A Survey of Electronic Literature Collections

In their article "A Survey of Electronic Literature Collections" Luis Pablo and María Goicoechea describe characteristics and functions of collections of electronic literature and analyze descriptors used and the way information can be accessed. Based on their observations, Pablo and Goicoechea advocate a database structure which is flexible and can produce a dynamic archiving model as texts are registered and collected so that tags form a close set for the texts in the collection and this set can expand as new texts make new tags necessary. Further, the organization of tags into ever more complex taxonomies seems inevitable, since this provides an accurate description of knowledge accumulation with respect to the field's richness. They postulate that the study of tagging practices applied to digital works provides us with guidelines not only to describe texts of electronic literature, but also to demonstrate the wide variety of forms which a literary text can embody. Luis Pablo and María Goicoechea, "A Survey of Electronic Literature Collections" page 2 of 9 CLCWeb: Comparative Literature and Culture 16.5 (2014): Special Issue New Work on Electronic Literature and Cyberculture. Ed. Maya Zalbidea, Mark C. Marino, and Asunción López-Varela Luis Pablo and María Goicoechea A Survey of Electronic Literature Collections One of the problems in the preservation of born-digital material is the rapid development of new technology and consequences of the obsolescence of older technology. For example, html coding becomes obsolete on certain browsers, desktop computers no longer read floppy disks, many URL-s are unstable, web pages disappear because of the non-payment of server fees, authors remove their web pages, etc. For these reasons, governments, institutions of higher learning, and private companies realize the importance of preserving born-digital material. For example, there are for the preservation of born-digital material in the internet archive WaybackMachine , the portico and clockss archival systems for the preservation of born-digital scholarship, the Government of Canada Electronic Collection of Library and Archives Canada , the Government of France L'Institut National de l'Audiovisuel , etc. In the study at hand we discuss aspects of the following digital platforms established for the preservation of electronic literature: Electronic Literature Directory , Electronic Literature Collection 1 , Electronic Literature Collection 2 , NT2: Le Laboratoire de Recherches sur les Oeuvres Hypermédiatiques , Base de Récits Interactifs et Autres Oeuvres Interactives , ELMCIP: Electronic Literature as a Model of Creativity and Innovation in Practice , CELL: Consortium on Electronic Literature , Hermeneia: Estudis literaris i technologies , and Literatura Electrónica Hispánica . We pay special attention to the manner in which information about the born-digital material is retrieved and to the way new literary works are tagged (a tag is a descriptive label assigned to individual works in a database). We also discuss the presence or absence of access points to the resources themselves and that allow users to read the literary texts referenced. Our main purpose is to foreground the effectiveness of the cataloging choices implemented in born-digital preservation platforms and we hope to contribute to the debate around the development of a common standard of metadata in the preservation of electronic literature. Information compiled in the above listed directories includes authors, work descriptions, scholarly articles, information about gatherings, as well as didactic practices for born-digital literature. Since they all share the same subject matter and have similar objectives, the connections among them are tight. Given different institutional origins, each directory has a different amount of registered works: for example in the Base de Récits 176 authors and texts to the more than 700 compiled in Hermeneia to the nearly 4000 in NT2. The objectives of Electronic Literature include the identification and description of electronic literature and their bibliographical resources. The reason behind its creation was to call attention to these new types of texts and to scholarship produced about them. Because electronic literature is a recent form of art, Electronic Literature also has the objective of showing the diversity of texts leaving for a second stage the analysis of patterns. However, since 2011 Electronic Literature has been transformed and now offers the possibility of using wiki software to include tags, descriptions, and discussions collaboratively. Implementing this transformation meant downloading the information from the old system and adding the descriptions and missing tags. As Joseph Tabbi points out, the inclusion of tags proved to be of great utility and the creation of the collaborative tool gave new study perspectives, in accordance with the original objective of the organization: "The critical function of 'tagging,' though it may have seemed a minor element in my 2007 essay starts to look more central when one recognizes