorandum by Kernighan ?Programming in C: A Tutorial? which shows the first known version of the program.
Although the 30-year voyage of Hello World is less impressive than Lorem Ispum?s 500 (or 2050) year journey, both were circulated by agents largely ignorant of their origins. Both examples are not merely text, but the combination of a conventional text with a cultural practice within a creative community. Both practices emphasize the explication of design syntax ? in the case of Hello World, this design syntax being the programming language itself.
The difference between evaluating the layout and the code however is the difference between the generally ?framing? or ?surface? status of the layout and the generally ?underlying? or ?prior? status of the code with respect to text. Whereas in graphic design evaluation takes place by considering the holistic page (without being distracted by the meaning of words, which exist within that page), in programming evaluation proceeds more along the lines of Norbert Weiner?s information theory, by testing the channel of message passing (which must emerge out of or because of the compiled code). Thus, while the text ?Hello, world!? is as much pseudo-content as ?lorem ipsum,? the practice of Hello World brackets the question of reading differently. Rather than suspending reading, it essentializes reading by presenting a minimal-length message which is equivalent to a binary return ? a ?yes? or ?1? which is a response to the unspoken question ?did the program work, or didn?t it?? This message takes the form of a greeting from the ghost in the machine.
3.3 Adventure and Cloak of Darkness
A practice of benchfic in elit, however, would not just be an examination of outputs or ends, but an examination of models of interaction. If there is a precedent to benchfic in software, it probably begins the ur-text of interactive fiction (IF), ?Adventure,? a text whose history is described by IF scholar Nick Montfort as intimately tied to the rise of the internet, computer culture, and the game industry [14]. Co-authored and widely adapted / reimplemented from its inception, ?Adventure? later entered a new stage of ubiquity when it was repackaged as the commercial Zork series by Infocom and became the landmark text of the emerging commercial computer game industry.
The repackaging of ?Adventure? arose out of a culture of reimplementation that was already widespread, however the interest here is in Infocom?s strategy of distributing their interactive fictions via a virtual machine, abstracted from that first text and subsequently implemented on countless hardware platforms.
Like later virtual machines (such as Java), z-code instructions could be executed on any platform by an interpreter that met the specification. Whereas the specification Hello World is almost the simplest imaginable, the z-code interpreter was quite complex, and specifying it represented to some extent a formalization of the possibilities of the genre of IF as imagined by Infocom.
Today the distribution of IF virtual machines is a cultural tradition in programming and specifically within the community of UNIX/POSIX distributions, where inclusion in an operating system is seen as a point of historical pride.
With IF interpreters however, much like the significantly simpler Hello Wold, the goal was not to examine variations in behavior, but rather portability, achieving nearly identical behavior by means of widely varying code. Today, simple z-code interpreters are available as everything from command-line applications to Firefox browser plug-ins, however their behavior remains largely the same.
However it is perhaps not surprising that IF?s culture of portability gave rise to another project with the specific purpose of providing a kind of benchmark or comparative metric. Roger Firth?s ?Cloak of Darkness? website implements the same short scenario about a mysterious message through a variety of interactive fiction langauges, with some discussion of the varying capabilities and limits of each langauge. [4]
This site tries to help in your evaluation, by presenting the same (very small) game using a range of authoring systems. The implementations have been made reasonably consistent, so as to facilitate comparison. As well as the game source... we sometimes provide information on how it was compiled, present a transcript showing it being run, and try to mention some real games that you might also like to try.
As with Hello World the emphasis in ?Cloak of Darkness? is on achieving similar behavior ? yet unlike it the purpose here is also to reveal dissimilarities of effect (reader experience) rather than simply cause (author implemented code). In this Cloak seems like a synthesis of these two practices, and a more sophisticated model of evaluating variation in interactive media. However, substantial differences from the concept of benchfic remain. The target audience for Cloak is primarily authors looking to evaluate potential languages for use in future projects, rather than critical readers considering the reading/playing experience in varying contexts. In addition, Cloak conceives of itself as operating within a genre (IF) rather than across genres (IF, Hypertext Fiction, Chatbots, etc.). To that end a ?specification? ? a descriptive list of component features to be included in the scenario ? rather than a source, something which must perhaps be conceived of more flexibly in order to operate across disparate genres. A retelling of the same Cloak scenario in widely disparate forms would expand the comparative focus from IF?s varieties to its limits, highlighting what is characteristic of IF and what lies beyond.
