I’ve no wish to defend this mindset, or some librarians for that matter (in fact I can think of several I could happily poke in the eye with a sharp stick), but it is worth noting that the National Library of Australia has embraced Flickr, setting up two groups, PictureAustralia: Australia Day and PictureAustralia: People, Places and Events in order to get more images of contemporary Australia into PictureAustralia, their searchable database of Australian pictorial collections. Metadata for PictureAustralia is harvested from Flickr in the form of titles, tags, descriptions and EXIF data. There were concerns about the quality of this metadata (legitimate concerns, as images with crappy or non existent metadata will be a lot less discoverable in the PictureAustralia environment), so the National Library posted guidelines and examples in the groups’ descriptions. There has been a noticeable improvement in the quality of metadata on images submitted to these groups subsequently, indicating that people are willing to follow guidelines (if they see a return for doing so I guess).

I’m an advocate of so-called social tagging, and dislike prescription, but providing minimal guidelines on tagging in the Flickr environment, for example, which people are free to follow or ignore, might improve the user experience.

i didn’t mean “break” as in “it all falls apart”…but a wrong tag will show up as a “blip” when searching for that tag. if done systematically enough, on a large enough scale (e.g. a whole group of people aggressively tagging their own shots wrong and going around en masse to add wrong tags to the shots of others – maybe even with a nice script that takes advantage of the flickr API), it can make those particular tags completely useless. however, at that point, the managers of the “taggable” system (i.e. flickr, in this case) would probably have more of an argument to remove those tags due to the malicious nature of that practice.

@ben – Agree. Tagging can be highly effective and also counter-effective. It suits some topics and projects and not others. It can also sit along side traditional categorisation or hierarchy methods as an alternative navigation system. It doesn’t need to be one or the other.

@pat – The beauty of tagging is that anomalies don’t break the system because the system is looser and more flexible, it can cope with things that do not fit into neat boxes or may be placed in the wrong box (although it could still be argued that a totally incorrect tag may have meaning to the tagger).

I have actually conducted experiments on this. Take this flickr photo for example where we tried to add masses of tags to test numbers and relevance of tagging.

PS. thanks for the tabindex – I had missed that little sucker!

p.s.: your tab order is screwy…the submit button doesn’t come after this textarea, as would be expected

I think the key difference between tagging and traditional meta-data is that tags are “human”; and they form a concensus and not a catalogue. It’s democratic, which means it’s arbitrary and unfair! But it’s surprisingly effective for certain topics.

It does bomb when you get naming clashes though – eg. trying to look up posts tagged “opera” (or “thong” ) will return wildly unrelated posts. That’s not a fault of the system though, that’s just life. Humans don’t live by a controlled vocab.

And yeah, any librarian worth their salt really should at least give it a liten too…..

http://www.longnow.org/projects/seminars/

(search for Shirky on that page)