that collecting works of born digital writing means recognizing, and revising 'the imaginative qualities of actual things' ... The application of tags to the study of electronic literature has generated quite a bit of critical discussion about works that previously had been viewed through narrower perspectives" (). In the Electronic Literature Directory file cards are compiled about published electronic literature, as well as information about the individual texts. In order to know which texts are included and see their file cards, the Directory offers several Luis Pablo and María Goicoechea, "A Survey of Electronic Literature Collections" page 3 of 9 CLCWeb: Comparative Literature and Culture 16.5 (2014): Special Issue New Work on Electronic Literature and Cyberculture. Ed. Maya Zalbidea, Mark C. Marino, and Asunción López-Varela points of entry: the texts can be accessed through an alphabetic list of their titles or through a list of texts sorted by year. Another tab allows the user to see an alphabetically ordered list of authors and other contributors. Finally, a search engine allows users to make a simple or advanced search of all the content at the site. Another way in the Electronic Literature Directory to access texts is through tags and a section offers an alphabetic list of all the tags used to describe the texts gathered. However, the inclusion of these tags presents several problems: 1) They do not send the user to a closed set of controlled vocabulary and thus any word can become a tag including proper names such as Argentina, Alex Lemon, Brad Liening, Mark Bernstein, etc. This is positive as it allows freedom when tagging a text, but it hinders the retrieval of information; 2) The tags are not lemmatized and this means that the same concept can have variants which impede the retrieval of similar cards depending on the word which has been used: a) in singular or plural: antecedent/antecedents, anthologies/anthology, blog/ blogs, b) morphological variations: environment/environmental/environmentalism, c) semantic variations: prose poem/prose poetry; combination adventure/combinatorial combinatory/ combinatory poetics; d) digital literature/digital media/digital narrative/digital narratives/digital poetics/digital poetry/digital textuality/digital writing, e) interactive/interactivity/interactive art/ interactive cinema/interactive fiction/interactive motion graphics/interactive novel/interactive poetry/interactive visual art); and 3) Nearly all tags are in English, but when there are others in different languages owing to the origin of the commentators, there is no correspondence between cards from different languages. Similarly to tags, the lists offered to access the works through authors' names present duplications: sometimes there appear different cards for the same author. Thus, if one searches for Katherine Hayles instead of N. Katherine Hayles, the results are different and one has to access both cards for the complete information. Further, beginning with Alan Sondheim's "Text," the poems written in the programming language Pascal, and the Chinese poem "Clown" by Shaolian Su, the Electronic Literature Directory compiles works from 1979 to 2012. Under the epigraph "Antecedents" there are five more prose texts which share features with electronic literature although they were printed on paper, such as "100,000,000,000,000 Poèmes" by Raymond Queneau and "Un coup de dés jamais n'abolira le hazard" by Stéphane Mallarmé. Every text included in the Electronic Literature Directory has a card composed of the following elements: 1) A brief description of the work, signed and dated by the person who composed the card plus some data about the work, author, year of publication, language, and technical specifications like the platform and software used, and finally the tags; 2) Screenshots are habitually included, as well as the URL where the work is located. However, not every card has this much information and some, like that of "Clown" (1981), do not have tags as they come from the previous version of the directory; and 3) the discussion section permits registered users to add comments about the work described and to open a chain of comments about a specific aspect of the work. The Electronic Literature Directory a periodical publication which gathers the most significant examples of electronic literature in a concrete period. It is published under a Creative Commons license on the world wide web and as CD-ROM or DVD. The first volume was published in October 2006 and the second in 2011. The collection is a valuable repository of electronic literature, although its focus on English-language sources has caused ELMCIP to publish another anthology in 2012 with a more global vision: "The language used is predominately English, but several works are produced in other languages" (), "18 works of electronic literature from 10 different nations in 10 different languages" (). Although it might not be intentional and a response to the great influence and number of Englishlanguage texts, the fact is that out of the sixty works collected volume 1 of the Electronic Literature