3.4 Gabriella Infinita
Whereas ?Cloak of Darkness? is our most sophisticated example of a collection of adaptations proceeding from a specification towards a concept, other sets of adaptations have been the work of a single author, proceeding from concept into specificity.
One useful example is Jaime Alejandro Rodr?guez Ruiz?s ?Gabriella Infinita? [15] a set of crossmedia works that are available as ebook (PDF), hypertext, and hypermedia. For Rodr?guez Ruiz, a professor of electronic literature at the Pontificial Universidad Javeriana in Bogota, the various intermedia iterations of the tale follow his own eductation in hypertext and hypermedia. He refers to the entire production as an ?obra metamorfico? or metamorphic work, and indeed the story has had various formal incarnations. He began with a novel, or as he puts it, ?It began as a bud, greedy about its body, realizing its fragility and contingency and in the end realizing that it was destined for volatility.? (Of course, this translation of Rodr?guez Ruiz?s words also possesses such volatility.) Indeed ?Gabriella Infinita? seems to yearn for an infinite, boundless state. The novel begins with Gabriella arriving at the scene of her now missing lover, Frederico, a tormented genius author. On that note of the author gone missing, the text uses multiple points of view, metafictional twists, and all the self-reflexivity and narrative complexity of the works of postmodernism. It is a book very much at odds with or in contest with its form. In 1997, Rodr?guez Ruiz encountered hypertext as a literary form at a conference on the novel and postmodernity. As his book seemed to be yearning for this new form, he adapted the work. Following the current of technology, he adapted the hypertext to hypermedia with the collaboration of voice actors, visual artist Clara In?s Silva, and programmer Carlos Roberto Torres.
Although Rodr?guez Ruiz presents his works as moving ?towards? hypermedia, by posting all of the forms with commentary, his work begs for comparative analysis. In many ways, the fixed sequence text form of the story presents more versatility than the later forms. Its pages can be accessed in any order, its voices and images supplied by the user. Also, as a PDF, the full text is searchable. On the other hand, the addition of film and voice clips adds layers of information and intonation that cannot be found in the print-only text. In any case, this work in its current online form stands as the archeological strata of cities built upon each other, retaining basic infrastructures, while adding new technology. Though presented in a narrative of progress, these proto-benchfics offer valuable lessons in creative adaptations and revisions in different electronic forms.
3.5 Our First Source: The Lady, or the Tiger?
When selecting a first text for the Benchmark Fiction Project to use, we had a number of criteria. To enrich the introspective process of creation, we hoped for a story whose themes resonated with new media or eliterature. To make the process of producing benchfics accessible, we wanted a piece that was short, yet to keep them rich we wanted one that was thematically substantial. A range of characters, times, locations, thoughts, and physical actions were desirable, but again that range needed to be limited. Finally, we wanted one that was free and clear of all the sundry complications of international copyright law - and, if possible, available in some widely disseminated electronic text form such as via the Guttenberg project. In practice, this meant either something published under a recent and explicit copyleft license (such as Creative Commons) or else something old enough to have lapsed into the public domain.
In the history of the short story, Frank Richard Stockton 1882 short story ?The Lady, or the Tiger?? (LOTT) stands as a kind of anomaly [16]. It is the tale of a ?semi-barbaric? kingdom with a uniquely quantum system of justice. Any alleged criminal is brought into the arena and faced with two doors. Behind one waits a beautiful woman whom he will marry (as all criminals are men, presumably). Behind the other paces a hungry and ferocious tiger. On a bed or on a plate, justice is served.
