Revising digital library content in response to user requests

Alan Dawson

Centre for Digital Library Research, University of Strathclyde, Glasgow G1 1XH

July 2006

alan.dawson@strath.ac.uk

This article appears in Library Review Volume 55, Issue 9, 2006

Structured abstract

Purpose - This article uses real examples to show how issues such as authority, currency, preservation, sustainability and trust are inter-related and can be affected by user contributions to digital libraries.

Methodology/approach - Several specific user requests to change digital library content are summarised, along with the relevant responses, to illustrate issues of library management and policy.

Findings - The examples demonstrate the need to adopt consistent but flexible policies for content updating, so that conflicting principles or requests can be handled effectively.

Practical implications - The issues raised are likely to be relevant to anyone responsible for managing or updating digital library content.

Originality/value - By illustrating the relationship between specific small-scale actions and significant long-term issues, the paper highlights the need for digital libraries to follow a holistic approach to content management, and to think globally while acting locally.

Keywords - Change management, Collection policies, Digital libraries, Sustainability, User contributions.

Paper type - Case study

Introduction

Several publications have considered the issues involved in designing and creating digital libraries (e.g. Andrews & Law, 2004; D-Lib; JoDI). Some writers have also discussed the issues involved in sustainable management of a digital library (e.g. Hamilton, 2004; Middleton, 2005; Smith, 2003). This article considers one particular aspect of digital library management that has hitherto received little attention: how to respond to requests by library users to add, remove, update or annotate material.

It is usually encouraging to know that users are sufficiently interested in library content to bother getting in touch, but it is not always obvious what a suitable response should be. This can be a sensitive issue. For example, suppose that a library published a digital archive of the Danish newspaper Jyllands-Posten. Clearly the job of the library would be to hold a complete and accurate digital copy of the original paper publication. Who could disagree with that? Yet following publication of the September 2005 issue, and subsequent reproduction of cartoons of the prophet Mohammed, embassies were attacked and burned, editors and journalists were arrested, dismissed or forced into hiding, several people died in protest marches in Nigeria, while in the UK demonstrators were arrested for displaying placards carrying death threats to those who insult Islam. Is it really in the public interest for a digital library to publish material that is offensive to millions of people and has the potential to arouse public disorder?

Yet surely a digital archive can not rewrite history and pretend that the publication did not occur. One would expect the physical material to be securely lodged with legal deposit libraries (a copy of Jyllands-Posten is held by the British Library), so why should digital libraries be any different? Surely the historians, scholars and students of the future will require access to this material in order to understand the reasons for the widespread reaction to it. Censorship can have consequences too. It has been reported (BBC, 2006) that the Swedish foreign minister Laila Freivalds resigned after ordering closure of a website that published cartoons of the prophet. Presumably freedom of speech rather than digital preservation was the main issue, but the principles were closely related in this case (the issue of integrity was also relevant). On a less dramatic note, Butt (2006) reports that copies of the Cardiff University student newspaper Ghair Rhydd were withdrawn from distribution and pulped after the editors (who had decided to reproduce the cartoons) heard about the Danish embassy in Beirut being set alight.

These recent incidents illustrate the potential delicacy of the issues. They highlight the need for digital libraries and other websites to have clear policies so they can respond quickly and consistently to requests to revise or remove material. Over the past five years the Glasgow Digital Library (Dawson, 2004) has had to deal with a variety of such requests. Most have been polite, friendly and relatively mundane, with the most extreme case involving potential legal action rather than death threats, but each one required a considered and appropriate response. The remainder of this article summaries some of these requests and the responses to them.

Example 1: Transcription error reported in ebook

Content: The ebook 'Glimpses of old Glasgow', digitised from the printed book published in 1894. Like all ebooks in the Glasgow Digital Library, this was published as standard web pages rather than as PDF or image files, which means that transcription or proofreading errors are possible (though rare).

Request: A user researching family history asked if a sentence could be checked against the original book as it did not make sense.

Response: The printed book was checked and the user was found to be correct, as three words had been omitted. The text was updated, the ebook was recreated and republished, and the user was informed and thanked.