Collection only seven are in a language other than English. In these cases, the tag used is "Multilingual or Non-English," that is, an opposition which places English against all other languages (on this see Marino ). In Electronic Literature Collection 2 the number of bilingual or multilingual works together with texts created in other languages increases: fourteen out of the sixty-two texts collected are not in English which shows interest in other cultures and languages. Sometimes these new additions are triggered by prizes received after Luis Pablo and María Goicoechea, "A Survey of Electronic Literature Collections" page 4 of 9 CLCWeb: Comparative Literature and Culture 16.5 (2014): Special Issue New Work on Electronic Literature and Cyberculture. Ed. Maya Zalbidea, Mark C. Marino, and Asunción López-Varela texts were selected for volume one and this is the case for Isaías Herrero's "La casa sota el temps" and Chico Marinho's "Palavrador" which received prizes in 2006. Access to the texts the Electronic Literature Collection is different from that in the Electronic Literature Directory. Since the Collection offers a reduced showcase of texts, each work can be accessed from the opening screen, as well as through the alphabetic lists of authors, titles, or keywords under which the works are classified. There is no simple or multiple search option available. Once the user selects a work from the Collection, the next screen displayed is the text's title card which introduces it and provides key data about it. Another difference between the Collection and the Directory is that the Collection offers a closed list of keywords. Since it presents a reduced corpus of texts, each tag is followed by a description and links to tagged texts. As technology has evolved from volume 1 of 2005 to volume 2 of 2011, the descriptors used to tag texts in the volume 2 have changed slightly. This change between the volumes provides a fertile ground for the analysis of the evolution of digital literature in scarcely more than five years. In the following chart we compare the two lists of tags and indicate in bold the additions and eliminations between them: Collection 1 Collection 2 (3D in this list is positioned after "text movie") 3D ambient ambient animation/kinetic animation/kinetic appropriated texts appropriated texts audio audio augmented reality CAVE CAVE chatterbot/conversational character chatterbot/conversational character children's literature codework codework collaboration collaboration combinatorial combinatorial conceptual conceptual constraint-based / procedural constraint-based / procedural critical/political/philosophical critical/political/philosophical catabase database documentary documentary ergodic / interactivity / participation essay/creative nonfiction essay/creative nonfiction fiction flash flash games games gender / race / sexuality generative generative hacktivist hacktivist hmtl/dhtml hypertext hypertext inform installation installation interactive fiction interactive fiction Java Java JavaScript JavaScript locative locative mash-up memoir memoir multilingual or non-English multilingual or non-English music narrative network forms network forms non-interactive non-interactive parody/satire parody/satire performance/performative performance/performative place place poetry poetry processing processing Luis Pablo and María Goicoechea, "A Survey of Electronic Literature Collections" page 5 of 9 CLCWeb: Comparative Literature and Culture 16.5 (2014): Special Issue New Work on Electronic Literature and Cyberculture. Ed. Maya Zalbidea, Mark C. Marino, and Asunción López-Varela QuickTime retro Shockwave Shockwave Squeak Storyspace StretchText StretchText TADS textual instrument text movie text movie 3D time-based translation viral video visual poetry or narrative visual poetry or narrative VRML women authors wordtoy wordtoy Figure 1. Differences between Electronic Literature Collection Volume 1 and Volume 2 As we can see, some old tags describing volume 1 works are no longer used, for example those referring to less used software programs like Inform, QuickTime, Squeak, TADS, or VRML. Other tags appearing in volume 1 have been omitted, such as "children's literature," "time-based," or "viral," since there were no works tagged with them. And new ones have been added, e.g., "augmented reality," "ergodic/interactivity/participation," "mash-up," "retro," and "video" showing the new genres and styles which emerged. In volume 2 eliminated are "music" (which is now included in the more generic "audio" thus avoiding semantic duplication), "women authors" and it is substituted by the category "Gender/Race/Sexuality" meaning that gender has become descriptive of works themselves rather than placing an emphasis on the author's gender as the tag in volume 1 implied. Further, the tag "women authors" is associated with only three works which shows the tendency to use tags to describe the texts themselves. Another change affects the substitution of the tag "fiction" in analogy with print by "narrative" in volume 2, a category which is to encompass the treatment of time and space in