Unfortunately, for the princess, who is also ?semi-barbaric,? her lover is caught in a crime. She finds out what is behind each door, and when the criminal looks up to her before choosing, she gestures to the right. But we do not know that she has gestured towards the door with the woman, because the Princess knows the lady behind the door and suspects that she is no lady. To be sure, the Princess suspects that that woman has had designs on her lover. But how could she feed her lover to a tiger? The story ends with the author declining to resolve the question, and asking the reader to choose the outcome. We are left in the world of the indeterminate.
Instructors in writing workshops attempt to stave off a tide of similar open-ended stories by telling their students that ?there can only be one ?The Lady, or the Tiger.?? Yet Stockton?s story prefigured the rise of an insistence on indeterminacy that became one hallmark of contemporary postmodern fiction.
LOTT is also unusual in its final request that the reader intervene as writer ? a second person mode that is the hallmark of both the multilinear ?Choose Your Own Adventure? gamebooks and IF. As an early exemplar of the indeterminate story, LOTT is often seen as a precursor to the multiple paths realized in later multilinear storytelling. Indeed, hypertext author Michael Joyce wrote LOTT into hypertext history by referencing it in his seminal ?afternoon: a story.? [9] His lexia labeled ?The Lady or the Tiger? teases:
It comes down to that, doesn?t it? Despite what we think of our techno-philosophical advancement? Love or death. Risk with two faces. Go on, press the button, treat it all as if it were real. The lady? or the tiger?
Choosing ?the lady? however sends you to a lexia labeled, ?you have no choice.?
Joyce?s invocation of LOTT at the inception of literary hypertext, establishes the tale as one of the literary influences while his evocation of ?techno-philosophical advancement? questions any illusions we have of progress in storytelling through electronic forms.
4. BENCHFIC IN PRACTICE
Our source selected, the Benchmark Fiction Project begin experimenting with various elit forms for rendering the text. Though relatively new to media history, electronic literature has produced a large number of genres and subgenres, with even more media-specific channels for producing those genres. To offer a brief, but by no means all-inclusive list, we might conside ?The Lady, or the Tiger? via: Interactive Fiction (Inform, TADS), Hypertext (Storyspace, HTML), Hypermedia (Flash, PowerPoint), Chatbots (AIML, Yapanda, Personality Forge)? the list becomes even longer the more specificity we include in an implementation. What version number of the Inform language? AIML implemented via javascript over the web, or rather over an instant messaging network?
4.1 Benchmarking as Database
Our decision was to subsume this general discussion into a lightweight framework for contributions ?a database cataloging system that could serve as a prototype for thinking through our critical practice. Our move to database modeling follows Lev Manovich?s thesis in ?The Language of New Media? that the database is the essential conceptual form of new media art, and performs benchfic as a kind of meta-art ? a new media creation in its own right. Designing the catalog led to developing the following project vocabulary:
Source: A source is simple digital text with minimal markup - as close to "content" as we can get. Sources may be the literal raw text, or may be made up of subordinate units such as pages, paragraphs, chunks, lexias, or various adaptations such as a dialog, a lexicon, etc. A given unit might be identical or almost identical across various sources, and we can compare differences. We currently have sources of ?The Lady, or the Tiger?? which are broken up by paragraph, annotated with wikicode, and parsed as word-pairs, among others.
Interface: An interface is a system for rendering content of a particular type in a particular form. It may be a print text, wiki, CYOA, chatbot etc. In conception, this might be compared to other functional form / content systems such as CSS / HTML. In the practice, the interface behaves as a function that takes a source and returns an edition, and as such includes not just display but also some kind of navigation or interaction model which might head into the AJAX end of what is possible via HTML, or beyond. We already list several PHP displays, a CGI script, and a few remote web APIs. Interfaces are in principle reusable ? by dividing the source from the interface, the benchmarker makes a particular assertion about where the essence of the text ends and the essence of the form begins. A raw text module or a wiki module might take the same source collection of paragraphs and render them differently. Not only can an interface render many sources, but one source might be rendered through several different interfaces. We currently have two such alternate wiki display engines, which draw on the catalog database, as well as interfaces for browsable paged text, printable versions, a Pandorabot, and some whimsical external examples that make use of programs such as GoogleFight and Life vs. Life.
