The act of registration seems ingrained in all of us, especially where dealing with ‘authority.’ That seemingly random agglomeration of letters and numbers on the back and front of our cars? A registration number. We visit the Registry Office in order to make births, marriages and deaths official. Indeed, in the dark days before it was permissible to marry between the frozen peas and the fish fingers, the Registry Office was the place to which the atheistic, agnostic, and otherwise religion-averse turned in order to conduct the wedding ceremony itself.

As with so many words, registry is one that the technology sector coopted and quickly began to render almost meaningless by adding layer upon layer of obscure nuance. For Windows users (at least back when I last used Windows), the Registry was the place you went to disable whatever was responsible for slowing your machine to a painful crawl. On the web, registry shares some of the attributes of library, directory, repository, and similar concepts; it is typically a place into which some resource, function, or package is registered for later discovery and use by others. Personally, I’ve always tended to presume that registries are differentiated from those similar terms by being intended wholly or predominantly for use by machines rather than people; a registry a place for software to find the data it needs, isn’t it?

Registries crop up regularly in discussion of curated online resources, typically when someone says “We need a registry for x” or (increasingly), “The Web is the registry for y.” To synthesise some of the different perspectives here, and to really look at whether or not the Web could be the registry, the UK’s Joint Information Systems Committee (JISC) has asked me to conduct a short desk-based project which is now well underway.

In the education/ cultural heritage/ government contexts within which JISC operates, there are plenty of ‘registries’ to look at; the venerable Information Environment Service Registry, the CKAN instance sitting behind, America’s new National Learning Registry, Australia’s Registry of Research Collections and more. But what do they have in common, how are they differentiated, and where are the similar examples from the business world? Since the decline in enthusiasm for standards like UDDI, are formal registries in decline?

I’m certainly not seeing mainstream uses of the term that seem ‘wrong’, but there’s clearly scope to develop some sort of simple taxonomy to capture the various functions for which someone might propose a registry-based solution. Having done that, I can then finish by looking at the extent to which the architecture of the Web can reliably and sustainably support those functions.

So… what do you think a registry is, and do you have any great examples of that use case? Is it best done in some centralised and organised fashion, or can resources spread out across the web simply self-assemble as required?