multimedia texts and not exclusively the storytelling practices of print. The revision of these tags and/or keywords is a sign of the desire to improve and evolve and to create a better correspondence with the collected texts. As Tabbi states, "We recognized that a critical vocabulary, to be relevant, needed to change with each new literary genre created for each new technological platform" (). NT2: Le Laboratoire de Recherches sur les Oeuvres Hypermédiatiques includes in English and French a directory of digital literature in its Répertoire des Arts et Littératures Hypermédiatiques. As is the case with the Electronic Literature Directory, the NT2 Répertoire is just another work area developed for matters which range from bibliographic compilations of born-digital material to the organization of events about cyberculture, a blog, and the Cahiers Virtuels. The NT2 Répertoire manifests a broad perspective since it gathers also hypermedia and multimedia texts where neither written text nor spoken words appear and that we could consider examples of net art. The Répertoire's purpose is to show which texts are created for the computer screen be they artistic or literary. Although NT2's is focused on cataloguing Francophone works, it also includes other languages, mainly English, since English-language works are found in larger numbers as the following figure with the distribution by language in the cards shows: Luis Pablo and María Goicoechea, "A Survey of Electronic Literature Collections" page 6 of 9 CLCWeb: Comparative Literature and Culture 16.5 (2014): Special Issue New Work on Electronic Literature and Cyberculture. Ed. Maya Zalbidea, Mark C. Marino, and Asunción López-Varela Figure 2. Language distribution of texts in NT2 Repertoire (2005-2014, detail). Source: . Copyright release to the authors With the exception of "Eliza" by Joseph Weizenbaum (1966), the texts collected cover a period between 1986 and 2014. The Repertoire offers many access points to its information and in 2012 these access points found in lateral panels located at the right hand side of the webpage had information about the following criteria: the type of content: reduced card/amplified card (the amplified cards include a larger amount of data such as biographic information about the author, contact information, and thematic classification which leads to a dossier of similar works); type of item recorded: a work, a website, a virtual gallery, information about an event, etc.; year of creation; device: web address/CDROM; type of interactivity; format: main feature or characteristic in the text such as video, sound, animated image/text, hypertext, photo album, etc.; genre the text according to its content and textual modality: poetry, narration, photography, abstract art, etc. At the bottom of these right-hand tags there is another option to filter information which orders the entries by relevance, title, type, author, and date. Initially, the NT2 Repertoire offered the advantage of using a closed set of tags or descriptors for all the cards organized in a taxonomy containing four categories: nature, interactivity, format, and content and this closed set facilitated the retrieval of information. However, not every tag included in the card descriptions were navigable, that is, some tags were not visible in the search option offered in the main page of the Repertoire even when their frequency was higher than those which did. For example, in the case of tags about interactivity to the twenty descriptors offered explicitly one needed to add at least eleven more found in the cards. This meant that a large portion of the information search and retrieval possibilities were hidden or not explicit for the user accessing the Repertoire. This problem was solved by making all tags navigable and allowing the user to expand or reduce their list in the Repertoire's search function. The taxonomy has been perfected by including three more categories and refining their descriptors, which now are nature, interactivity, format, themes, media, language(s), and digital data processing. In our opinion, this system is efficient although it poses some problems to a beginner who would find it easier if the tags which compose the taxonomy were described in a glossary (as is the case in the Electronic Literature Collection). ELMCIP: Electronic Literature as a Model of Creativity and Innovation in Practice is for the study of electronic literature in Europe. It has been developed by a consortium of research groups at the universities of Bergen, Amsterdam, Ljubljana, Jyväskylä, the Blekinge Institute of Technology (Sweden), the Edinburgh College of Art, and the University College Falmouth at Dartington. ELMCIP's purpose is to provide information about authors, texts, scholarly articles, events, and didactic resources of electronic literature. In November 2014 it contained 2407 records about creative works and 2721 records of scholarship. Structurally it shares many characteristics with the Electronic Literature Directory such as its collaborative character, while it also has some common features with NT2. For example, with respect to