The Long And Winding Road

I was recently an invited speaker at Intute’s first Staff Conference, which was described in a blog post on Intute’s newly launched blog service. The title of my talk was “What If Web 2.0 Really Does Change Everything?“. Before exploring the challenges which the range of externally hosted Web 2.0 service would pose to a JISC-funded service such as Intute I took the opportunity to revisit the early days of Intute, when, in the days of the eLib programe the services were known as Subject Based Information Gateways (SBIGs), before becoming known as the RDN (Resource Discovery Network) prior to their current name.

What, I asked, was the key to Intute’s success? Was it, I wondered:

• ROADS: the open source software which formed the basis of services such as SOSIG in the early days?
• The lightweight whois++ distributed searching protocol supported by ROADS, which would allow users to cross-search across the various SBIG services?
• The MySQL database, which formed the core data management tool for ssome of the services?
• The PostGres database, another open source relational database management system, which provided richer functionality than MySQL?
• The distributed approach to development and hosting, which enabled a diversity of technical approaches to take place?

From today’s perspective, we can see that the only technical component of the Intute service from the list given above which is still critical is the MySQL database. ROADS is now festering on SourceForge and the whois++ protocol seems to have dropped off the radar screen, having been superceded by the SRU/SRW cross-searching protocols which were designed for a Web environment. And the distributed development and hosting approach has been replaced by a centralised service, hosted at MIMAS.

At the conference I argued that the success of Intute wasn’t due to the initial technical choices. Rather it was due to the effectiveness of their outreach activities, with staff from SOSIG, EEVL, OMNI and the other hubs regularly appearing at conferences, giving seminars, running training sessions and writing articles for many publications.

There was, however, one piece of technical innovation which has shown itself to be sustainable, which was described in a short paper on “RDN-Include: Re-branding Remote Resources” by myself, Pete Cliff and Andy Powell published in May 2001 in the WWW 10 Conference Poster Proceedings. RDN-include allowed the RDN service to be embedded in third party Web pages. The initial development made use of a CGI script which needed to be installed on the institution’s server. However we realised that there was always likely to be a SysAdmin barrier (”no third party script to be allowed on my server”) so a lightweight JavaScript alternative was also developed, RDNi-lite. And, as described in a post on Integrate Intute content on the Intute blog, this service is still being provided, although under a new name and using, I believe, rewritten software.

A focus on users? A lightweight approach to embedding content? This sounds pretty much like Web 2.0 to me. As I said in my talk, I think the success of Intute was due to the Web 2.0-style approach they took, before the term was coined.

But in the light of what we now know, how might Intute have developed? We can see that the distributed approach taken initially wasn’t sustainable, and the emphasis on cross-searching would have been misplaced in a more centralised model. Looking at The History of Yahoo! it strikes me that, in an alternative universe Intute could have been the Yahoo! of the planet.

We thought we were at the start of a long and straight Roman road in the days of eLib. Looking back, we can see that it was a long and winding road, and occasionally we’ll realise that we’ve been heading in the wrong direction and retrace our tracks. If we were starting all over again, which way would we go?

JISC Capital Circular 2/07: Call for Proposals

The JISC Capital Circular 2/07: Call for Proposals was announced on Friday 27^th July 2007.  This circular invites institutions to submit funding proposals for projects in the following areas:

• Enterprise architectures
• e-infrastructure
• Users and innovation

Proposals may be submitted by HE institutions funded via HEFCE and HEFCW, and by FE institutions in England that teach HE to more than 400 FTEs.

Further information is available at at: http://www.jisc.ac.uk/fundingopportunities/funding_calls/2007/07/circular0207

The “Circular 02/07 appendix F: Next Generation Technologies and Practices call” MS Word document in particular includes a variety of issues which relate to the topics which have been addressed, including the role of standards in emergent technologies, accessibility, risk assessment, etc.

I’ll discuss some of these issues in more depths in forthcoming posyts over the next few weeks and months. 

From The DNER To Web 2.0

My former colleague Andy Powell was one of the key developers of what was originally known as the DNER (Distributed National Electronic Resource) and was later rebranded as the JISC Information Environment (IE). Andy produced a diagram of the IE architecture, an early version of which is illustrated.

This diagram (and subsequent versions which further developed the initial model) illustrate how JISC’s development strategy recognised the importance of the network as a platform for providing access to services across the higher and further education communities.

I was involved in some of the early discussion about the JISC IE. And the following diagram (taken from a talk on The Web In The 21^st Century given at the JUSW 2001 workshop on 4-5^th September 2001 at Loughborough University) gives my interpretation of how the JISC IE might develop.

It should be noted that in this diagram I floated the idea that the JISC IE could be enhanced to include access to application services and not just middleware services such as authentication. It is interesting that my vision was for access to lightweight services such as spell-checks and bookmarking services. The idea came to me after reflecting on services such as HaL’s Web-based HTML validation service which was announced way back in 1994 and was subsequently mirrored on the (now defunct) national HENSA mirror service. It struck me back then that this concept (based on simple REST interface) could be applied more widely.

Back then I didn’t envisage that it would be possible to deploy networked versions of full-scale applications such as a word processor. But this is now available, as the Google Docs service (and many other competitors) clearly illustrate.

I also did not foresee that the service we use within the higher and further education communities could be provided by the commercial sector. But del.icio.us, and many other social book marking services, also clearly demonstrate that the model of networked access to bookmarking services, which I suggested in my diagram, can be deployed on a global scale.

On reflection I think the vision for the JISC Information Environment, which was devised and developed by UKOLN and JISC colleagues including Andy Powell (who now works for the Eduserv Foundation) and Liz Lyon (UKOLN) and Rachel Bruce (JISC), can be seen as an architecture which has strong connections with Web 2.0. The JISC IE vision, however, probably missed out on the importance of social networking and user generated content and, indeed, generating interest which will encourage users to adopt new technologies (indeed, as Andy Powell commented recently “One of the … problems with the JISC IE diagram is that it was largely technology driven“). But the initial technical architecture that was devised (especially syndication using lightweight technologies such as RSS) seems to have been validated by the success of Web 2.0.