Edition: An edition is a specific pairing of a benchmark source with an interface ? any interface can display many sources, and a source may be displayable by many interfaces. While most editions are dynamically generated, editions may also be external, or even static.
The possibility of static editions in the Benchfic Catalog is particularly important, as it means that bringing an adaptation into consideration next to other benchfic does not require following this methodology of sources and interfaces. Allowance for listing external sources also means that strong centralization of the catalog is not required. Contributions that follow such a methodology but cannot be subsumed under our current database schema can likewise list their results in the catalog. Furthermore, the current technical specification of Linux-Apache-MySQL-PHP (LAMP) is a free, widely supported standard for open source and academic projects, which should ease dissemination of Benchmark Fiction software if it reaches a stage where it is suitable for distributed educational use.
4.2 Wikifiction: TheLady or TheTiger?
Perhaps the most complete benchfic to date is the wikifiction interface used to render ?TheLady or TheTiger?? Wikifiction is a branching text much like hypertext fiction, yet taking advantage of the peculiarity of wiki syntax that the bracketed word like [this] or ?CamelCase? word LikeThis is automatically rendered into a link. Such a link leads to a new chunk or lexia whose title is the same as the link text.
While in modern wikis such as Mediawiki (which runs Wikipedia) this constraint can be removed through the use of extended syntax, the principle that ?links lead to named objects? is a foundational one for the wiki, and creates an interesting aesthetic for rendering forking or garden-path fiction through a wiki interface.
For this wikifiction interface, we chose TiddlyWiki, a DHTML wiki display script which represents the wiki entries as floating boxes which can be reshuffled by opening or closing. The result is a reading effect not unlike what Ted Nelson described as ?stretchtext.? The source preserves a default order for reading the text by ?naming? each paragraph after a wikified word or phrase in the previous paragraph, thus rendering the text as a chain or accordion.
For the LOTT benchfic done via this interface, one of the facts this highlighted was the extent to which each paragraph does not logically queue the next one, creating a need for mid-paragraph linking and an unusual titling convention. However it is serendipitous that the story?s final line ??the Lady? Or the Tiger?? is the same as the title ? allowing the final click to scroll to the top of the screen, and the wikified text to be gracefully rendered as an infinite loop. This emphasizes in part the rhetorical flourish of the original short story ? in part the tendency of wikifiction to a kind of unending interiority that is similar and yet different to that much-noted characteristic of hypertext fiction.
Equally important to the wikifiction genre is the open editing and versioning ethic which is fundamental to the wiki concept. What a wikifiction LOTT would become given an engaged reading community might be an interesting experiment in a classroom setting.
4.3 Chatbots: Turing or the Tiger
The choice of how to adapt the chatbot to the tale is almost as perilous as the choice within the LOTT itself. Should the chatbot tell the story when questioned? Should the chatbot be one of the characters (e.g. the Princess or the King)? This case proves a fundamental difficulty in benchmarking fiction. The issue is how much license the author take when adapting the story to new media forms that, not being linear, are structurally different from the original print text.
To date we have dealt with this question by allowing ourselves the maximum freedom within each form, realizing that merely using one medium like another does not do the testing justice. Just as early filmmakers learned that merely positioning a camera before a dramatic scene was not the best use of the camera (a movie is in fact not a play seen from the best seat in the house), new media artists explore the unique properties of the medium rather than simply remediating. For example, a chatbot could tell the entire story in one response to any given input, but that would not be using the chatbot?s functionality. Nonetheless, we realize that this freedom may contaminate our benchmarking process on some literal level that is frankly outside of our aesthetic interests in new media. In effect, we have added many more doors in the process of adapting LOTT.
In one adaptation of a chatbot to LOTT, the text on the HTML page presents the user with a variation on the prisoner?s dilemma.
