Tony Ross, writing in the current issue of Ariadne, Lost in the JISC Information Environment, gives a nice summary of some of the issues around the JISC Information Environment technical architecture. He hits a lot of nails on the head, despite using that diagram twice in the same article! On that basis he seems anything but lost. That said, one is kind of left feeling "so what" by the end.
The IE does not, can not, have existence. The term is a description of a set of interoperable services and technologies which have been created to enhance the resource discovery access and use for users in HE/FE; it exists to aid conceptualisation of this ephemeral subject. No more, no less.
Well hang on, either it exists or it doesn't (that paragraph says both!) but the creation of services tends to indicate, to me, that it does. Whatever... it's an angels on the head of a pin type of discussion, best moved to the pub.
In his response, Paul Walk suggests that all models are wrong, but some are useful, the JISC IE architecture being one of the useful ones, and broadly speaking I agree, though one might argue that the prescriptive nature of the architecture (or at least, the prescriptive way in which it has often been interpreted) has got us to a place where we no longer want to be? And, leaving the diagram to one side for a moment, the technical standards certainly were intended to be prescriptive. I can remember discussions in UKOLN about the relative merits of such an approach vs. a more ad hoc, open-ended and experimental one but I argued at the time that we wouldn't build a coherent environment if we just let people do whatever the hell they wanted. Maybe I was wrong?
Referring to the "myth" quotes in the Ariadne article, I don't have a problem with trying to be prescriptive but at the same time recognising that what ends up happening on the ground may well be different in detail to what the blueprint says.
Looking back, I do find it somewhat frustrating that the diagram came to epitomise everything that the JISC IE was about whilst much of the associated work, the work on UML use case analysis for example (which was very much focused on end-user needs), largely got forgotten. Such is life I suppose? But let's ignore that... the work certainly had impact, and by and large it was for the good. Think about when the DNER effort first started, way back at the end of the last century (yes, really!), a time when any notions of machine to machine interaction were relatively immature and not widely accepted (certainly not in the way they are today). The idea that any service provider would care about exposing content in a machine-readable form for other bits of software to consume and display somewhere else on the Web was alien to many in the community. Remember Lorcan Dempsey talking about us needing to overcome the information brandscape? :-)
If nothing else, the IE architecture helped contribute to the idea that there is value in going beyond the simple building of an HTML Web site. In that sense, it had a Web 2.0 flavour to it well before Web 2.0 was a gleam in anybody's eye. The world has come a long way since then... a long, long way. The IE architecture got things wrong in the same way that most digital library activities got things wrong - it didn't anticipate the way the Web would evolve and it adopted a set of technologies that, with the exception of RSS, were rather non-Web-friendly in their approach (OAI-PMH, Z39.50, SRW/SRU, OpenURL and so on). The Web Architecture, the Semantic Web, Web 2.0 (and in particular the emergence of the Web as a social environment and the predominance of user-generated content), REST and so on never really got a look in - nor could they, since the work on the JISC IE came too early for them in many ways.
With hindsight, the appearance of registries down the left-hand side was probably a mistake - what we missed, again, was that the Web would become the only 'registry' that anyone would need. But it is the largely exclusive focus on resource discovery through metadata rather than full-text, as though the Web was a library of physical books, that is the JISC IE's most harmful legacy - a legacy that we still see being played out in discussions around open access repositories today. If I've done harm to the community through the work on the JISC IE, then that is where I feel it has been worst. Remember that in the early days of the JISC IE the primary aim was around the discovery, access and use of commercially valuable content that was not being exposed (to anyone) for full-text indexing, so the initial focus on metadata was probably excusable. Unfortunately, the impact of that design choice has now gone well beyond that.
The addition of the 'indexes' box to the diagram (it wasn't in the original versions) was recognition that Google was doing something that the IE could not - but it was too little, too late - the damage had been done. That's not to say that metadata doesn't have a place. It certainly does. Metadata is about much more than resource discovery after all, and in any case, it brings things to resource discovery that are not possible with full-text indexing alone. But we need balance in the way it is adopted and used and, looking back, I don't think we properly had such balance in the JISC IE.
Towards the end of his blog entry Paul says:
Turning to the reworked diagram which Tony offers at the end of his piece - I presume this is not offered too seriously as an alternative but is, rather, meant simply to show an ‘non-deterministic’ version. It is interesting that this version seems to miss what is, in my view, the most important issue with the original, in the way it simply copies the same depiction of the client desktop/browser.
That diagram was created by me, initially for a small group of JISC-people but then re-used in the presentation that Tony cites. It originally had the caption, "what the user sees" and was preceded by the usual diagram with the caption, "what the architecture says". So yes, some humour was intended. But the serious point was that every variant of every box on the diagram necessarily offers a human Web interface, irrespective of whether it also presents a machine-interface, so the user just sees a Web of stuff, some of which is joined together behind the scenes in various ways.
As to that "client desktop/browser" icon!? Yes, it's use was somewhat simplistic, even at the time - certainly now, where we have a much wider range of mobile and other client devices. But as with the rest of the diagram, there was/is a tension between drawing something that people can easily engage with vs. drawing something that correctly captures more abstract principles.
On balance, I think the UK HE and FE community is better off for having had that diagram and the associated work, around which a useful and significant programme of activities has been able to be built by the JISC, as described by Paul. Does the diagram remain useful now? I'm less sure about that tbh.