The Auricle

by Derek Morrison, 3 September 2008

Over on the JISC-PoWR (Preservation of Web Resources) site there’s a posting entitled Auricle: The Case Of The Disappearing E-learning Blog (1 September 2008). Yes, that’s right this blog is the ‘disappeared’. While it’s a new experience becoming such a case study the author, Brian Kelly certainly raises some interesting issues. Although I’ve already contributed a comment on the PoWR site in today’s post to Auricle I’m taking the opportunity to expand on the points I made there.

In his posting Brian describes how ghostly traces of Auricle continue to exist as a combination of past citations and dead hyperlinks to a now defunct University of Bath web site. Ironically, one such dead hyperlink in Brian’s article was an extract from an old Auricle posting titled The Weblog as the Model for a New Type of Virtual Learning Environment? (Auricle, 25 Feburary 2004). Readers of today’s posting will find the hyperlink on this site is live. In my 2004 posting I posited that blogs could “form the basis of a distributed, not centralised, information and learning object system”. It’s always an interesting exercise to revisit what I’ve opined and cited in the past and then reflect on whether and how experiences since have impacted on what I opin and cite now.

In my 2004 posting I quote the following:

“Portal and KM vendors could learn a few tricks from emerging technology segments like RSS, RDF, and the blogging community. These initiatives have stumbled upon [what I consider] the single most important aspect of network dynamics – the discrete addressability of information objects.”
http://myst-technology.com/mysmartchannels/public/blog/5936

“e-mail is where knowledge goes to die � In most companies annotations and observations are typically created in e-mail with some messages containing links that point out to specific information objects relevant to the message. Aside from the message itself, the knowledge dies a slow death in the inbox of office workers and executives. Creating a process so that annotations and business observations may live as uniquely addressable information objects, clearly has greater advantages; especially for portal users.” (ibid)

In 2004, I was was sympathetic to the “email is where knowledge goes to die” argument. In 2008 I’m convinced that email is an easy to use, easily abused, and increasingly dysfunctional, approach to communication; and at times a major impediment to reflection, knowledge acquistion and storage. But that can be the basis of another posting 🙂

In relation to Brian’s JISC-PoWR posting, what I want to focus on is the blog et al as representing one manifestation of “the discrete addressability of information objects”.

In 2004, I was fairly happy with the concept of the blog posting as an information object. Each blog posting (at least in the way I conceived Auricle postings) was meant to be a mini (and sometimes not-so-mini) quasi-essay with standard blog architecture automatically providing each posting with its own discrete address (or permalink). Discrete addressability, however, applies only within the context of the domain in which the blog is hosted and implies nothing about sustainability. Once the “lights are switched off”, i.e. the blog ceases to be hosted by the domain, the information objects (and their discrete addresses) will permanently disappear unless steps are taken to mitigate the effects.

There are a few things to consider. First, why are the lights turned off? As I indicated in my comment to Brian’s post the main reason is that without one or more champions to defend, maintain or enhance online artefacts like Auricle the, usually inflexible, IT policies of the institution will dominate thinking in the original host domain. Consequently, although there may well be a high value placed on the online application or service by external stakeholders, the hosting domain may well perceive blogs, wikis, social networking sites etc as now obsolete accounts in the same way an email account will be terminated once the owner(s) leave an institution. From a central IT support department perspective, there will usually be no assessment of potential ongoing value, but simply a technical housekeeping operation to take place. In fairness to the University of Bath they gave me plenty of warning that they would be terminating my various accounts and that I should relocate what was of value to me, which I have duly done.

Relocation, however does not solve the discrete addressability problem particuarly when moving from one blogging platform, e.g. pMachine to another, e.g. WordPress ,which use different internal addressing systems and which lack an easy migration route. For example, I tend to use lots of internal references to other Auricle postings but of course the permalinks have now changed from two perspectives, i.e. both the hosting domain is different and the post id have changed. Consequently, there is a lot of work still to do to ensure the integrity of both internal and external references. This will be problem mainly of the legacy postings imported from the University of Bath domain with the increasing quantity of more current postings hiding this issue because they will have up-to-date internal links. So ironically, although, as long-time Auricle readers will know, I am no fan of putative ‘solutions’ that ‘lock-in’ or, to use the current vernacular, ‘tether’ users to one system or device but I would now include “migration support” as a key attribute when choosing a blogging engine in the future. pMachine was a good blog engine in its time but migration support was effectively at the level of RSS export and import to WordPress. The result? Each new information object was given a new discrete address and all internal links and some external links broke as a result.

My migration experience has also prompted me to reflect on the nature of my internal linkages.

