What are the issues delivering taxonomy support in SharePoint 2007 and SharePoint 2010?

Thursday, 15 July 2010 11:26

At a basic level, out-of -the-box SharePoint 2007 and SharePoint 2010 do not even provide a tree control that can show the parent-child structure which elevates a taxonomy from a flat look-up list.    Standardizing taxonomy-driven metadata across site collections, automatically adding taxonomy-metadata to documents are "must have" requirements for anyone serious about information management and that's where Semaphore for SharePoint comes in.

Key Issues

Standardization of metadata across site collections

SharePoint 2007 and SharePoint 2010, design requires that site columns (the typical container for taxonomy terms) are duplicated in every site collection that needs them.  This adds a huge administrative overhead to typical implementations.

Semaphore provide Central Admin menus to distribute the site column definition across all the site collections on a farm.  The definition references the taxonomy which is available external to SharePoint as a Web Service.

Over-loading users

Often introducing SharePoint introduces a whole new way of working to your users.  One of the main changes is the "edit properties" dialog that appears every time a user saves or uploads a document.   This is something they never had to worry about when saving to the shared network drive, but now they are being asked to fill in a load of properties (which the organization recognizes are critical for information management and governance).

A taxonomy often drives one of these [subject] properties.   Getting a user to populate this well can be difficult.   If the taxonomy is small they may pick some terms, but then across the site each term has thousands of documents, so findability is not greatly improved.   If the taxonomy is large the user has to work to understand where to find the relevant terms.  The "cognitive load" is high, often too high for all but dedicated experts, indexers or information architects.

Semaphore solves this issue - Auto Classification suggests terms to the user in the background or as part of the edit properties process.  The user can search for terms, using synonyms and signpost terms as well as the actual taxonomy labels.   The user can browse a Semaphore tree-control and select the terms.

Common metadata over any list type

SharePoint content comes in many list types.  A custom list of items, a publishing site with page definition lists, a document library, a blog site, etc.   For information management these are all pieces of content that should have the common metadata applied to them.  From an integration perspective, each list type can behave in a different manner.

Semaphore can classify all the information in SharePoint, not just that in a Document Library.  The integration will, for example allow you to choose the relevant text columns from a publishing site or wiki entry to be sent for classification (title, byline, body).