languages, ELMCIP offers the possibility of grouping texts according to the language in which they have been written. This makes ELMCIP a more comprehensive database than its counterparts as it has thirty language categories including Arabic, Basque, Bulgarian, Cantonese Chinese/Mandarin, Catalan, Czech, Danish, English, Farsi, Finish, French, Galician, German, Greek, HeLuis Pablo and María Goicoechea, "A Survey of Electronic Literature Collections" page 7 of 9 CLCWeb: Comparative Literature and Culture 16.5 (2014): Special Issue New Work on Electronic Literature and Cyberculture. Ed. Maya Zalbidea, Mark C. Marino, and Asunción López-Varela brew, Hungarian, Icelandic, Irish, Italian, Japanese, Korean, Norwegian, Polish, Portuguese, Romanian, Russian, Slovak, Slovene, Spanish, Swedish, Turkish, Vietnamese. Despite these categories, not every language has many works in its category. For example, the database contains only two Hebrew works and there are no works in Bulgarian at all. ELMCIP presents a systematization of texts which seems more favorable to the information retrieval than the other directories, since it allows the sorting of the texts alphabetically by title, author, tags, and date. Through a more complete questionnaire than in the other directories— although still not an advanced search—it allows the user to filter and classify cards by languages, by chronological periods with the beginning and end year, as well as type of publication such as CD-ROM or DVD, a webpage, an online gallery, or a magazine. Further, ELMCIP allows the user to filter works according to whether they were exhibited in a conference or festival, etc., and it offers an alphabetical list of tags which lead directly to the individual texts described. It is an open list which also includes tags such as "critical," "political," "philosophical," "hacktivist," etc. ELMCIP cards are linked to the two volumes of the Electronic Literature Collection and similar to the Electronic Literature Collection, there are some issues. For example, the possibility of singular/plural alternation as in "bibliography/ bibliographies," "computer game/ computer games," "community/ communities"; morphological duplicates such as "blog/blogging/blog-literature," "canon/canonicity/canonization"; synonymic duplicates such as "cave/cave writing"; the presence of proper names e.g., "Borges, Boston, Bratislava"; etc. This is a common problem when allowing free tagging. The Base de Récits Interactifs et Autres Oeuvres Interactives is different from the above discussed collections and directories in that it has been created and is maintained by one individual, Serge Bouchardon, at the University of Technology Compiègne (see also Bouchardon and López-Varela ). Although a large number of texts in French are collected in Base de Récits, there are also many works in English. The collection is listed in a kind of excel page which allows the ordering of works alphabetically per author, title, year, web address, language, device for which the work has been created (i.e. web, off line in CD-ROM, or installation), and the date of creation. In the title field, one can include commentaries about the work. Sometimes these commentaries are included by Bouchardon himself, others stem from other researchers or visitors of the page. Despite the low-tech appearance of this database design, we find it efficient, direct, and we find especially fitting its feature inviting commentaries. Hermeneia: Estudis literaris i technologies also offers on its website a directory of electronic literature in two formats: and anthology which contains works in any language and a directory of electronic literature in Catalan. Some of the texts included in the latter directory are translations of works originally written in other languages. The title cards for each text only offer a link to the work without other information. When a text is a translation, a cross reference to the original work is provided. Users can access the directory through three entry points: alphabetic list of authors, titles, and tags. The tags are the same ones used in volume 2 of the Electronic Literature Collection with some additional ones such as "Catalan," "Spanish," "Portuguese," "Traducció al Català." However, few entries are tagged, thus little information can be gathered through this point of entry. The main problem we observe in this case is that the title cards provide scarce information about the texts. Our last example is Literatura Electrónica Hispánica, a collection of electronic literature in Spanish. Unfortunately the collection is limited and the twenty-two texts of the collection are traditional genres, namely electronic formats of the novel: hypernovel (4), hypermedia (9), webnovel (1), blognovel (3), collective novel (2), and wikinovel (3). This classification ignores the most prolific genre in electronic literature, namely poetry. Nevertheless the collection's importance lies not in its reach, but in its pedagogical value because it offers a sample of works in open access. Further, it is of note that the collection includes Latin American texts such as by Argentinian