The aim of this master's thesis is to investigate and to explain the concept of folksonomy in the perspective of knowledge organisation. By creating a conceptual explanation and defining what folksonomy as a concept consists of, the intent is to establish a basis for evaluating folksonomy and to discuss what relevance folksonomy may have in the field of knowledge organisation. By applying a qualitative content analysis a review is performed on how folksonomy is perceived and discussed in a selection of articles written on the subject. Employing the theory of classification, subject indexing and bibliographic retrieval, folksonomy as a method for knowledge organisation is analysed. The practice of tagging in existing systems is also studied. The study shows that folksonomy as a concept consists of documents, tags and users as well as the relations between these factors. The essential attributes are the lack of rules regarding creation of tags, the freedom from authorities imposing guidelines on how to determine the subject matter of documents and the participation of the users in both the indexing and retrieval phases. The term folksonomy represents both the browsable collection of documents and tags, the activity of assigning tags to documents and the vocabulary that is constructed. The relativistic and subjective nature, the simplicity and the lack of rules are inherent qualities of folksonomy, by some perceived as its major merits and by others as its flaws, hence the reasons why the application of folksonomy in knowledge organisation may be problematic. / Uppsatsnivå: D

The aim of this thesis is to, through an empirical study ofthe social tags in the City of Stockholm libraries’ OPACwith critical discourse analysis, examine the social tags’linguistic design, usability, and communicative potential– all in relation to the context. It is a linguistic study andit is important to take into account that the tags are extremelyshort texts – single-word-texts. My starting-pointis that the tags are a form of utility texts.What I find is that a clear majority of the tags are nounsand function as index terms, but that there are differencescompared to authorized index terms. The study showshow the tag-authors solve their contextual premises by forexample word-compositions. I find that a balance is importantbetween specific and general as the tags lack aparent-child or sibling semantic relationship. Compared totraditional knowledge organization, there is a consistentfeature of subjectivity in the tags. I see a fourth type oforganization in addition to cataloguing, classification, andindexing – the personal comment. In the retrieval, the tagsare best suited for browsing. I also find that the social tagscould function as a complement to a controlled vocabulary,but not replace it. / Program: Bibliotekarie

The aim of this thesis is to study the phenomenon of read/write OPAC by looking at how students at Pennsylvania University use PennTags in order to influence document descriptions in the library catalogue. The main questions to be answered are to what extent PennTags is being used in different subject areas and what users contribute to document descriptions through tagging in different subject areas. Studying the posts created by a sample of PennTags users from a quantitative approach I try to find out if there is reason to believe that the activity of users and users’ possibility to make meaningful contributions are dependent of the subject matter being dealt with in documents described, as anticipated by the theory of pace layering. I find that user interest in influencing the library catalogue through PennTags appears to be low in general, with exception for in the subject areas of the main classes Language and literature and Social sciences in Library of Congress Classification. Users do however make meaningful contributions to the document descriptions through tagging in 59 % of the cases. The results support only vague evidence of the connection anticipated by the theory of pace layering and I suggest that other explanatory models must be sought in order to understand user behaviour in systems like PennTags. / Uppsatsnivå: D

Social indexing is an approach to manual indexing of web content. The purpose of this master thesis has been to investigate terminological and conceptual consistency in social indexing of web content. Conceptual consistency was investigated/measured focusing on synonyms and (the use of) singular and plural forms of words. In order to measure consistency in social indexing data were obtained from the Delicious website. The data set consists of five indexed objects, 3001 tags and 239 users. Synonym tags and singular and plural forms of words were identified and grouped. The indexing consistency was then calculated on five user groups, each containing twenty users. The users’ sets of tags were represented in vector spaces and each user’s vector was measured against the group’s average vector using the cosine measure. The indexing consistency was found to be low. An average terminological consistency of 15.4 % and an average conceptual consistency of 18.5 % were achieved. Synonym tags and singular and plural forms of words do affect terminological consistency but not to any great extent. The low (terminological) consistency may therefore not solely be attributed to the presence of synonyms and singular and plural forms of words among the users’ tags. / Program: Bibliotekarie