Purists may argue that an information object should not have any such dependencies and that internal linkages are thus compromising the standalone nature of the object. I argue, however, that whilst an internal link may break and that while that is unfortunate their should still be sufficient value in the content of that specific information object to make it a worthwhile read/viewing on its own. Also, the blog at least provides a search and filter mechanism so that the internal reference can be rapidly located, albeit with not quite the convenience of the hyperlink. Readers can help this author by notifying him of any such broken linkages in this blog (not the now obsolete pMachine version of Auricle which is also archived in this domain). While acknowledging the debate relating to the use of absolute or relative URLs in blogs after my migration experiences I would, on balance, opt for the use of relative URLs since that should reduce the amount of work when migrating the blog to another domain. In a relative URL the blog engine should complete the web address and while that appears to be true for links in posts it is not necessarily true in other parts of the WordPress environment, e.g. the WordPress links functions assumes the URL is always absolute whereas I use this facility a lot to support easy access to key internal postings in my Academy blogs.

So Auricle did survive. I am, however, guilty of having done little to promote its new domain, to date, mainly because I have also been supporting a multiplicity of other blogs in my work for the Higher Education Academy. Included in my new portfolio of blogs are:

• Benchmarking of e-Learning Exercise
• Pathfinder Programme
• Gwella (the HE Wales Enhancement of Learning through Technology programme)

Auricle survived because I am an adherent of the ethos (if not the technical architecture) behind Lots of Copies Keeps Stuff Safe (LOCKSS) ethos. LOCKSS explores the application of technical solutions to create self-healing or self-repairing systems based on the distribution of data across multiple servers in a way similar to local PC or server RAID systems (Redundant Arrays of Inexpensive Discs). Similarly, the work of the Navigators (distributed systems research team) into Fault and Intrusion Tolerance in Open Distributed Systems at the University of Lisboa, Portugal focuses on self-healing/repairing approaches. Unlike approaches based on keeping intruders out via firewalls the Navigators team believes it is better to assume intrusion and potential damage will occur and focus on approaches based on redundancy and self-repair.

I apply the above principles in a rather more prosaic way by preserving the designs and databases associated with Auricle in multiple alternative sites and then using this redundancy to rebuild Auricle if it proves necessary.

But what could the alternative scenarios have been?

1. In much the same way as establishing an electronic journal the ideal would have been for the hosting institution to have sufficiently valued Auricle to have continued and enhanced the work. The reality was, however, that whilst some internal personnel were happy to utilise the output of the small team who contributed to Auricle not many were willing to become contributors on a frequent basis. Also the team were e-learning specialists who were absorbed into an amended University structure. In the end no one was left to keep it up-to-date or who could analyse its value. Also, whilst I enjoy reflecting, analysing, and writing, not everyone does. Expecting people to do can place unwelcome demands on their time although I could argue a case for appointing people in some support roles who have demonstrated their willingness and ability to reflect and communicate online using tools such as blogs and wikis, etc.
2. A variant on the above would have been a scenario would have been the emergence of either a self-sustaining University-wide internal community or to open the blog up to interested individuals from other institutions and across a range of disciplines. I suspect this may be rather easier now but in 2004 blogs were a relatively new vehicle which were at best tolerated and at worst were treated with deep suspicion. In my posting Psst … want to ‘see’ a good podcast? (Auricle, 15 June 2008) I highlighted research work that suggests that even major, apparently egalitarian, enterprises like Wikipedia are actually driven by a relatively small cadre of active contributors in comparison to the mass of readers, e.g.

“… only one tenth of one percent of Wikipedia editors account for about half of the content value of this uber Web 2.0 exemplar… So far from the “widsom of the crowd” a significant proportion of the value of Wikipedia appears to derive from the contributions of a relatively small proportion of authors/editors, i.e. the wisdom of an ‘elite’? 🙂 “

3. Had a central agency like JISC offered a blog hosting service with guarantees of editorial control and sustainability for authors then I would have seriously considered that route. Although reasonable Acceptable Use Policies (AUPs) would be necessary for this approach to suceed editorial control would need to be invested in the blog owners so that content remained uncensored even if expressing potentially unpopular but legal opinions, e.g. criticism of funders or agencies. Where the contributors to Auricle were based would then not have mattered. I often cite the University of Harvard Acceptable Use Policy as providing a good reference model for statement of policy which facilitates, rather than inhibits, technology use, while still setting clear boundaries for behaviour. From a technical and usability perspective, the commercial sector already offer useful rapid activation personal blog models on which such a JISC service could have been based, e.g. wordpress.com.
4. I could have opted for one of the ‘free’ services which were emerging at the time. I eschewed this approach because of the risk associated with being dependent on a service which traded easy creation of a blog on one day for absolutely no guarantee that the service and all its data would continue to be available on the next. Nevertheless, while the Auricle concept could certainly have been tested first on a free blog from blogger.com, wordpress.com, or pbwiki.com etc, the functional and presentation constraints would still have left me desiring the flexibity that an installation under my direct control provides. Interestingly, a few of the Pathfinder projects found it difficult to negotiate a local blog installation in their institutions but having gained experience and proved the case from the use of a free service have now managed to persuade their institutions of the value of a hosting and supporting their blog engine.

Anyway, thanks to Brian for hunting down the ‘disappeared’. Hopefully his posting will help make Auricle a little more visible than it has been of late.