Sunday, December 11, 2005

"Millions of Games"- example of structured tagging

Search Engine Watch,a web site dedicated to the search engine industry, has had very little good to say about folksonomy-type tagging. In folksonomy tagging users annotate pages, images and other web content with descriptive keywords without the aid of controlled vocabulary. For their view on this controversy see "Tagging Not Likely The Killer Solution For Search"

Recently, Search Engine Watch cited an example of a site called Millions of Games, a search tool for finding games on the Internet. Unlike other user-tagged sites, the developers of this site provide controlled vocabulary categories to assist users assign unambiguous category tags.

Search Engine Watch comments: "Rather than allowing users to tag anything with any words they fancy, [we think] adding structure, defining a mutually agreeable taxonomy and relying less on idiosyncratic language seems like a very promising approach. Millions of Games is an example of a good start to a more useful form of structured tagging. See "Where Tagging Works: Searching for a Good Game"

Value of the faceted taxonomy

Faceted Taxonomy - Creating a Better User Experience through Information Discovery - is taken up by Todd Peters in a paper that is part of KMWorld's Best Practices in Enterprise Knowledge Management (Nov/Dec 2005).

This is a compact article that lays the basis for understanding taxonomies as "a way of consistently organizing and classifying large amounts of data through a controlled vocabulary of terms". It describes the building-in of facets with Human Resources as an example, and the resulting navigational components. Author recommends this approach when "considering a site re-design or the implementation of a Web content management system".

Article is available after registration from http://www.kmworld.com/Categories/CategoryIndex.aspx?CategoryID=59

Tuesday, December 06, 2005

Tips about Taxonomies

I Column Like I CM: What’s so funny about Taxonomy CoPs, FAQs, and Tips? EContent (Dec 6, 2005) - Bob Doyle has created a web site named TaxoTips that he hopes will provide answers to questions people have about taxonomies and controlled vocabularies. He says that, "At TaxoTips, I'm going to tackle these and other more sophisticated options like multiple hierarchies, thesauri, and faceted classifications, as well as the ontologies, semantic networks, topic maps, and conceptual graphs that underpin the new thinking of the Semantic Web that aims to add machine-readable meaning to content."

Tuesday, November 29, 2005

CMS selection is a jungle of possibilities

EContentMag.com: Bob Doyle of EContent Magazine performed a study of the number of Content Management Systems (CMS) available for purchase. He came up with the astounding figure of 1,879 choices. CMS selection is a jungle, Doyle says, and you may well need a seasoned guide who will bring you back alive. See Bob's article as he provides links to directories of CMS products with lists of features.


Note to those shopping for a CMS: These CMS systems need to have some form of taxonomy and auto-categorization capability. Know what a taxonomy is and what kinds of features you require.

Autonomy Purchases Verity

AUTONOMY CORPORATION PLC ANNOUNCES AGREEMENT TO ACQUIRE VERITY, INC.: CAMBRIDGE, ENGLAND - 4 November 2005 - Autonomy Corporation plc, a global leader in infrastructure software, and Verity, Inc., a leading provider of business search and process management software, today announce that they have entered into a definitive agreement under which Autonomy will acquire Verity.


Autonomy and Verity are both major vendors of content management systems with auto-categorization and taxonomy functionality built into their product offerings.

Friday, November 18, 2005

Siderean extracts metadata

Siderean Announces New Solution for Metadata Generation Econtent Magazine (Nov 18, 2005)

"Siderean Software, provider of Seamark Navigator, an enterprise content navigation solution, has announced a product and services solution for extracting and aligning metadata from large amounts of unstructured and structured data based on the IBM Unstructured Information Management Architecture (UIMA)."

Views on Tagging

Tagging is very hot right now on the public Web. Del.icio.us and Flickr.com are most famous for it. People at Furl.net do it. Technorati has adopted it. And those using the new Yahoo MyWeb2.0 can do it. Tagging is the act of adding keyword terms for an article or bookmark, or, in the case of Flickr, a photo. Collectively these tags are called folksonomies.

There have been a few articles on this phenomenon (or craze depending on your point of view.)

