To access the “Design Pages” referenced in this report, please either scroll to the end of the post (low res) or download this pdf (high res).
Information Needs of Intended Users
Intended users of the Author Collection of Max Barry are readers of fiction, and beyond their shared characteristic of fiction reading, the users will typically be interested in writing fiction and understanding the process of writing fiction from an author’s perspective. English speakers with at least high school level proficiency in reading are the primary intended users; although some documentary units will be available in additional languages. Users will share eclectic interests that connect to questions of personal autonomy in different settings, and entities that address problematized scenarios of personal autonomy in those settings. Such questions will include: How does Lexicon’s parallel society reflect the targeted-marketing practices in highly industrialized countries? How does Company address the creation of discrete corporate and public identities by modern employees? What can Diablo III teach us about individual’s market demands in a centralized economy?
In order to engender high maintainability, the intended users are limited to those who share interests in the aforementioned area of questions, and who share an interest in the fictional works of the author Max Barry; users who may be interested, for instance, in the duality of work and non-work identity but prefer non-fiction will not be targeted in this index. By limiting the intended user pool, design decisions such as the data visualization of facets on (Design Page 1) can reflect targeted areas of interest. Further, targeted audience increases usability by offering specificity and depth to users. As there is no competing collection for this curated set of indexing units, even within the indexing units themselves, the users will see this collection as a primary point of reference. Accessing units via a “search” bar limits users to terms and ideas they can already articulate and which appear explicitly in the text of the e-book or blog post, but the facetted index of the collection also address the implicit ideas of the text and allow users to navigate to indexing units that address a broader set of their interests.
Subject scope is limited to increase maintainability and usability by two factors: single authorship of documentary units, and use of Social Theory as a guiding theory of facet selection.
Limiting the collection to the documentary units of one author limits the range of subjects that can be explored by virtue of the limited time and output that any given author can provide. Although Barry addresses a range of eclectic subject exploration across his novels and blog posts, he has a limited output and will eventually pass away and close addition to the collection entirely. Further, as an author, he has a professional and artistic inclination to explore similar ideas frequently. Once these limited facets are established firmly, it is likely that most of Barry’s work will in some way address them as part of his consistent artistic practice.
The use of index terms and subject headings consistent with the discourse in Social Theory further limits the subject scope. This policy is explained to users on (Design Page 2) in the first paragraph under the title “About the Collection…”. As with limited authorship, a specific theoretical framework will shape the scope of the subjects by providing firm and established facets, but allowing for new facets as new work is done. Thus, the subjects in the index can be updated to reflect current discourse in the space by consulting the new vocabulary and associations within Social Theory.
Users will value this framework for subject scope because it will remain relevant (as it can be updated to reflect any new ideas), but it is appropriately limited to those topics which will best address their potential questions and interests.
Within the collection itself, the primary limiting factor to scope is the limited authorship of indexing units to one author makes the collection highly maintainable. Further limiting the scope to specific formats, books and blog posts, improves maintainability as well. Barry produces regular blog posts monthly, and produces books roughly every 2-3 years. These restrictions for choosing indexing units in the collection allows the number of indexing units added at any one time to be kept to a reasonable amount.
A limited number of documentary units means a limited, manage-able number of documents to categorize for the indexer. Further, the focused attention that can be provided to each documentary unit will allow users to experience optimal usability of the index; because of the limited scope and domain, users will be able to access greater depth of categorization of each unit.
Documentary units for this collection are sourced from web log posts available on maxbarry.com, and books published by Vintage Books (or similar presses, such as Penguin). Selection of the units is based on authorship: limited to the author Max Barry.
This highly restricted domain controls the amount of documentary units included and allowed into the collection to be highly maintainable by a limited number of indexers. Although this limits the collection’s usability for broader scholastic or commercial use, it addresses the audience need for concentrated and deliberate categorization of Barry’s many eclectic works.
Secondarily, as a by-product of the limited domain, the collection can be used to access highly specific documentary units that may help the user navigate another collection with more success. For instance; consider the record on (Design Page 7 and 8), accessing such a documentary unit regarding marketing and socialization in my collection may encourage the user to further explore other units in the collection regarding either facet, and this term may be more helpful to them in exploring other collections or making successful searches for similar resources in databases. Rather than using their search string from (Design Page 6), they may be encouraged to use “behavioral marketing” or “group socialization” to find documents of further interest, and this type of terminology will likely increase their success in accessing relevant documents.
The collection consists of a facetted index, where top-level categories are the facets and headings, and sub-headings are displayed as either is: a part of or is: a type of semantic relationships, with limited associative relations. The top-level categories are facets of Social Theory and reflect the common usage in the field.
