The Tagging Tangle: Creating a librarian s guide to tagging Gillian Hanlon, Information Officer Scottish Library & Information Council
Introduction Scottish Library and Information Council (SLIC) advisory body to the Scottish Government on library and information matters members from library and information institutions Chartered Institute of Library and Information Professionals in Scotland (CILIPS) professional body supporting individual members from all sectors affiliated with CILIP UK SLAINTE (Scottish Libraries Across the INTErnet) shared website: http://www.slainte.org.uk Scottish Library and Information Council (SLIC) advises the Scottish Government on library matters and is also a membership based body, drawing members from library institutions in all sectors of the profession. So, on the one hand, we work closely with all library services in Scotland developing new initiatives and supporting innovation and, in addition to this, we work with the Government to influence policy and decision making on issues affecting libraries. However, SLIC also has a shared working agreement with the Chartered Institute of Library and Information Professionals in Scotland (CILIPS). CILIPS is the professional body representing individual librarians and supporting professional development, and is affiliated with the UK body CILIP. Under the working agreement, CILIPS and SLIC share offices, some staff and collaborate on many projects and initiatives, like the work I m going to speak about today. We also share a website, SLAINTE which features in the work I ll be discussing. 2
SLAINTE 2.0 http://www.slainte.org.uk/slainte2/ Moving on from SLAINTE to SLAINTE 2.0, the SLIC and CILIPS Web 2.0 services. We started experimenting with Web 2.0 in Oct 2007, initially focusing on flickr as a solution to internal image storage and sharing. This was then followed by all the services you see listed here and, most recently, twitter, which we ve just started using in the last month or so. Our aim in developing and maintaining these services is perhaps slightly different to a traditional library service. Rather than engaging with library users, our users are professional librarians so our aim has been to promote professional engagement and improve communication. But also, as a professional organisation, a further aim has been to encourage experimentation and innovation among librarians in Scotland we wanted to show librarians what s out there and how it can be used. This then led to the issue of tagging and the creation of guidelines. 3
Creating a librarian s guide to tagging: first steps What are we tagging? freely available Web 2.0/Social networking services Library management systems Why do we need guidelines? benefits of tagging professional standards and good practice Before I discuss the specific tagging issues, what are we tagging? Firstly, the range of freely available web services like the ones I ve just mentioned. But also within library management systems which are increasingly being developed to accommodate tags. Although the overall aim is to encourage users to assign tags, it s likely that librarians will have some role to play. Evidence from some early adopters indicates that librarians input is needed to guide and prompt users and, perhaps, also to guide acceptable use. But why do we need guidelines? Some people might argue that there are many benefits to pure, user generated classification that would be lost with guidelines. You cold argue that, with tagging, items can be identified as being about a number of topics simultaneously; that it s more democratic in coming from a bottom up perspective; that it s more responsive to changing terminologies than controlled vocabularies and that it allows for items of interest to be found by chance through related tags. So why would you want to lose these benefits by applying rules? Well, the trouble is that all of these benefits bring problems that could have serious implications for retrieval. So, as a professional library organisation, we felt that it was important to apply some professional standards by creating guidelines.
Creating a librarian s guide to tagging: background issues Scope of the project local level guidelines Review of other work in the field limited research undertaken elsewhere Open to collaboration within Scotland low level feedback Still low level of Web 2.0 use in Scotland starting to change More activity in the academic sector but very limited in public libraries access issues Given the scope of Web 2.0 services and the breadth of users, we were conscious that any efforts to impose control could only function at a local level and so our target was to establish guidelines for use within the SLAINTE 2.0 services. Our first step was to look for examples of other work in this area to guide our progress. This was over a year ago and we found very few examples of similar work, certainly in the English language. The only example we found was a US based librarians group on flickr that offered contributors some very basic guidelines for tagging images submitted to the group. Since then, the model for libraries engagement with the tagging of images altered somewhat with the launch of the Commons, the Library of Congress flickr initiative. As many of you will know, the Library of Congress has used existing Dublin Core records (exporting these to the description field in flickr) and then invited users to assign tags, rather than the library being involved in the tagging process. Although this has been a very interesting and successful project, this approach doesn t match our aims. Given the lack of work in this particular area, we ve been trying to stimulate some debate and encourage collaboration on this topic. However, we ve had limited feedback so far. A possible reason for this may be the low level of adoption of Web 2.0 services within Scotland, although there are signs that this is starting to change. At present there is wide variation according to library sector, with more work underway in academic libraries and public and school libraries trailing behind. One of the main reasons for this is, perhaps, that public sector IT departments block access to many (or even all) social networking and Web 2.0 websites. SLIC and education bodies are trying to resolve this with the Government so hopefully this will change soon.