The introduction to the conversation reads:
Behind the door lies a lady, pacing the floor in anticipation or a tiger, cleaning its teeth for its close up. The king has offered you the chance to pass notes scrawled on scraps back and forth between you and one of the doors. The king?s mage has cast a spell allowing the tiger, for the moment, the powers of speech. From your interactions and questions, you must decide who is behind this door. The Lady or the Tiger...
http://www.pandorabots.com/pandora/talk?
botid=f92e9ba34e35ea30
The user can then attempt to discern which one is behind the door. Asking, ?What is your gender,? yields the reply, ?My gender is Tigress.? Asking, ?What color is your hair,? elicits, ?My hair is striped.? The playful ambiguity of the answers returns to the indeterminacy of the story and its infamous ending. The adaptation uses the situation of the questions to challenge the interactor?s position. Asking, ?Are you a lady,? elicits ?You wouldn?t ask a lady a question like that.? The piece also refers to its own quantum state. Asking about the King elicits, ?The King, who made this system of divine justice, making the captive choose between Schrodinger and the Cat, taught his daughter every thing she knows.? With the answers, the chatbot becomes more of a complement to and a commentary on the tale than a strict re-presentation, but that is an aspect of new media as well.
This chatbot is an AIML Alicebot, using a customized version of Dr. Richard Wallace?s 2002 ALICE code and hosted on the Pandorabots site. One of the interests here is that Alicebots come with a large vocabulary of predefined responses, many of which have been left in place. As a result, scraping below the surface of the Lady or the Tiger reveals the personality of the mage that is their mystical medium. More importantly, ALICE?s response algorithm has not been changed except in the routine customization of responses. Other chatbots, whether from scratch or using another authorware system (such as UltraHal), would result in a quite different benchfic.
4.4 Google Fight
More playful still are our interpretations of the ?source? of LOTT in the Google Fight and Life vs. Life interfaces. These experiments push at the limits of what can be considered a legitimate adaptation, or benchfic.
Google Fight is a website that uses the Google search engine to put two search terms against each other in the context of a physical battle. After the terms have been entered, a brief flash animation plays of stick figures battling. Google is searched for each term, and the resulting page counts are represented as bar graphs, with the more commonly matched search term crowned as winner.
As a first attempt at feeding LOTT through an external interface, we took significant word pairs from the story and fed them into the GoogleFight interface to see the result. When confronted with the choice between ?The Lady? or ?The Tiger,? what would the PageRanked internet find collectively more interesting?
Our ?Lady Vs. the Tiger? edition of LOTT is that particular fight. More precisely, it a link leading to the CGI which asks that particular question, meaning that each user to accesses it witnesses a result which may be similar or different. This Google search battle will fluxuate throughout the life of the internet (and the site). As of 6/23/05 at 12:24 pm, PST, the Lady was ahead by a slight margin (34,200,000) to the Tiger?s (33,500,000). After watching these battles, we began to question whether LOTT was really about a choice between the Lady and the Tiger or between Trust and Suspicion in which case, as of this writing, Trust wins by a landslide. Left Door vs. Right Door? King vs. Princess?
Google Fight allows for variability in each instantiation and keeps the story in flux as the internet is. It also complicates the idea of the binary choice by resolving it through an enormous quantitative calculation, itself the result of a qualitative (PageRank) algorithm. Of course, GoogleFight work better integrated into a more contextual adaptation of the tale, as the two search terms offer no backstory, exposition, or narrative outside of combat. Nonetheless, the central indeterminacy of the story as well as an accentuation of chance have been brought out in this whimsical adaptation.
4.5 Life vs. Life
If GoogleFight resolves the final chapter of LOTT through leveraging the aggregated digital text of the internet, Life vs. Life is an algorith for resolving conflict between individual pixels.
Life vs. Life is a website which provides a head-to-head rendition of the Life cellular automata algorithm. In it, a red and a blue pattern compete, and, like in the game Othello, the color with the most dots wins when no more moves can be made. The system is set up with a pattern authoring environment and a competition rankings board wherein any two patterns may compete. With a fixed grid and a limit of 60 units per pattern, a few simple aesthetics of warrior-patterns have emerged over time to dominate the rankings.