In keeping with Wellisch’s ideas of consistent semantic relationships as the preferred display, the collection seeks to minimize associative relations and maximize clear relationships to make the index more readable to users. On (Design Page 3), both fully consistent and hybrid-associative arrangement is shown. The “Corporatism” facet is fully semantically consistent, as each subheading is a part of a Corporatist political philosophy. By contrast, the “Capitalism” facet has the subheading “Regulation (of) Capitalism” which breaks the convention of being a part of Capitalist economic theory. Limited associative relations are acceptable for this index because of the highly limited scope and domain; the relationships available for users to browse at any given time will be small and more easily parsed than they would in an index with broader scope and domain.
As intimated in Arrangement and Subject Scope, the vocabulary for the index is managed by limiting the facets and displayed terms to standards of language in use in Social Theory discourse. Further, as illustrated in (Design Page 4), the scope notes for each term are available to the user and allow the index to explain Social Theory as the authority file controlling the vocabulary. The sample text on (Design Page 4) is in reference to the displayed index term “Popular Culture,” and would allow a user who does not have an interest in this particular definition of the term to seek out an alternative term. Further, it illustrates how common synonyms in Social Theory discourse are addressed using “see also” references. These pop-up definitions will allow users to understand terms that may be unfamiliar, and, further, it will allow advanced users to evaluate if the index reflects the definition of the term they are seeking to explore. The aforementioned factors increase usability by making the vocabulary clear to users, and increase maintainability by establishing a clear framework for vocabulary selection.
Analysis and Indexing Methods
The index is maintained through manual indexing. Due to the limited number of documentary units, manual indexing is a maintainable practice, and manually indexing each unit will increase usability for users by providing much needed evaluation of the aboutness of each eclectic document. Consider the record shown in (Design Pages 7 and 8); the web log shown is a blend of marketing and socialization facets, with a particular focus on an artifact of culture, a video game. Automatic indexing and full-text search alone would provide users only the terms most frequently used/parsed in this document, which would not include the term “socialization” at all. If B.F. Skinner were selected by the automatic index as a subject, users who were not familiar with his work would not seek it out, nor have a good idea of how his work may relate to the marketing practices explored in the piece. Thus, not only is manual indexing feasible in this index due to its limited scope and domain, it provides users with more precise (and thus relevant) results. The index combines this manual indexing with full-text search because users expect full-text search, however, and it will allow them to access documents that may contain terms of interest not contained in the facets of Social Theory.
Depth of Indexing
Using the record on (Design Page 8) as the example documentary unit, it is clear that Exhaustivity is relatively high for the size of the collection; 7 terms are assigned to the document displayed and the available terms in the index are 65. As the collection strives to address a variety of user interests, it is important that relevant terms be included at the expense of speedier browsing. Further, due to the overall size of the collection being so limited, the speed loss to high exhaustivity felt by users is expected to be low.
Thematic Specificity is low, and is designed to connect users with more documentary units with some relationship to key ideas in each facet, rather than focusing on the smaller subthemes within each facet. For instance, in (Design Page 3), the main idea of many documents is labelled under the sub-heading of “Free-market Theory,” which contains within it numerous different subthemes, such as “trade,” and “regulation,” not expanded on in the index. Due to the small scope and domain, more relevant documents can be displayed to users by focusing on main ideas within each unit, rather than dissecting them to their most specific sub-theme, and in order to avoid search results of 0 low thematic specificity is appropriate.
This combination of high exhaustivity but low thematic specificity, prioritizes recall over precision, as users of the index are primarily concerned with accessing documents of general interest rather than accessing specific documents of particular interest.
Syntax for Representation of Displayed Indexes
The collection uses pre-coordination of terms into facetted headings with subheadings (Design Page 3). The collection also uses post-coordination via advanced search through “Limits” fields (selected categories from the xml records and the facetted headings and subheadings respectively) and Boolean operators (Design Page 5). The facets are displayed in title case and the sub-headings in uppercase when displaying as the browse by subject index (Design Page 3), while all index terms are in title case for the alphabetical subject index (Design Page 4). The browse by subject index display allows for smaller subheading text to remain legible and establishes the facets as the “titles” to each chart, which increases usability by making a clear distinction of what the data visualization intends. The alphabetical subject index title case display helps reinforce the alphabetical categories by having each index term share the same capital letter as each alphabetical category. Both design decisions reflect a desire to increase usability for the user, by making it obvious to them how terms are connected visually as well as syntactically. Standard word order is the norm, with few exceptions, in both displays to help combat the controlled vocabulary and syntax of Social Theory terminology.
