I’ve previously commented on this topic and think I’ve made my views fairly clear. However, it’s comforting to know that I’m not alone in questioning whether IPSV serves any useful purpose.
I picked up a report on a recent meeting of local authority webmasters and managers held in Birmingham (England), where most present appeared to conclude that IPSV, now the official Government Metadata Standard, served no useful purpose and should be ignored and not implemented. As a delegate at the meeting pointed out, search engines don’t use government metadata – at all. IPSV is not used. So really then, what is the point? The report went on to say that delegates wree not sold on formal taxonomies for websites, and definitely not centralised taxonomies like this Vocabulary. The ability to produce a site map from a taxonomy is a benefit, but a fringe benefit at best. Current thinking appears to be arriving at a much less prescriptive model of metadata, in which, rather than forcing editors to select a term from a taxonomy for the area of business the content relates to, they’re provided a ‘finger buffet’ of metadata to choose from, including schemes for geographic, demographic, subject (i.e. topic) and business tags. (I think we’re heading into the realms of folksonomies here).
However, this will no doubt upset those that want to see a tidy hierarchical view made possible by a formal taxonomy, but does that matter if it provides vastly richer possibilities in terms of interrogating and presenting content?
One of the central selling points of LGCL and now IPSV was the broad view of government services relating to a given term (e.g. see everything that all government agencies and local authorities have on animal welfare), but the reality is that’s both a pipe-dream and of absolutely no use to the vast majority of users. A broad view of government services however that relate to, say, a single mother, under 30, recently made redundant, with children under 5 living in Birmingham would be of real value. This is only possible though if one gets away from thinking like librarians and stops trying to neatly categorise every single one of a council’s services and information nodes based on universally understood terms.
Steve, many of those words are mine, and I’m pleased to see them getting an airing here. The debate about the value of IPSV or any other centrally prescribed, subject-based taxonomy needs to be had.
We (local gov web managers) have been expected to invest time and energy in implementing and maintaining the LGCL and now IPSV on our sites, for nothing more than a promise of riches to come. We should be questioning why we’re making that investment.
Just to clarify, folksonomies are not what I’m suggesting we should be pursuing. It would make an interesting experiment to have our customers tag and describe our services, but the hard part would be finding a sufficient number of willing participants. I’m more interested in describing our information in ways that make sense to our customers. That requires a structure and a degree of centralisation that a folksonomy doesn’t offer.
I’m also glad to see this issue raised. A few observations:
It would be interesting for local authority websites to use folksonomies and tag clouds – this would capture the search terms of the *users*. There is no substitute for this; you need real users rather than surrogates. It would be a good start for every local authority to have to publish the top 50 terms entered into their local search engines on a monthly basis. This would be a great way to ensure the content of the website matched what people were looking for, in the language that they use.
During the IEDISS project we did workshops with real users. One of the things we asked people to discuss were terms used by councils such as ‘scrutiny’ ‘customers’ – people didn’t have a clue what they meant in a local government context. In fact most of the participants were confused as to what the difference might be between ‘local authority’ and ‘council’ with some people thinking they were different things.
Categorisation in itself can have some uses – e.g. in forming an online directory where top level categories (in conjunction with other information retrieval tools like search engines) can give the user another option to find things. The problem I’ve always felt with IPSV is that it has become a hybrid of a thesaurus and a taxonomy. The two things are different and don’t work if merged in the way IPSV tries to.
From a practical point of view; whenever I finish a document and need to add keywords or a ‘subject’ definition, I rattle off the first 10 that come into my head, based on being the actual author of the content. I have tried to use IPSV and generally stop at the first term that remotely matches what I want. The overhead in terms of time and thinking is too great in trying to use a list which has too many terms (see point above about folksonomies).
But not to throw the metadata baby out with the categorisation bath water. More metadata would generally be a good thing in local government – it would be a big improvement in information management if every information resource produced by councils had at least the author, date of publication and publisher (e.g. Trumpington Council). As more content is shared and syndicated, those bits of metadata become more important.
Concentrating so much on categorisation as the key metadata element has always been, in my opinion, a mistake.
“It would be a good start for every local authority to have to publish the top 50 terms entered into their local search engines on a monthly basis.”
An interesting idea Danny, but without the context of knowing what the user was acutally looking for it would be of limited value. Here are the top 20 searches in the last year on the local authority site I’m responsible for:
vacancies, jobs, google, Population, job vacancies, council tax, school holidays, yahoo, schools, grants, employment, teaching vacancies, alva, recruitment, tulligarth, alloa, speirs centre, school term dates, alloa academy, swimming, local plan, term dates, building warrant, alva academy, dollar, tenders, planning, map, tullibody, teaching jobs
To do what you’re suggesting we would have to make assumptions about what these users were looking for, and in doing so would immediately contaminate their world view with our own. The search term only gives us half the picture.
It also ignores the long tail. Those top 20 terms account for approximately 20% of the total searches made. We’ve already optimised for those 20%, but what are the other 80% searching for?
The exercise you describe at IEDISS would provide much more usable data, but it’s labour-intesive and its success would depend on getting a broad range of participants who are willing to spend an hour or two describing how they perceive our services. Watch out for the rush!
I think what I had more in mind was the idea of tag clouds – for example http://del.icio.us/tag/
There is a very small overhead (in setting this up on the site) and it captures the terms being used by the users – in their language. Which is useful for creating synonyms for search engines (if the local search engine can be configured to deal with synonyms).
I’d be interested to hear of any local authority web sites which are using tag clouds or folksonomies.
Danny,
I intend using tag clouds/folksonomies for the IDeA Community of Practice site (www.communities.idea.gov.uk). I spec’d this as a requirement back in November and expect the functionality to be available by April this year. I’m not aware of any local authority sites using tag clouds.