What if we treated the pixel-fonts of ?The Lady, or the Tiger?? as material, and tried to resolve the conflict operationally? Rather than design killer geometric forms, the interface is here repurposed to run bitmapped text fonts, which allowing the rules of Life to determine which of two textual propositions overcomes the other ? which text propagates to success, and which text fails and is consumed. Life vs. Life is drive by the metaphor of bacteria, which are not only reproducing and competing (as in the original Life) but also consuming one another. ?The Lady? may defeat ?The Tiger? in these examples, but she does so by becoming that which consumes. The situation here mimics that of the princess in the original story ? a woman about to watch her lover consumed by one of two possible fates, either of which will separate them forever.
There are many other interesting qualities of the medium (mirroring, determinism, the performance of the battle over time), however many of the hard decisions in creating a benchfic relate to pragmatic decisions over choosing font point sizes, deciding on visual arrangement such as spacing and alignment, dividing up the text (what to do with the ?or??) and so on. The life algorithm is extremely sensitivity to initial conditions, making the composition process an odd echo of the precarious nature of the lover?s choice, whose fate balances on a knife?s edge.
5. COMPARISONS
Our early results prove some interesting lessons about LOTT. One outcome has been the evolution of our understanding of the story. As with any form of adaptation, the original must be understood in terms of what we determine to be essential elements. Through both literalist transcriptions and fanciful elaborations on those elements, we gain a greater sense of the stakes of the text. Yet we also come to appreciate what kinds of stakes are inherent in the forms of the benchfic. The energy of the insight lies in synergy.
Indeterminacy plays out differently in the different forms. In wikifiction, the topic headings highlight the changing stake of the conversaion. In GoogleFight, the seemingly random evolution of the networked pages of the internet come to substitute for the operations of chance or Fate. In Life vs. Life, our intervention in initial conditions propels us into the butterfly-effect unfolding a determinate universe, while in chatbots, that indeterminacy is the question of the unknown entity behind the interface and the play of the guessing game of Turing?s ?Imitation Game.? While this last example does not directly match the situation of LOTT, it does put the user in the same position of the interrogator in a situation where (in Turing?s proposal) choosing the wrong door may diminish his/her own life. The choice of the Tiger may very well match mistaking the man for a woman, or a computer for a human. It is a problem of ontology that by merely being posed can be fatal to a world view.
Already this analysis of benchfic begins to move away from the analysis of form and back towards the analysis of the content of the story. This reciprocal relationship will be fruitful. Further, the story has helped point out some of the essential characteristics of the form, such as the unknowability of the entire internet, or the experience of risk and chance when Googling, or the computational universe that underlies Life Vs. Life. That this commentary seems to be a kind of meta-analysis, involving layering on significance to the forms, also distinguishes this process from benchmarking in the computer science sense. Benchmark Fiction is a creative and interpretive activity that will yield different results to different heuristics, even when examining the same edition.
6. FUTURE DIRECTIONS
On a pragmatic level, the works will be accessible through a website, along with essays on their creation. Eventually others will be able to upload creative and theoretical contributions as well. Hopefully, our Benchmark Fiction Catalog will stand as an initial methodology towards systematic evaluation of the electronic forms available to authors of electronic literature ? although others would be welcome. No doubt, choosing other texts, other forms, or even other implementations of those forms would yield valuable results. Further, these results can serve as lessons to theorists and practitioners of new media who wish to understand the limitations and strengths of the various forms. In our example, benchmarking is applied creatively, as each adaptation reinterpreted this story. As an example of and call for a kind of creative practice, benchmarking and benchfic may also prove to be a source for re-examination of new media itself, much in the way adaptation and revision have operated in oral, print, and other artistic milieu for centuries.