The facets and subheadings beneath them are alphabetized according to NISO standards (Design Page 3, 4). Although the heading “jump to…” hyperlinks at the top are not organized by NISO standards – they are organized by the order in which the # and * keys appear on the QWERTY keyboard for the ease of users – the terms themselves would follow the arrangement of spaces, symbols, numbers, and letters if any such numbers or symbols were to occur. Articles would also be ignored if they occurred.
Locators and Links to Documents
As Max Barry maintains a personal website for web log posts, and book hyperlinks can be feasibly maintained due to the limited number of them, all documentary units in the collection can be hyperlinked to both displayed index terms and facets. Brief surrogate records are accessible via search, and hyperlinks are provided to less brief but not full surrogate records, as illustrated from (Design Page 6) to (Design Page 7). Hyperlinks are also save-able for users to use as personal locators; this feature is only available to those users who support cookies and have not cleared their cache, consistent with current web practices for such a feature. Once a surrogate record or the full Simplified Dublin Core record has been accessed by a user, they are given the option to use “save reading,” “return to current search,” or “return to advanced search hyperlinked locators. Wherever a hyperlinked locator is, the user is reminded that it can be used by the underlines beneath them; a standard practice of the web.
As the web logs are hosted on Max Barry’s website, full-text search is supportable through use of Google Search API/App Engine, as well as document previews. With support from the author and publishers, such previews and full-text search, although not full-text display, are also feasible for the novels. Where document preview is not available, the user will have the hyperlinked URL to the novels’ point of sale page. Using a Search API will increase usability and maintainability by removing the need to code full-text search for users from the indexer, and allowing users to search web log posts, in particular, in a way that is familiar to them. Locators are consistently underlined to increase usability by making them obvious to users, as well as improving maintainability by increasing visibility of any broken links to the indexers.
Search on the Max Barry Author Collection is divided into “direct” search, and advanced search. Direct search is available to users at any time from the search bar that appears on all Design Pages and advanced search is available via the global menu at the top of all Design Pages, and illustrated on (Design Page 5). Full-text is performed via the direct search, with the Boolean AND as the default, as this is common practice on the web, and fully supportable via Google Search API/App Engine. The advanced search allows for full-text search, but adds the manually generated facets (coined as TOPICS for increased usability by the users), index terms (coined as KEYWORDS for usability by the users), formats (as described to users on the opening page, and as used in the Simplified Dublin Core records), and date (shown as a slider for ease of use by users, and part of the Simplified Dublin Core records). Additional Boolean operators are offered on advanced search, making it more powerful to users than direct search; this also helps to ease the effects of the prioritization of recall described in Depth of Indexing.
Another powerful search feature that increases usability is illustrated on (Design Page 6): the “people who saved this also saved…” feature, which is presented to users as “More like this…” on (Design Page 7). As the collection is primarily built of web logs and digital users who share a common interest in the author, they are likely to appreciate features that show them what other users have found relevant to their searches. Further, this feature would allow the indexers to assess their records for completeness, as users who access certain documents may indicate where index term assignment needs to be rethought for optimal use, and they could manually add similar documentary units to the list if thought to be useful. As a result, maintainability is increased once the feature is successfully implemented.
The surrogate records displayed on (Design Page 6) reflect the indexes focus on high usability for users and the combination of manual indexing and full-text search. Each record lists Title, Topic(s) with any selected keywords or search terms highlighted, Date, Format, and a Text Preview with any search string from the full-text search highlighted. Highlighting terms allows the user to better understand why these records are displayed, while the other fields help them evaluate if these records seem appropriate for their interest. Users have control over the sorting of results; either sorting by date published, or popularity/relevance (as determined by search). The limited fields displayed reflect primary user interests: what terms from my search put this here, what format is the result, when is the result from (particularly relevant for topical information)? By focusing on user needs, usability is increased, while the limited displayed fields that are selected from the full text or from the xml record allow for high maintainability. As long as the xml record is consistent, all search results will display appropriately to users.
The collection uses the Simplified Dublin Core metadata scheme in order to balance usability and maintainability of the index. Simplified Dublin Core is limited in fields; and all the fields it is limited to are highly relevant to book and web log posts. By following the standards of Dublin Core, the index is made interoperable with linked data and the Semantic Web initiative, and these connections will allow the index to be more relevant and accessible to users via those methods. Further, Dublin Core is endorsed by NISO, so the metadata scheme and the alphabetization policy are in harmony. Maintainability is controlled by using Simplified Dublin Core because the amount of metadata and the style in which that metadata is entered are limited to 15 field types. As the collection is manually indexed, having a smaller amount of data entry required will make it more maintainable.