Policy: GDL ebooks are intended to be exact copies of printed books, published as accessible and searchable web pages. Any reported errors are checked, and if necessary corrected, as soon as possible. The automated publishing process makes it easy to update an entire ebook or an individual page.

Outcome: User satisfied and library enhanced.

Example 2: Factual error reported in ebook

Content: The ebook 'Curiosities of Glasgow citizenship', digitised from the printed book published in 1881.

Request: A user pointed out a factual error in the book, commenting 'I am aware that this was published from the book by George Stewart but I am in a position to advise you that he was wrong about the details of the Mount Vernon purchase... I have copies from the original documents to prove this beyond doubt.'

Response: The book was left unchanged. The user was informed that the digital version of the book is intended to be an accurate copy of the printed original, and that changing such a matter of substance would be contrary to the library policy.

Policy: Although the policy of retaining the original text is very clear, the request did illustrate the potential value of a system of annotations to incorporate user comments, provided they were clearly identified as such. However, the library sustainability policy precludes making manual alterations to web pages, so any system for handling annotations would have to be automated. While systems for annotating digital libraries do exist (e.g. as described by Arko et al, 2006), it is not clear how these could best be applied to specific ebook sections.

Outcome: The user acknowledged that 'of course you can't alter a book' and 'an annotation would be fine'. However, an automated methodology for handling annotations has not been introduced, so the case has not yet been dealt with to the satisfaction of user or library, and is therefore still pending.

Example 3: Events supersede ebook content

Content: The ebook 'Scotland and the Antarctic', published online in parallel with paper publication, 2003.

Request: The author noted that a reference in the book to the highest temperature ever recorded in the UK became out of date shortly after publication.

Response: After some consideration the text was changed, by agreement with the author, so the ebook was updated to record the new highest temperature (38.5ºC in 2003).

Policy: As this was a recent book, intended for current educational use, accuracy was judged to be more important than preservation, so the text was changed. The distinction between current and historical works represented a change in policy, introduced as a result of this case.

Outcome: Although the ebook is now accurate, it is not clear to users that the text in the digital version differs slightly from that in the printed version. In practice this probably does not matter, but in principle it is unsatisfactory. An annotation would be a better solution.

Example 4. Misleading descriptions reported in image captions

Content: Maxton Papers: a collection of images and documents about James Maxton, 1885-1946.

Request: A user wrote 'I am concerned about both misleading descriptions of the Conscientious Objection papers, and inaccuracies in the accompanying notes.'

Response: The comments were referred to the consultant who had written the image captions. The consultant acknowledged the authority of the user, commenting 'I would be inclined to bow to his superior knowledge'. Two image captions, and one image title, were therefore changed to the form of words recommended by the user, whose name was added to the credits page.

Policy: Requests such as this had not been anticipated. A new policy was adopted of referring any such request to the author of the caption, and to update the text only if both library and user were satisfied as to the authority of the user, and if the caption author agreed to the library making the requested changes (or was willing to rewrite the text personally). As the captions were specially written for this collection to accompany the digitised photographs and original documents, the issue of digital preservation was judged to be barely relevant, and less important than accuracy of content.

Outcome: User satisfied and library enhanced. When a similar case occurred for a caption in a different collection, the newly-adopted policy was applied, with an equally satisfactory outcome. A different consultant was involved, commenting, 'we should welcome input from experts ... [and] make every effort to keep the site as accurate as possible.'

Example 5: User offers additional digitised material

Content: The ASPECT collection (Access to Scottish Parliamentary Election Candidate Materials) of digitised election leaflets and ephemera.

Request: User offered to provide several additional (digitised) leaflets that were missing from ASPECT.

Response: The relevant images were copied from the user's own website and added to the collection, and the user's name was added to the credits page.

Policy: The aim was to have as complete a collection as possible. As there were no extra copyright issues involved, and the material was easy to deal with, the offer was gratefully accepted. However, additional material on offer was not used, as it related to UK parliamentary elections or by-elections and was therefore outside the collection policy for ASPECT (which covers only Scottish parliamentary elections).

Outcome: User satisfied and library enhanced.

Example 6: User offers original article and photographs

Content: Springburn Virtual Museum: An online exhibition of photographs and related materials from Springburn Community Museum.

