Tuesday, 8 April 2008

Seriously Seeking Significance

Yesterdays event at the British Library on Significant Properties in digital objects was a real gem. The programme was packed to the hilt, with eleven different speakers plus 45 minutes for audience discussion at the end of the day. It was a fantastic opportunity to hear more about the SigProp research that JISC has funded, as well as how significant properties are being explored in a number of other contexts such as in the PLANETS research into preservation characterisation.

Andrew Wilson started the day with a great and very targeted keynote that explored the background to significant properties research and what significant properties means to him. I was heartened when he almost immediately kicked off by talking about authenticity and significant properties, because authenticity is a big deal for me insofar as preservation is concerned! But this isn't a post about what we mean by authenticity in a digital environment, I shall save that for another day/time/place. Hot on his heels came Stephen Grace from CeRch, with an overview of the InSPECT project and future work, closely followed by David Duce presenting the conclusions of the Vector Images significant properties study. Interestingly, this used a slightly different measure of significance to the InSPECT project, which used a scale of 0 - 10 whereas the vector images study used from 0- 9. This imay be a minor deviation at the moment but might well assume more relevance when dealing with automated processes for migrating and assessing collections of mixed object types. Mike Stapleton then presented the results of the Moving Images study and, after a short break, on came Brian Matthews with the results of the Software study and Richard Davis with the results of the eLearning Objects study.

It quickly became clear that there was some variance in people's understanding of significant properties, particularly when one speaker stated that for different preservation approaches they would need different significant properties to acheve the desired level of performance. This is different to how I perceive significant properties. For me - and for several other speakers - significant properties define the essence of the object and are those elements that must be preserved in order to retain the ability to reproduce an authentic version of the object. To select different significant properties based on a given preservation approach surely means there is a diferent underlying understanding and use of the term significant properties.

What does this all go to show? That it's a case of different strokes for different folks? Well, to some extent, yes. It was widely accepted several years ago that different sectors had different requirements insofar as preservation was concerned - I remember attending an ERPANET workshop in Amsterdam in 2004, for example, that clearly illustrated just this point. And yesterday's audience and speakers represented an array of sectors with different requirements for preservation. So the whole concept of significant properties and use of the term across different sectors is something that I think we'd benefit from returning to discuss some more.

The afternoons sessions were, I think, intended to put a different perspective on the day. We heard from the PLANETS project, Barclays Bank, the DCC SCARP project, the SIGPROPS project from Chapel Hill, and a presentation on the relationship between Representation Information and significant properties. Cal Lee's presentation (SIGPROPS) on preserving attachments from email messages was fascinating and I suspect I'm not alone in wishing we'd had more time to hear more from him, but there simply wasn't the time. I suspect we would have had much more discussion if the programme had been spread out over two days - the content certainly justified it.

As always, the presentations will be available from the DPC website in due course. Keep an eye out though for the final report of the INSPECT project - they're not finished yet due to the change from the AHDS to CeRCH, but I expect it will be a fascinating read when they are done.


  1. Hi Maureen

    Indeed, this was cause for much discussion with regard to Learning Objects - I don't think it's any secret that we emerged with a much clearer understanding of LOs than SPs. I think the Sig Props concept is a very valuable approach to analysing requirements, but - however one interprets it - its application must have bounds.

    Much of the SP literature focuses on relatively easy, record-like targets - structured text, emails - that have been chewed over for a long time now at forums like Monday's. Two of the sibling studies focused on well-scoped classes of digital object whose names alone suggest where their essence might lie (moving images, vector graphics).

    Learning Objects may be any of those sorts of things (everything we found in JORUM was) and that's why considering their preservation in a purely abstract way, free of any context of who wants these things preserved and why, is something we struggled with. Without that context, why not simply preserve a Learning Object according to the generic digital object types/file formats of which it is an instance? That's why (like the software study) we didn't set out a finite, countable list of SPs, and instead just indicated where we felt further significances might lie when considering a collection of LOs destined for The Archive.

    The example that I suggested at the end - the electronic equivalent of an exam board archive of both question and answer papers - gives me, as an implementor, something more substantial to grasp (including the authenticity angle). I only wish I'd thought of it sooner!

  2. Hi Maureen,

    It was a really good workshop, thank you all for organising it. I too found the remarks on authenticity interesting. It's a concept that works differently for archives, it seems, compared to museums, but nevertheless Andrew Wilson's argument that it's essentially about the (transparent) integrity of the information being communicated is useful and clear.
    I piped up in the Q&A and I think failed spectacularly to make clear the point I had in mind, not least because I perhaps make a distinction between preservation and sustaining that not everyone would agree with. As a consequence I think that SPs, although they're being developed in the context of "preservation", may turn out to be very useful for the management of "living" resources (troublesome though this distinction is), but that we may find that we need slightly different SPs in that context.
    I've tried to work this through a little more on my blog, together with a few other notes on the workshop. Here's the post.


  3. ... a presentation on the relationship between Representation Information and significant properties

    It occurred to me at the time that although there is a relationship between these two things, there is also one important (I almost said 'significant') difference. Sigprops are indeed about some underlying essential characteristic of the work, using the terminology Chris referred to in his earlier post. They should be independent of its manifestation in some particular file format. But it's inescapable that Representation Information will be closely bound to the manifestation as well as to the work, and that it (the repinfo) will change on every migration, whereas the sigprops should not change if we get things right.

    As a simple example, the sigprops of a simple document will include the text, the order in which the text appears, and possible stylistic differences between headings, subheadings and body text (without being specific as to how to render those stylistic differences). The RepInfo cannot ignore the fact that the document is in PDF, or RTF, because that is essential to allow one to access the current manifestation of the object. But in most cases, the PDFness or RTFness of the object is not significant in and of itself.

  4. Thanks everybody for your comments. We've obviously still got a lot to discuss and pin down WRT significant properties. I hope that we'll have the opportunity to return to the subject in a more targeted setting in the future, perhaps through the future research opportunities that Neil alluded to towards the end of the workshop.


Please note that this blog has a Creative Commons Attribution licence, and that by posting a comment you agree to your comment being published under this licence. You must be registered to comment, but I'm turning off moderation as an experiment.