The aim for this master’s thesis is to investigate the indexing consistency in a folksonomy Delicious. This is done by investigating how users construct tags. By applying a semiotic perspective and use analysis of content the occurrences of inflected forms, hyponymy, meronymy and users own tag creations were established. Lancaster’s seven factors is the starting point for the discussion about indexing and the factors that affects it in either a negative or positive way. The indexing consistency is calculated for the first twenty users of the five objects that are investigated. For the calculation Rolling’s method has been used but since it focuses on the indexing consistency between two users, it was adapted for the calculation of the average consistency of the first 20 users. The thesis results show that users prefer to use much the same words which mean that the average consistency for all objects is 28.34 %. Like professional indexers, the users of Delicious prefer to tag the objects with nouns and they often use the singular form. Problems arise when users create their own tags because many of these tags only occur one or two times for every object. This affects the indexing consistency in a negative way.

The purpose of this bachelor thesis is to find how and if tags are used in Swedish library catalogues. Two different library catalogues from two different vendors were studied. One is from the Swedish company Axiell and the other one is from Open Library Solution. The tags were gathered from twenty books that were tagged in both catalogues. The books were a mix of both adults and children’s literature, and these were also from different genres.The following questions are asked: What features are user tags in comparison with classical indexing? What types of books are the most tagged? The theoretical framework of the thesis is based on F.W Lancaster´s theory about indexing.The result of this bachelor thesis is that tags are mostly used to identify the content of the books and the most popular type of tags is those who describe the content of the book. They are also used to identify what the books are or are used for, as a prizewinning novel or in an activity at the library. The group of users who are using tagging the most seems to be fantasy readers and possibly the library staff. The bachelor thesis ends in a discussion if libraries can learn from them what makes tagging interesting for library patrons.

Syftet med denna kandidatuppsats är att utreda användargenererad metadata (taggar) som komplement till kontrollerad indexering (ämnesord och klassifikationskoder) av skönlitterära tecknade serier. Studiens relevans grundas på ett behov av att utveckla riktlinjer för att ämnesbeskriva- och skapa sökingångar till tecknade serier via bibliotekskataloger (1), samt att implementering av taggningsfunktioner, via gränssnitt i onlinekataloger (OPACS), blir allt vanligare vilket förutsätter utredning för att överväga en kombination mellan systemen (2). För att besvara studiens syfte utfördes tre delstudier. Först analyserades 48 serietitlar representerade via ämnesord och klassifikationskoder i MARC-format via Libris (I), därefter analyserades de 10 titlarna - med mest metadata från steg I - representerade via taggmoln från LibraryThing (LT) (II) för att slutligen göra en jämförelse mellan de 10 MARC-posterna och taggmolnen (III). Kvalitativt orienterad innehållsanalys har använts för att samla in, dela upp samt jämföra ämnesord-/klassifikationskoder med taggar efter fasetterna: genre, tid, plats, ämne, person och form, framtagna av bl.a. Svensk Biblioteksförening (2004): ”Att Indexera Skönlitteratur”. Aboutness-teori och semiotisk-teori har applicerats för att dels skapa en djupare förståelse för problematiken i att indexera skönlitteratur (-och bild) konsekvent men även för att jämföra systemen (dvs. kontrollerat indexeringssystem med folksonomi-baserat system) ytterligare. Resultatet av studien visade att få termer, via MARC-formaten, representerade fasetter som karaktär, ämne, tid och plats medan taggmolnen från LT innehöll desto mer metadata om seriernas ”innehåll” såsom: karaktärer, (sub)genrer och ämnen. Studien visade även tendenser som att taggarna var desto mer konnotativt präglade samt att de representerade dokumentens aboutness i högre grad, jämfört med klassifikationsnotationerna-/ämnesorden, samtidigt som det skulle behövas mer ingående studier för att närma sig taggarnas egentliga betydelse. Taggarna synliggjorde även ett mer vardagligt språkbruk samtidigt som inkonsistens gällande ordstavning och syntax uppenbarades. För att kunna besvara om användargenererad metadata kompletterar kontrollerad indexering krävs fortsatta studier kring varför användare taggar samt vilka termer som används vid informationssökning av tecknade serier. Analys av metadata från lokala bibliotekskataloger skulle även behöva göras för att ge en mer rättvis bild av hur ämnesordsindexering av tecknade serier ser ut i praktiken.