Request: A user offered a new photograph and a separate article with accompanying photographs.

Response: The user was asked to provide a three-paragraph caption to accompany the photograph. When this was supplied, the submissions were accepted and added to the virtual museum, and the user credited.

Policy: This circumstance had not been anticipated, so a new policy was adopted of accepting original user-contributed content, provided that library staff were satisfied about the integrity and authority of the user and the authenticity of the new material. Evidence presented by Jones (2005) suggests that this is a sound policy; he reports on the success of three archives that rely entirely on user contributions.

Outcome: User satisfied and library enhanced. Photographs from different contributors were subsequently dealt with in the same way; they were accepted if supplied with accompanying captions, as required by the collection policy.

Example 7. User requests removal of content owing to copyright violation

Content: The ASPECT collection of digitised election leaflets and ephemera.

Request: The following email was received from an election candidate: 'Please be kind enough to remove my election leaflet from your website. Copyright for this document rests with me and you do not have my permission to publish my personal data on your website.'

Response: A check was carried out on the file of letters received in response to copyright clearance requests. No letter could be found from the relevant political party. As no permission had been granted, the election literature was removed as requested. However, the election results for the relevant constituency, comprising non-copyright data, were retained.

Policy: The existing policy had been to attempt to obtain copyright clearance from all political parties and independent candidates (but not from individual party candidates) and to add material if permission was granted or if no response was received. In view of the fact that only one request to remove election material had been received, this policy was judged to be valid and successful, and was therefore retained.

Outcome: User satisfied but library diminished.

Example 8. User requests removal of old images of his work

Content: 'Sculpture in Springburn: Towards Glasgow 1990'. This physical exhibition from 1989 had been commemorated with a 50-page booklet, which was digitised to create an online exhibition (with permission from the booklet compiler and publisher).

Request: One of the sculptors featured requested removal of illustrations of his creations, as they were not representative of his current work.

Response: It was explained to the sculptor that the digital exhibition was intended to be a full record of a historical event, and that the library would be reluctant to remove any material from it. The sculptor understood the reason for archiving it but was understandably unhappy at the prospect of being judged by his early works. A compromise was agreed whereby the material was retained but a link was added to the sculptor's current website.

Policy: No policy seemed particularly relevant, other than a willingness to be flexible and compromise. It is not clear what would have happened if the sculptor had been adamant about seeing his material removed. Perhaps the images could have been retained but the text made non-indexable, so that the material could not be found via a Google search.

Outcome: User partially satisfied, library satisfied and not diminished.

Example 9: User demands removal of libellous content

The final example is drawn from the Sapiens publishing service (Wallis, 2004) rather than Glasgow Digital Library.

Content: A book review published in The Angry Corrie magazine in October 2003. The Sapiens service incorporates a complete digital copy of every issue of this magazine, dating back to 1991.

Request: The author of the book (first published in March 2003), wrote demanding that the review be removed, as the book had been reprinted in September 2004, with 'corrections of all the critical areas contained on your website'. He acknowledged that the review contained fair comment about the first edition, but threatened legal action for libel under the UK 1996 Defamation Act, writing 'now that you have been informed of this libel we require you to immediately remove the offending article'.

Response: Differences of opinion between the departmental manager, the digital content manager and the magazine editor meant that more time was spent dealing with this request than on all other cases combined. Although the request appeared self-evidently absurd, the repeated and specific threat of legal action was sobering, and the departmental manager favoured removing the review pending legal advice. This was resisted by other staff, but as a compromise the title of the review was changed to include the date (March 2003), making it clear that the review applied only to the first edition. When legal advice was finally obtained, it confirmed that there was no case to answer, and the book author was therefore informed in writing that the review would not be removed.

Policy: Unclear.

Outcome: User extremely dissatisified, archive mildly diminished, management relieved, other staff somewhat dissatisfied with the response. As legal action has not been forthcoming, the case appears to be closed. However, the retrospective title change remains in place, meaning that the digital version of the magazine is no longer an exact copy of the paper version.

