Thanks.

I think there are two different things in action here–research and discovery–and two different audiences (though with some overlap, for sure.

Researchers are looking for very specific things and yes, they need vetted data, curatorial input, and good solid cataloging. They’re the ones our database serves and everything we can do to improve it is worthwhile.

People who are in discovery mode cast a wide net, using tags and keywords, and don’t necessarily expect that everything they find is relevant. What they want is the ability to browse, rather than drilling down through data to the one perfect answer. Serendipity is a big part of what they’re doing (and having fun doing it.

Just to let you know where I’m coming from, since I haven’t posted on this before — though I have definitely given it some thought. I manage the Digital Lab at the Museum (and used to be the Museum Archivist). We’re the ones who create the images you see on the website, and we run the Museum’s image database.

When I first started working on the data template for that database I was convinced that we needed to include subject and authority terms developed by professionals. But cataloging at that level is time-consuming and expensive and Collection on the Web (and Flickr) went live without the extra data.

I’m a tagging convert. I’ve been watching the process now for several months and am more and more convinced that the tagging that people are adding to our website and our Flickr images is way, way, way more valuable than I ever thought it would be and in a lot of different ways. We’re getting a lot of good information added to the records, people are looking for and finding images, and they’re interacting with us about their questions and the information they have to add.

In other words, it’s not all about finding a picture, it’s about interaction, or community, if you like.

As to giving an institutional “stamp of approval” to tags as you caution against: I’m a firm believer that 1) there’s no such thing as a good tag or a bad tag and that 2) people know that they are not vetted. They’re all just tags. It’s an informal, additive process–the value is in the aggregate, not in the “correctness” of each individual tag. When people see something wrong or conflicting, they let us know (and believe me, they do.

I have read the report from LOC P&P, and it does not answer the questions posed above.

Until there are objective assessments of the quality of crowdsourced data, I strongly advise caution in their use by museums and libraries. You risk putting your institutional stamp of approval on erroneous and persistent public information.

Outreach projects using Web 2.0 are powerful and effective; they do not need to be justified by positing a curatorial value which as yet has no demonstrated basis in fact.

I’m going to refer you to the Library of Congress on this – the LOC has done excellent long format report on the benefits and findings regarding the issues you are talking about here. Read it here. I believe the Powerhouse is working on a long format report as well that will be released soon, so keep your eyes out for that. It’s a lot easier to report these kinds of things when you have some room – not easy to do in either a blog post or a comment like this one.

A few more questions:regarding “crowdsourced” data on images in your collection:

— Do you verify user-supplied tags and comments?
— How many crowdsourced leads were staff unable to verify?
– How many items were tagged with conflicting if not mutually-exclusive terms?
– What is the “arrearage” or backlog of crowdsourced tags, comments, and notes to be verified?
– What kind of disclaimer do you provide patrons about the unverified crowsourced data that appears on the Web with your institution’s imprimatur?
– When will you do a cost-benefit analysis of crowdsourcing versus professional curatorial research and cataloging?

Web 2.0 is great for outreach, and is certainly of use in widening curatorial networks of knowledgeable collectors, scholars, and amateur enthusiasts, but I am concerned that there has been no impartial independent review of the quality and value of “crowdsourced” data in the museum and library environment.

These comments prompted me to go back to the copyright page on our website. I don’t actually think there’s a conflict here. What we say is:

Non-commercial use of text and images in which Brooklyn Museum holds the copyright is permitted, with attribution, under the terms and conditions of a Creative Commons License.

Emphasis on “in which Brooklyn Museum holds the copyright.”

The images we put on Flickr Commons are all either in the public domain or old enough that they probably would be, if we knew the date and creator, thus the designation “no known copyright restrictions.” This trumps the Creative Commons license noted on our website, so there isn’t any conflict between the Museum website and what we say on the Commons that I see.

Deb Wythe