Belén Gache and Hernán Casciari (blognovels), Colombian Antonio Rodríguez de las Heras's and Ecuadorian Jaime A. Rodríguez's or Ecuadorian Leonardo Valencia's texts. The collection follwos the traditional pattern of bibliographic description including author, title, edition, series, the website in which the text is integrated, language, and subject matter following the Universal Decimal Classification. Literatura Electrónica Hispánica is not updated frequently, but news about electronic literature are disseminated through its blog "Literatura Electrónica" which has continued to collect works of electronic literature up to the present and provides links to fifty-three texts in Spanish. Luis Pablo and María Goicoechea, "A Survey of Electronic Literature Collections" page 8 of 9 CLCWeb: Comparative Literature and Culture 16.5 (2014): Special Issue New Work on Electronic Literature and Cyberculture. Ed. Maya Zalbidea, Mark C. Marino, and Asunción López-Varela Our observations on the above collections of electronic literature are as follows. Electronic literature is presented in some although not all collections as part of a wider field which covers digital art and multimedia. The convergence of image and sound transforms electronic literature into a hybrid product with its own characteristics, coming close to video and audiovisual art. As Amanda Starling Gould points out, the field of electronic literature is international, interdisciplinary, and animatedly emergent and the artist and theorist converging on the field come from different traditions—literary studies, media studies, computer and information science, art history, etc.—each bringing the assumptions, approaches, methodologies, and questions particular to his/her home discipline (). Our hope is that some of the questions raised in our study would help to improve information in file cards and the access systems, for example through advanced search engines which allow the filtering of information through several search options. While the collections offer tags with descriptions of the texts and their connection to other works with similar characteristics, they rarely establish genre classifications and tagging would be one option to remedy the lack of genre classification. At the same time, we are aware that one reason for this might be that it is probably too early to establish specific genre classifications. However, there is increasing evidence to support the impression that new media are causing genres to converge and obliterate generic distinctions which previously applied to traditional genres, a trend which should lead to a reorganization of tags. Classifying "narrative" or "poetry" under the heading of "theme," besides tags such as "ludism" or "love" might cause further problems of classification. In conclusion, the tagging of electronic literature goes beyond classification and allows for retrieval from a database. It is indeed a crucial enterprise for the development of electronic literature since it provides a framework of reference for both creators and researchers of electronic literature. It allows researchers to analyze the evolution of creative tendencies, the formation of genres, as well as to detect outstanding original work. The choices which designers of electronic literature collections make affect the way information is retrieved and how electronic literature is approached by researchers. On the one hand, we observe a tendency to classify tags into taxonomies (NT2) or to treat them as horizontal markers (Electronic Literature Collection); they can form closed sets (Electronic Literature Collection) or open ones (Electronic Literature Directory). Too much closeness and hierarchization can produce a myopic perception of the field (Literatura Electrónica Hispánica) and make certain unclassifiable works unreachable. The opposite—too much openness and horizontality—produces a similar result, namely that the search can leave a significant portion of works undetected. As in most cases, finding a middle position is the key. We advocate a database structure which is flexible and can produce a dynamic archiving model as texts are registered and collected so that tags form a close set for the texts in the collection and this set can expand as new texts make new tags necessary. Moreover, organizing tags into ever more complex taxonomies seems inevitable, since this provides an accurate description of knowledge accumulation with respect to the field's richness. The study of tagging practices applied to digital works provides us with guidelines not only to describe texts of electronic literature, but also to demonstrate the wide variety of forms which a literary text can embodies. Note: Research for "A Survey of Electronic Literature Collections" has been funded by the Ministry of Education Spain in the projects Escritorios Electrónicos para las Literaturas 2 (FFI2012-34666) and by a PICATA Postdoctoral Fellowship at Complutense University's Moncloa Campus of International Excellence (2011-2013). The authors thank Kieran Meteyard (Bath University) for his suggestions in the research and writing of the study.