7. REFERENCES
[1] Bakker, Jan-Hendrik. (2001) ?Hypertext and the human factor. Narrativity after Modernism: Jan-Hendrik Bakker in conversation with Michael Joyce.? Kunsten/Literatur. http://www2.eur.nl/fw/cfk/kunsten/hypertext.shtml.
[2] Bolter, J.D. and R. Grusin. (2002) "Remediation." MIT Press. pp 65.
[3] Dena, C. (2005) ?Texts, Worlds, Realms and Channels: Towards a Taxonomy of Polymorphic Works? presented at ?SCAtharsis Monthly Seminar Series?, University of Melbourne, Victoria, 17 Aug, published by ?SCAtharsis? [Online] Available at: http://www.sca.unimelb.edu.au/scatharsis/Events/Seminars2.htm
[4] Even-Zohar, I. (1990) ?Polysystem Theory? in ?Polysystem Studies, Poetics Today?, Vol. 11, 1, pp:9-26 [Online] Available at: http://www.tau.ac.il/~itamarez/works/books/ez-pss1990.pdf
[5] Firth, R. (2002) ?Cloak of Darkness? [Online] Available at: http://www.firthworks.com/roger/cloak/
[6] Hayles, N.K. (2002) ?Writing Machines?, MIT Press, Cambridge, Mass.
[7] Herman, D. (2004) ?Toward a Transmedial Narratology? in ?Narrative Across Media: the languages of storytelling? (Ed, Ryan, M.-L.) University of Nebraska Press, Lincoln, pp. 47-75.
[8] Jenkins, H. (2001) ?Convergence? I Diverge? in ?Technology Review?, June, Cambridge, MA, pp. 93. [Online] Available at: http://web.mit.edu/21fms/www/faculty/henry3/converge.pdf
[9] Jenkins, H. (2003) ?Transmedia Storytelling? in Digital Renaissance section of ?Technology Review?, 15 January, Cambridge, MA. [Online] Available at: http://www.technologyreview.com/articles/03/01/wo_jenkins011503.asp?p=1
[10] Joyce, M. (1987, 1991) ?afternoon: a story.? Eastgate Systems.
[11] Klastrup, L. and S. Tosca (2004) ?Transmedial Worlds: Rethinking Cyberworld Design?, ?Proceedings International Conference on Cyberworlds 2004?, IEEEE Computer Society, Los Alamitos, California, published by ?Klastrup Cataclysms? [Online] Available at: http://www.itu.dk/people/klastrup/klastruptosca_transworlds.pdf
[12] Lempke, J. (2004) ?Critical Analysis across Media: Games, Franchises, and the New Cultural Order? presented at ?First International Conference on Critical Discourse Analysis?, Valencia, published by ?Jay Lemke?s Personal Webpage? [Online] Available at: http://www-personal.umich.edu/~jaylemke/papers/Franchises/Valencia-CDA-Franchises.htm
[13] Manovich, L. (2003) ?New Media from Borges to HTML? in ?The New Media Reader? (Eds, Wardrip-Fruin, N. and Montfort, N.) MIT Press, Cambridge, Mass., pp. 13-25.
[14] McClintock, Richard. ?Lorem oopsum: Dummy text is Latin, after all? in ?Before & After? issue 19, pp 3.
[15] Montfort, N. (2003) ?Twisty Little Passages?, MIT Press, Cambridge, Mass.
[16] Rodr?guez Ruiz, J. (2002) ?Sobre el proyecto.? ?Gabriella Infinita.? http://www.javeriana.edu.co/gabriella_infinita/proyecto/historia.htm
[17] Stockton, F.R. (1882) ?The Lady, or the Tiger? in ?Century Magazine.? Project Gutenberg, Etext#396. [Online] Available at: http://www.gutenberg.lib.md.us/index/by-title/xx849.html
[18] Walker, J. (2004) ?Distributed Narrative: Telling Stories Across Networks? presented at ?Association of Internet Researchers 5th Annual Conference?, Brighton, 21 Sept, published by ?jill/txt? [Online] Available at: http://huminf.uib.no/~jill/txt/AoIR-distributednarrative.pdf