Conclusions

User expectations of digital libraries are clearly different from those of physical libraries. The global accessibility of digital libraries makes them a more obvious target for comment or complaint. Users know that content updating is relatively easy, and most digital libraries are in effect publishers, as well as being libraries. It is unlikely that a physical library would be asked to remove a critical book review from a periodical, and almost inconceivable that a physical library would make changes to a printed book in response to factual errors reported by users. On the positive side, there is great potential for enriching digital libraries with user-contributed content, although this raises difficult issues of trust and authenticity, as well as methodology and sustainability. A discussion of the concept of the 'Internet trust model' (as opposed to the academic trust model), is provided by Bilder (2006), who argues that publishers 'can learn much about approaches to handling Internet trust' from the actions of large and successful Internet-based services such as eBay, Google and Wikipedia.

Users may perceive a digital library as just another website, but it should be far more than that, with collection and updating policies that reflect its nature and purpose. The above examples demonstrate the importance of adopting a consistent approach to user requests, with responses based on clear policies and underlying principles. In practice, however, policies can not cover all possible circumstances and may not help in dealing with every user request. Sometimes a new policy may be needed, or an existing policy adjusted, in response to a request, so some flexibility is useful. However, consistency of response is necessary for library credibility, as well as for keeping staff, contributors and users satisfied, so any revised policies should be adopted and applied to subsequent cases, to retain a balance between consistency and flexibility.

Where compromise and flexibility are insufficient for keeping all parties happy, it may be necessary to take a utilitarian view; failing to comply with a specific short-term request (or threat) may make a specific user (or group) unhappy, but this can be justified if there are judged to be long-term benefits for the majority of library users. Where conflicting demands arise, e.g. between accuracy and preservation, or between expediency and social responsibility, it will be much easier to resolve such conflicts, and make appropriate decisions, for digital libraries that have a clear sense of purpose as well as a clear set of principles and policies. On the other hand, faced with potential threats to career, livelihood, or even life, the most highly-principled digital library manager may be forced to remove material. Even the best policies may count for nothing in an extreme situation. The role of digital library management can be far more complex, subtle and politically sensitive than one might expect.

References

Andrews, J. and Law, D. (2004), Digital libraries: policy, planning and practice, Ashgate Publishing, Aldershot.

Arko, R. A. et al. (2006) "Using annotations to add value to a digital library for education", D-Lib Magazine, Vol 12 No 5. http://www.dlib.org/dlib/may06/arko/05arko.html

BBC. (2006) "Sweden FM quits over cartoon row", BBC website, 21 March 2006. http://news.bbc.co.uk/1/hi/world/europe/4828052.stm

Bilder, G. W. (2006) "In Google We Trust?", Journal of Electronic Publishing, Vol 9 No 1. http://hdl.handle.net/2027/spo.3336451.0009.101

Butt, R. (2006) "Hold the front page! The students have a story", The Guardian, 20 June 2006.
http://education.guardian.co.uk/egweekly/story/0,,1801163,00.html

Dawson, A. (2004), "Building a digital library in 80 days: the Glasgow experience", in Andrews, J. and Law, D. (Ed.), Digital libraries: policy, planning and practice, Ashgate Publishing, Aldershot, pp. 131-153. http://cdlr.strath.ac.uk/pubs/dawsona/ad200301.htm

D-Lib Magazine. http://www.dlib.org/

Hamilton, V. (2004) "Sustainability for digital libraries", Library Review, Vol 53 No 8, pp. 392-395.

JoDI: Journal of Digital Information. http://jodi.tamu.edu/

Jones, P. (2005) "Strategies and technologies of sharing in contributor-run archives", Library Trends, Vol 53 No 1, pp. 651-662.

Middleton, K. (2005) "Collaborative digitization programs: a multifaceted approach to sustainability", Library Hi Tech, Vol 23 No 2, pp. 145-150.

Smith, A. (2003) "Issues in sustainability: Creating value for online users", First Monday, Vol 8 No 5. http://www.firstmonday.dk/issues/issue8_5/smith/

Wallis J. (2004) "Scottish Academic Publications Implementing an Effective Networked Service (SAPIENS) project", Library and Information Research, Vol 28 No 88, pp. 27-33.