You are currently browsing the monthly archive for October 2009.
One of the good things about being at an HE institution is that you get the opportunity to find out first hand about some of the research that is going on. Often that provides a fascinating glimpse of a range of new ideas and concepts.
This week I had the chance to see the current state of play with some research into methods of using images to support learning. The research took a collection of photographs, loaded them into a learning system and then asked learners to categorise them against a set of criteria. Interestingly the system used that data to visualise the results and feed back to the learner. This process helps the learner by getting them to identify and explain patterns from the data.
A couple of prototype systems were discussed that used Magic Studio software from Lexara and images from the Bridgeman Art Library. The first system used 19th and 20th century images to investigate the role of women. The system was tested in four schools with 12-14 year olds being asked to consider and categorise each of the images. Although it seemed to have worked well as a learning activity it did flag up the need to provide more in the way of support in interpreting the images and identifying patterns.
A second prototype was developed covering aspects of sustainability. In this example a set of images of buildings was used and learners had to tag them using a set of keywords and then answer a series of questions in a quiz about their attitudes to sustainability.
Learners were also asked to plan a town using the various buildings and were then able to look at factors such as the environmental and economic impact of the decisions they had made. They could also compare with other people in class and collaborate.
The final part of William Kilbride’s seminar talked about some research into content created as part of JISC projects. Questions included the length of time that the data created as part of the project would continue to be made available. There were a wide-range of responses, some of which were quite alarming (and quite short). Although JISC require data to be made available for 5 years, many have seen that as the maximum period of time and not the aspiration that the data should continue to be available.
Other questions that the survey asked included: ‘Where is master copy of data and metadata?’ – ‘How long will content be available?’ – ‘How it will be created, maintained and preserved?’
The emerging conclusions from the study include:
- a written preservation policy is critical
- format is not specifically an issue
- good content management is good content management
- the size of an organisation isn’t a measure of success – but partnerships are crucial
- the synchronisation of content between delivery service and archive is hard
- projects focus on users – which is a strength
- the programme has driven institutional development in preservation policy
William ended with a few final thoughts including ‘Our legacy will be what we make it – it cannot be taken for granted’ and ‘Digital Preservation is not impossible’
All in all it was a useful run through the issues around Digital Preservation. It was interesting that one of William’s previous roles was as an archaeologist and there was some discussion over what a ‘Digital Archaeologist’ or maybe a Forensic Archaeologist might do in the future.
Given the continuing delays in implementing the Legal Deposit of digital media and the apparent loss of a proportion of material digitised through JISC projects it is a concern how much of the digital record is likely to be preserved. Add to this the increasing amount of data that is locked up in proprietory electronic databases, it starts to point to a very real possibility that future Digital Archaeologists will find very large gaps in the written record. Although it is clearly impossible, impractical and possibly even undesirable to try to capture a complete digital record it seems quite likely that there will be a much smaller proportion of the digital world being preserved than we find with the modern printed word.
This starts to make me wonder if a future Digital Archaeologist might have much more in common with a Dark Age archaeologist than we might expect, or hope.
William Kilbride’s lunchtime talk on the Digital Preservation Coalition today. William is from the Digital Preservation Coalition (sounds like a surprisingly good name for a Rap artist!), based in Glasgow. www.dpconline.org His talk gave a good run through the work of the DPC and was a good primer on some of the issues around Digital Preservation. A useful follow-up to some of the ECDL sessions where digital preservation was touched upon.
The DPC’s statement ‘Our digital memory accessible tomorow’ is a great aspiration and William explained how the DPC was going about trying to help to support and encourage making that a reality. Describing long term preservation as ‘trickier than we expected’ he first flagged up four issues:
- the growth in digital data
- the complexity of data
- the requirements of data
William illustrated the problems with a couple of great examples. 60% of the URLs in Hansard reports between 1997 and 2006 no longer work. And, the tapes from the Viking lander had to be recreated from a bitstream copy as the original tapes weren’t usable.