The aim of this thesis is to study social tagging in an OPAC by looking at the tags of Ann Arbor District Library’s catalogue. In this thesis I analyse the possibilities of social tagging. The main questions to be answered are what is the distribution of tags in different categories, what are the differences between fiction and non-fiction and how do the social tags differ from the terms of the professional indexing practice. Studying 500 tags I find that subject matter was the most frequent category for tags assigned to fiction and non-fiction. The tags assigned to fiction were more multi-dimensional than the tags assigned to non-fiction. The tag categories experience and task, which are unique to the social tagging practice, were not used frequently. I suggested that they still had potential to be of interest to users with common interest and taste. The study shows that the users did not, to any higher extent, use the terms of the catalogue. The tags were in general less specific than the professional terms. In the thesis I use communication theory to analyse the social tagging practice and two different terminologies. I find that the social tagging practice has the potential of being a complement to the professional indexing, allowing users to use more associations and their own terms. I find that the manga and anime community used their own terms that in many ways were different from the professional terms. / Uppsatsnivå: D

The aim of this bachelor thesis is to study and analyze the impact of LibraryThing's tagmoderation process and tag cloud limits in the library catalog, and assess the differencesbetween tags on their web page and tags in “LibraryThing for Libraries”.Communication with personnel on LibraryThing regarding their moderation processwas established, and the results were then compared with the collected data in thisstudy. Tags from 20 different books, collected from 2 different library catalogs, onewith a tag limit of 15 and one with 25, and from LibraryThing's web page werecollected and analyzed. The theoretical framework of the thesis was based on J. T.Tennis, and the data were organized using a modified version of Golder & Huberman'sseven different functions found in tags (2006).The results show that LibraryThing's claim that they only remove the inherentlypersonal tags are only partially true, since some other kinds of tags have disappeared aswell. In some cases a certain concept has been included in one book and excluded inanother. Tags concerning opinions or attributes of a book, have often been excluded,these could however serve as valuable forms of recommendations to other people,despite their personal context. The comparison between the library tag clouds limited toeither 15 or 25 tags suggested that tag clouds with an increased number of tags gave amore pronounced picture regarding the contents of the book. However, the number ofproblematic tags with synonymous or redundant information also increased when thesize of the library tag cloud grew.

Young LGBT people rather search for LGBT books in library catalogues privately than ask a librarian for help. But words that describe different sexualities and gender identities change rapidly, and controlled vocabularies do not, which make it difficult for members of the LGBT community to find what they are looking for when using more modern and informal terminology. We wanted to know the difference in how users in a folksonomy and librarians restricted by a controlled vocabulary describe the content of young adult LGBT literature. Who are most suitable to describe this literature? To do this we did a small comparative study using 50 randomly selected LGBT books, focusing on the LGBT subject headings in the Swedish national library system Libris and the LGBT bookshelves on the public social cataloging website Goodreads. We also wanted to know which LGBT related words can be found in SBI’s controlled vocabulary, to see what words librarians are restricted to. Conclusively both the users and the librarians, controlled vocabularies or not, used an excessive amount of synonyms. This led us to the idea of making synonyms searchable for the users, whilst librarians still use the structure of a controlled vocabulary. Users grouped books under acronyms for the LGBT community as a whole, whilst librarians rather split up the sexualities and gender identities into separates.