Ugh. No, it begins with people who have web sites realizing that if they want their site to be found (also known as SEO) then they need to add metadata. Hopefully, CMS’s and software like Dreamweaver will start making it easy to add this metadata. The metadata is then spider-able by anyone who wants to spider it, and if the data makes use of things like URIs, it also becomes linkable to data in Wikipedia, geonames, id.loc.gov, etc. Since each web site automatically has an identifier, anyone else (like your librarians) could create more data that is associated with that identifier (like making connections to VIAF). But the “it starts with librarians” is a non-starter. There are how many billions of sites on the Web, and how many that are new or that change each day? I’d just be happy if librarians would start thinking about how to make use of the micro-data that is out there today, INCLUDING the WorldCat linked data. For that latter, I have a very brief video “walk through” — for human access, not machines:
I’m now looking at the WorldCat file that was exported, and hope to have some “how to” related to that before too long.
Concerning libraries and linked data, I think we can all assume that if and when linked data really begins to take off (although such a development is still doubtful), that is when Google, Facebook, Bing, Yahoo, and the rest will dive in in such a way that libraries will be elbowed out completely and won’t have a chance. The coding used: RDF, microdata, RDFa or whatever will make no difference to the outcome.
To combat this, libraries must find a path for themselves. They can do this by making something different from what the Googles make–and what the Googles don’t want to make–and that everyone knows the public wants. One way of creating something that the public wants, and that many are beginning to demand, is to provide selection–“reliable selection” in all kinds of meaning of the term. People are starting to understand that the touted secret algorithms can always be cracked for the advantage of some group or individual (and the algorithms were never that great anyway). Certainly the idea of library selection would have to change, and those changes would lead to other changes, but at least we would be providing something that the public wants and no one else provides.
I would also hope that people would begin to appreciate reliable metadata as well. The average person doesn’t know how to make coherent metadata and the vast majority couldn’t care less about it: the “authors” added often reflect a bureaucratic need to make sure that the bosses can fill out their CVs, so those names are added to practically everything in their own departments, or they add everybody who just looked at the resource. (I’ve seen this happen more than once) An untrained person cannot even begin to analyse a subject, much less assign subject descriptors from a thesaurus in a coherent way, and I won’t even mention the complexities of LCSH. Relying on the public to assign metadata will provide something like these two examples of tags on Amazon:
Going rogue by Sarah Palin http://amzn.to/3YL2CD
Dreams from My Father by Barack Obama http://amzn.to/QyFyxR
Then with spammers, advertisers, blah blah, we would wind up with exactly what we have now. Why wouldn’t we see more of the same?
To me, believing that linked data and SEO, especially without some kind of reliable selection, will be any improvement over what we see in the search engines now is just wishing. It will be the same characters doing the same things they do now, just using slightly different methods.
I still believe that there is a lot that librarians can do to improve what the public currently has and librarians can become vital parts of the information universe, but they need to take different directions.