He outlined five good reasons why long-term access was required – regulatory requirements; financial risk; opportunity cost; reputational cost and heritage loss, and gave the example of NOFDigi, where only a proportion of that data (created at significant cost) was still available.
The DPC see their role as enabling and agenda-setting. They have a broad range of institutions involved from Universities to the BBC, to various research organisations, Museums, JISC and the MLA. In their enabling role they run the Digital Preservation Training Programme and produce Technology Watch reports, for example. In the Agenda-setting role they have been involved in reports including ‘Mind the Gap’ and ‘Digital Britain’.
William identified seven long-term challenges.
- He saw that ‘Digital information has little inherent meeting but is only rendered comprehensible by a combination of technologies and skills’. Access depends on configuration of hardware and software and the skills of the operator. Mitigation strategies are to document the configuration, use emulation and migration to access the content.
- Technology continues to change which creates obsolescence. Examples include changes in file formats. Mitigate by watching for changes that would lead to file formats becoming obsolete. Use tools such as Pronom or Droid from the National Archives to help with analysing which file formats are present on the network. Migration and emulation strategies can help
- Changes in storage media, e.g. the Domesday disk. Mitigate by refreshing and using slef-check methods
- Digital Preservation systems are also subject to obsolescence. Mitigate with standards such as OAIS and be modular
- Data being altered, corrupted or deleted. Mitigate with security and checksum tools
- Digital resources are intolerant of gaps in preservation. He suggested that automation and economies of scale have a role to play.
- There was also limited experience and rapid technology change.
It was commented that there was another issue around deciding what should be preserved. Issues around how much risk you would accept might determine your preservation strategy. One interesting comment was that a recent JISC report had noted that ingestion cost 70% of the overall costs. Although storage might be cheap to buy the environmental costs could be high.
One of the interesting follow-ups was that the difference between born digital and digitized content was that with born digital content you don’t have the original source material.
I recently had the opportunity to go to the European Conference on Digital Libraries. This three-day conference, held in Corfu this year, has been running for a number of years and is of interest to a range of different disciplines, from computer scientists to librarians and archivists. This year there were around 400 attendees, mainly from Europe but with a sprinkling of attendees from Asia and the Americas. With two keynote speakers, 30 papers and a range of posters and demonstrations the conference covered a good deal of ground and I’m going to put down a few of my impressions.
ECDL is quite scientific and academic in approach. Although the conference sub-title was ‘Digital Societies’ most of the papers approached the subject of Digital Libraries from quite a narrow technical view, investigating areas such as DL software performance, semantic search techniques or the impact of data-loss on images. A significant number of the papers were from Doctoral Students and many others outlined the latest state of research by groups of researchers.
‘The days are past when scholarly authority alone determines what is saved, learned and used’
The two keynote papers from Gary Marchionini and Peter Buneman provided two quite different perspectives on Digital Libraries. Marchionini’s paper flagged up how social networking is likely to impact on future Digital Libraries whilst Peter Buneman concentrated mainly on the need for Digital Libraries to reach out to the research communities and offer them a way of safeguarding their research materials.
Both saw the long-term storage cost of DLs as a significant issue and identified preservation as key issues for the future. Marchionini identified key differences between curated and community DLs, whilst Buneman concentrated on the importance of curation of research material as a key challenge.
Other conference papers
The conference also had a number of interesting sessions including a paper on the performance of DL systems – Fedora, DSpace and Greenstone – building large collections of material and stress-testing the systems to analyse their performance. Some of the papers presented covered quite complex technical solutions and models or methods, e.g. such as using visualisation or sound clues to aid search and retrieval.
Although the conference sub-title was ‘Digital Societies’ on reflection few papers other than the keynotes really got to grips with much beyond the technical issues surrounding Digital Libraries. The impact of Web 2.0 was touched upon but much of the content was still around building and managing monolithic database structures to contain vast repositories of data. Whilst that is a challenging activity in itself, the domain perhaps needs to move to a wider discussion of the impact of ‘Digital Societies’ It is possibly with this in mind that there are moves planned to change the name of the conference from 2011.