Want to try for yourself a new information retrieval method that is powerful yet very simple to understand and implement? One that will affect everything from enterprise search to your own desktop search? Follow these simple steps . . .
- Make up a random string of letters and numbers.
- Paste it into your favorite search engine. You should get no results.
- Now paste the same string into one of your serving Web pages. You might ask colleagues to paste it into some of theirs.
- Wait a few days for the search engine robots to crawl your pages and add your latest content to their full-text search databases. Sigh. It's not instant gratification, but worth the wait.
- Search again. Presto! Results now include all the pages you have tagged with your own unique string.
What are we to make of this? What shall we do with this extraordinary ability to retrieve tagged pages from anywhere on the Web?
First let's note that huge numbers of people are adding metadata to their photos at Flickr and their bookmarks at del.icio.us. They are tagging with ordinary words, which are ambiguous. If people use the same word to tag very different material, it's not easy to retrieve exactly what you want. For years it has been said that content contributors would not add metadata to content, that special metadata editors would be needed. Not so. Nowadays "folksonomies"—built by do-it-yourself "free taggers"—show us that motivated people can and will add metadata to content if it helps to make that content findable again.
Yet metatdata must be usable to be useful. Flickr and del.icio.us have proprietary servers that use their metadata to find your content. Web page metadata lives "outside the docs." It does not show up in standard search engines. Enter the meme, memography, and embedded aboutness.A meme is an idea, "the smallest unit of cultural transmission," according to Richard Dawkins, who coined the term meme to capture the evolutionary nature of ideas, comparing them to biological genes. Think of all the content on the Web as divisible into these ideas, or memes. Memography is a method of mapping and linking all these memes. It starts with creating meme IDs, globally unique strings that identify a single meme. Something like the DNA codons (nucleotide sequences) that map to a single gene. The great part is you can create your own meme IDs following steps one through three above. The Memography Wiki (www.memography.org) suggests how to design your meme IDs so they don't clash with others.
Next you need a page somewhere that describes what your meme is about, so you and others know when to embed it in a Web page. The Memography Wiki is modeled after Wikipedia and let's you create a public "aboutness" page there. When you embed the meme ID in a document, you can wrap it in a hyperlink to the meme aboutness page. This is called a memelink. Users can click on it to understand what is being tagged.
Next you need a page somewhere that describes what your meme is about, so you and others know when to embed it in a Web page. The Memography Wiki is modeled after Wikipedia and let's you create a public "aboutness" page there. When you embed the meme ID in a document, you can wrap it in a hyperlink to the meme aboutness page. This is called a memelink. Users can click on it to understand what is being tagged.
Tim Berners-Lee's hyperlink made the Web possible. Larry Page and Sergey Brin collected inbound links to a Web page in a database to make Google PageRank possible. The memelink will extend search engines like Google to make ideas more findable. It's the next big thing in enterprise search.The Semantic Web uses XML, RDF, RDF schemas, ontologies, and still-to-come AI inference engines to discover machine-readable meaning in Web pages (placed there by humans, of course). The new "Memetic Web" lets you add your own machine-readable meaning to a page, with a link to the meme aboutness page, so inference engines could also discover something about your meaning.
Of course you don't need to publicize what you're doing. Corporate intranet applications will want to keep their memes private, along with their specific meme IDs (to prevent others from spamming the meme IDs) and their aboutness pages. As Lou Rosenfeld aptly blogged about the Memetic Web, "I do think there's promise in tagging a document with metadata that's both unique and meaningful, especially in closed sites like intranets, where abuses are likely to be minimized."
Memography can even add value to desktop search. With a search tool like Apple Sherlock or Google Desktop, you can embed meme IDs in your files, then retrieve them later when you have forgotten which folders you put them in. With surgical precision, a memetic search makes your work findable, on your hard drives or anywhere on the new Memetic Web. And findability is what search is all about.