Creating a librarian s guide to tagging: the challenges Variations in interface and functionality Use of additional fields title, description Specific tagging issues: Synonyms and homonyms no control Uncontrolled vocabulary wide variation in tags Specificity varies according to level of expertise Longevity of tags in terms of evolving vocabularies Spaces and multiple words variation in process Personal versus social popularity of tags such as mystuff and toread The first challenge we encountered in trying to adopt a standardised approach was the variation in interface and functionality across services. Many diverse services from different providers are covered by the term Web 2.0. It therefore becomes very difficult to find one rule that fits all. This has meant that our guidelines have had to include some service specific sections as well as more general rules. For example, additional guidelines for tagging images in flickr. We also quickly came to appreciate the limitations of tagging and the importance of using the other fields effectively. When describing items using social media services, there tend to be just three fields : title, description and tags. As both description and title will be used to aid retrieval, our guidelines set out a standardised process for both, including a set format for assigning descriptions. Now, looking specifically at the tags themselves. I spoke earlier about the benefits of tagging when outlining the reasons for creating guidelines. I d now like to raise some of the problems with tagging and the impact on creating guidelines. The issues that I ve listed here are commonly agreed difficulties with tagging, both from research and personal experience. Basically, tagging allows for all the problems that traditional cataloguing approaches were developed to address.
Tagging guidelines in practice This example of our tag cloud from flickr demonstrates how we ve tried to overcome some of these problems: You ll notice that we ve decided to use plural forms rather than singular version, in keeping with traditional cataloguing. Also, we ve opted for using lower case throughout which is the preferred option on the web and so seemed like the natural choice. Other things we had to make decisions on include tags for organisations names. We decided to use acronyms so the CIGS, for example, is for Cataloguing and Indexing Group in Scotland. We also stipulated two specific tag areas: geographical location and library sector. We deliberated for some time over the correct way to handle multiple word tags, since most of the services were set up to accept single words only. Based on what seemed to be the general trend, we decided to run the words together. The other main option we looked at was separation with an underscore. We tested both options and found that each returned a different set of results, no space returning the greatest number in most cases. The most difficult issue, however, is that of handling synonyms certainly in the English language where there are often a number of words to describe the same thing. In recognition of this, we ve tried to be as exhaustive as possible. However, this can only ever have limited success.
Tagging without Guidelines On the other hand, here s an example of a tag cloud from the National Library of Scotland, where no guidelines are applied and users are free to contribute tags: We have both singular and plural forms appearing map/maps, pirate/pirates Notice the variation in the description of the wars the various abbreviated ww forms but also world and war as separate tags We also have douglas and haig as well as douglashaig, also riverforth and forth Both plane and aeroplane are used Also NLS and the organisation name in full And there is even germany and gernany in larger font which tells us that this is the more frequently applied version! This shows the contrast with the previous slide. 8
Future developments: new guidelines Engage with key partners and wider community to encourage input. Promoting uptake in the library world: Edinburgh City Council Further Education Standards Promoting uptake beyond the library world: the demand for guidelines? facilitating sharing raising awareness and promoting use So, our aim now is to develop the guidelines for more general use. This will involve 2 main steps: encouraging input and promoting use. We want to make it easier for librarians to contribute so we re preparing to host some short polls on SLAINTE to capture opinions on specific tagging issues to feed back into our new guidelines. We re also working on 2 practical projects to promote uptake within library services. The first involves the development of a Web2.0 model for public library services, and we re also working to include processes for new technologies within a new performance monitoring framework for Scotland s colleges. In terms of wider demand for guidelines, we ve recently been approached by some journalism students regarding incorporating our guidelines within the social networking site associated with a student newspaper, which indicates that other people are interested in learning about good practice tagging. A recent UK report examined how social networking is increasingly used for workplace objectives. This could create a wider demand for a more structured approach to tagging, as retrieval or findability is likely to be more important when an organisational rather than personal objective is involved. We plan to open the new guidelines up to multiple purposes by distributing them online via a Creative Commons license. We ve also been considering promoting the new guidelines using short how to videos on YouTube and SLAINTE (in the style of Common Craft: http://www.commoncraft.com/show).
Future developments: horizon scanning Library 2.0 model increasingly adopted as user s expectations shift. Evidence from research suggests arrival at a tag consensus over time. Technological developments that are likely to change the systems for processing tags and need for guidelines towards universal tags? Looking forward to some of the wider developments that may affect our work in the future: Although many people still see web2 as a passing trend, I think it s becoming increasingly evident that there s more to it than that. The concept of web2 is more than the individual websites and services, rather, it s about how we as users and creators of content view the web and its possibilities and what we expect to be available. For this reason, I think that it is inevitable that Web2 will become increasingly embedded within, most web delivered services, including library services. As more people adapt to working in this way and tags become common practice, it seems fairly likely that a tag consensus will develop over time a logical and standardised approach as much as the same words being applied to the same type of items. I believe that this process will be guided and shaped by technology. Already there are systems being developed to simplify the tagging process electronically. For example, I was reading about one US patent that has been lodged for a computer program which aims to aggregate existing tags in order to predict tags for similar items, moving towards a model of universal tags (which supports the consensus theory). Who knows what the future will bring? But in working with technologies as they emerge and evolve, we are able to keep pace with developments and apply our professional skills in the ever changing environment of the web.
Contact details Presentation on slideshare Follow tweets on twitter for work updates See http://www.slainte.org.uk/slainte2/ for details Existing guidelines: http://www.slainte.org.uk/slainte2/index.htm Email: Gillian Hanlon slic1@slainte.org.uk