I finally got around to reading the preliminary SharePoint 2010 taxonomy and metadata info on MSDN. As most of you know, MOSS 2007 had some serious challenges in this regard (e.g., no hierarchical metadata, no sharing controlled vocabularies across site collections, etc.) -- and they caused information architects like me much consternation.

It would appear that SharePoint 2010 includes many new features under the umbrella of Enterprise Metadata Management, and they seem to have addressed some key issues. Here are the highlights.

1. Terms & Keywords

SharePoint 2010 has a few types of vocabularies with different levels of control:

Terms: Basic construct -- a word or phrase that can be associated with content. A term can become a managed term or a managed keyword.

Managed Terms: A controlled term that can only be created by those with appropriate permissions. Term sets (think of them as taxonomy facets) are collections of related terms that can be hierarchically structured.

Managed Keywords: User-generated keywords (aka tags) kept in a non-hierarchical list called the keyword set.

What's nice is that you can easily turn a managed keyword into a managed term, which essentially sets up SharePoint 2010 as a decent platform to blend taxonomy and folksonomy approaches (more on this later).

2. Managed metadata & the Term Store

Once you have managed terms in place, you can create a new column type called "managed metadata". This column type is like the 2007 lookup, but points to the managed term sets instead of lists.

One of the biggest limitations of MOSS 2007 was the site collection boundary: lists were specific to a site collection, so in order to share metadata and taxonomy across multiple collections, you had to essentially reproduce the same lists in each collection.

Managed term sets (and managed keywords) now live in what's called the Term Store Management Tool -- a centralized database that allows term sets to be shared across a farm.

SharePoint 2010 -- Centralized Metadata Repository. Image courtesy of EndUserSharePoint.com

3. Taxonomy & Metadata Management

Within the Term Store, there are some metadata and taxonomy management functions enabled, though the documentation is not clear on how many of them are achieved.

Feature Comment
Create or delete term sets Term sets can also be thought of as facets
Add, modify or delete terms It is unclear however how modifications and deletions are propagated throughout content that is already tagged - this is a big problem in 2007, as changes are not reflected in tagged content
Arrange terms in hierarchies It would appear that this uses a drag & drop mechanism
Define synonyms Unclear as to how this works and whether it is used in the search thesaurus
Import terms Via a .CSV file
Promote managed keywords into managed terms Great news for those of us who want to use tags as a source of candidate terms for the taxonomy
Create multi-lingual     taxonomies Unclear if this is at the term level or you have to create language-specific versions of each term set... likely the latter

My biggest concern is the whole question of integrity and change propagation, but overall this is a huge leap forward from 2007 where there was no taxonomy management to speak of. This is actually more functionality than I've seen in many very expensive CM suites, so right now on the surface, I'm rather pleased with this set of features.

4. Tagging

I haven't seen a screenshot of the interface for tagging from a hierarchical term set, so I'm still a bit in the dark on whether it's a tree view or a panel pick list. But, Microsoft does point out that there is also type-ahead functionality for managed metadata fields, meaning that you can start typing and it will present suggestions from the managed term set, as well as show you the term's position in the hierarchy (contextual placement). This is a massive improvement over the previous tagging, which was limited to flat pick lists - cascading lists were even a customization.

Type-ahead is also available for managed keywords, which is excellent in that it will suggest previously entered keywords as suggestions - encouraging cross-user consistency in tagging is always a boon. You can ignore the suggestions however and add a new keyword at any time.

SharePoint 2010 Tagging Interface
SharePoint 2010 -- Tagging Interface for Managed Keywords. Image courtesy of EndUserSharePoint.com.


Still a lot of details to discover about how a lot of these metadata and taxonomy functions behave IRL, but for now I'm secure in my assessment that Microsoft has taken a sizable leap forward in fixing some of the ills of MOSS 2007 giving us taxonomy and metadata nerds many of the tools we need to create effective content structures in SharePoint 2010. Looking forward to seeing more as the beta program kicks off. If any of you have direct experience and more information to share, please do so in the comments!

I'll write separately about my take on these functions and how they encourage hybrid approaches to taxonomy and folksonomy.