Daniel Terdiman at CNET feels that 'Tagging' gives Web a human meaning (Nov 16, 2005). Cory Doctorow of the popular blog BoingBoing is quoted as saying, ""We've had this decades-long program of top-down metadata. People (were asked) to go out and become familiar with one ontology and to make sure data is categorized like this. But people are not very good at this ...". With tagging, people use words that mean something to them. It's not mentioned that one person's category for technical support may not the be same as another's.

These tags may be more useful to communities that share interests - perhaps smaller work groups. Brad Hill sees tagging as a "tool for collaborative social use". There is mention in this article that some corporations are adopting tagging internally.
""In a corporate environment, the interests are narrower than all the human interests on the Web and the vocabulary becomes narrower," said Dave Weinberger, a fellow at the Harvard Berkman Center."
There are other examples that show the potential in tagging - especially the one where people re-classified George Orwell's 1984.

Will these tags help in general web search? The editors at SearchEngineWatch are sceptical.

Danny Sulivan wrote that Tagging Not Likely the Killer Solution for Search (Mar 22 2005) - tags are wide open to abuse by spammers, and search engines have learned from past experience to ignore them. Sullivan, himself, doesn't find that tags (or categories in directories for that matter) much help in search. He prefers, as do so many, keyword search - the "warp drive to zip you to what you wanted". In Yahoo My Web Tagging and Why (So Far) It Sucks (June 30, 2005) , he describes the many weaknesses to tagging, among which is the sheer effort involved in browsing tags and trying to assess which would be most relevant. People have been disinclined to use directories even though directories contain hand-picked and reviewed sites. Why would they spend time with other people's tags? Automated classification is probably more effective as a search aid.

Where will it work? Maybe, just maybe, where people can choose terms from an established list. Chris Sherman says in Where Tagging Works: Searching for a Good Game (Nov2) that it can work if there is a controlled vocabulary in the background and points to the success of Millions of Games.
"The site uses controlled vocabulary (called "Gameology") to describe categories (arcade, shooter, puzzle, etc). Although you can also add your own free-form tags, these category tags are well known to most users, so there's little ambiguity about what the tags mean.""

Will people spend time tagging with or without a controlled vocabulary? For photos they will, since, as many have noted, that's the only way you can find them again. But will they for articles, documents, and other text items? Would people even tag their email? We'll see. This will play out on the Web first.

Sunday, November 06, 2005

Categorizing the Web

The general purpose Web search engine (Yahoo, Google and others) does not serve people needing specialized, industry-specific information well. This article mentions two companies that use categories to tailor the collection of information and display of results to particular business needs.

Web Search Gets Down to Business, by Cathleen Moore, Infoworld (Oct 31, 2005)

Convera's private label search engine, Excalibur, seems especially advanced with capability to do "deep web" indexing employing millions of topical categories and displaying results as "facets". Convera is targeting Excalibur to "vertical industries such as media, financial services, research, and legal".

The Excalibur web page states that "Convera has assembled a vast storehouse of commercially recognized taxonomies, plus privately compiled classification systems and other knowledge sources, to give Excalibur one of the largest and most detailed Web indexes on the planet. This Excalibur "intelligence index" links users to topical, targeted Web results in real time for a complete and comprehensive perspective."

Exalead, from France, also uses categorization to help searchers see how "provide an integrated search experience that serves up results from multiple sources in a single view".

Saturday, November 05, 2005

PaperThin: New Taxonomy Software

EContentMag.com: PaperThin, Inc., a Web publishing and content management software provider, has announced advanced content classification and discovery capabilities in CommonSpot Content Server version 4.6. The product's taxonomy module now provides taxonomy management for Web content administrators and contributors. New features include a taxonomy term editor and a taxonomy API, along with a facet-based navigation element.

Saturday, October 15, 2005

Inxight Smart Discovery

Announcements about new capabilities of Inxight Software's Smart Discovery have been appearing.

KMWorld noted in Text analysis takes off (Sep 28) that "
Inxight Software has introduced SmartDiscovery Analysis Server 4.2, which is said to offer faster, more flexible and accurate extraction, categorization and search indexing capabilities."

The entry at EContent (Sep 16) - Inxight Announces Availability of SmartDiscovery Analysis Server 4.2 - singled out the ThingFinder for entity extraction (people, places, companies etc) which also allows users to define their own entities using special syntax.

The Inxight web site has descriptions of company use of Smart Discovery under the Solutions tab.