Striking achievements from the National Library

I was so immensely impressed by the National Library of Australia's new Australian Newspapers BETA search service that I sent a message out about it in the irregular newsletter I send to staff. Someone found a relative in a newspaper overnight (they wouldn't do that kind of searching at work) and reported in. And since then I have come across several more happy users.

The service - see http://ndpbeta.nla.gov.au/ - is available and working now. It is freely available to the public and currently (several weeks ago) contains 73,000 out of copyright newspaper pages (approx 730,000 individual articles) from 1803 onwards. Another 20,000 digitised newspaper pages will be added each week to the service. The goal is to provide the full text of a daily newspaper for each state and territory up to about 1955, in both image and machine-readable form. You can also go and look at the Australian Newspapers Digitisation Program (ANDP) website for more information – at http://www.nla.gov.au/ndp/project_details/ The site has a huge amount of information about the program.

Most of you, I assume, will just want to go to the site and try it out. One of the great strengths of the ANDP is that it is national; if your subject of interest is not nailed down in a specific state, the default search is national, across multiple newspaper titles. The text of the newspapers appears in all cases both as an image, and as scanned text using OCR. And because of the vagaries of OCR, if you spot errors (and there are plenty) and itch to correct them, you can. You can register as a user, and this enables you to correct the text - more crowdsourcing. The service even has a list of Top Text-Correctors.

You can add tags, ask questions, and use facets to refine your search, which is relevance ranked. In other words, the tools for accessing information through the Australian Newspapers service are definitely superior to the average library catalogue.

Word of the day

Today's word is crowdsourcing, and thank you to Rebecca for pointing it out, and to the JISC posting which we both saw. I was sure that I had used it before in this blog, but I haven't. I have just remembered where I did use the word, which was in the text of this library's IT Strategic Framework, developed early in 2007. So the word has been around for a while.

According to the Wikipedia, it means "the act of taking a task traditionally performed by an employee or contractor, and outsourcing it to an undefined, generally large group of people, in the form of an open call." Crowdsourcing has been coined along the lines of outsourcing, with which it is compared and contrasted. For us, crowdsourcing is most often applied to user-contributed keywords, or tags. There is an interesting article in Boingboing about the Library of Congress "using Flickr to crowdsource tagging and organizing its photo archive." And Flickr is the great example of user-generated subject descriptors - billions of them.

There are discussions amongst librarians about whether user-contributed tags are "better" than structured subject headings. Here is the definitive answer to that question: yes, and no. They do different things, both useful.