Supporting Research

• OrganizationalParadigms
  • 2 and 1/2 classification systems
  • The nature of Hierarchies
  • The nature of Faceted systems
  • Faking Facets in an Hierarchy
  • Changing data and the element of Time
  • How the cookie crumbles: Hierarchies in a state of nature
  • How the cookie crumbles: Facets in a state of nature
  • The nature of Tags

• Chandler_IA.ppt:
• Chandler_IA.pdf:

Hierachies, Faceted classification and Tagsonomies. Chandler's Hybrid Approach to Organizing Personal Information.

Takeaway: Don't throw out the baby with the bath water: A moderated, hybrid approach to information organization that plays to the strengths of traditional hierarchical systems as well as incorporating new innovations in tagging and faceted systems.

A. Given the nature of personal information in the age of personal computing:

1. Personal information (ie. email) is now incredibly easy to create, therefore, there is more of it than we know what to do with
2. The content and subject matter of personal information spans a broad spectrum from tasteless joke forwards to company financial updates.
3. The content and subject matter of personal information change in unpredictable. New projects to take on, new people to interact with, new problems to solve.
4. Information flows in large quantities and at high speeds.
5. The meaning and significance of information is not always clear and changes over time.

B. Given the way in which people deal with their personal information in the age of personal computing:

1. People organize information as they go,
2. Do not want to spend a lot of time reorganizing it and
3. Have trouble keeping up with it.

The combination of A and B leaves most people stranded between a data set that grows rapidly and unpredictably and no good way to iteratively organize and reorganize information on an ongoing basis.

This coupled with the increasing importance of personal information in the workplace today:

1. For knowledge workers, the output of their work is information: decisions, proposals, priorities and research. As a result,
2. People have a real need to mine their information for knowledge.

The wholesale transference of traditional organizational structures, namely hierarchical folders to the PIM domain has had limited success. While intuitive and generic, hierarchies quickly fail to scale to accommodate both the sheer quantity of personal information we deal with and it's amorphous, changing nature.

Hierarchies are simply too rigid to deal with PIM data precisely because they are too hard to organize and reorganize on an ad-hoc basis.

Furthermore, the adherence to physical, location-based metaphors in computer-based organizational systems (where a single item can only live in a single location) places an unnecessary hindrance to better, more facile organization.

However we believe there are still cognitive benefits to hierarchical systems, which partially explains why some people stubbornly cling to their email folder hierarchies, no matter how ornery they get and no matter how much time they consume in maintenance.

This presentation evaluates the three systems in terms of the following criterion:

1. Ability to provide a high-level narrative of the data set
2. Ability to provide a guided navigation experience
3. Flexibility
4. Facility at storing semantics
5. Ease of use, both with respect to human-computer interaction and cognitive overhead.

It then proceeds to explain Chandler's hybrid approach to personal information management and organization.

Conclusions

It is important to provide users with distinct workflows for processing, managing and organizing information and selective use of Tagging versus Facets versus Hierarchies is a good way to do that.

Hierarchies should be shallow, preferably no more than two levels and stay within the realm of conceptual elements unlikely to change over time. Things like: Kinds of items, Types of Attributes and Types of Collections.

Solution

Chandler's shallow, conceptual hierarchy provides a stable, but loose, conceptual framework and high-level narrative for organizing information, while leaving space for flexibility and user-defined growth over time.

Once we get into the realm of user content, however, more flexible systems like Faceted Classification and Tagging are required to build a cognitive and workflow ramp between the initial stages of Processing and Managing information and the more mature phases of Organizing information into a structure.

As a result, Chandler�s �hybrid� approach to organization is not simply a hodge-podge offering of Tags, Facets and Hierarchy as competing, alternate organizational affordances. Instead, the three emerge as phases in the life cycle of a unified notion of �collections of items�. This in turn sets up the framework for an iterative workflow structure that attempts to adapt to the way people want to work and interact with information.

Why do we care about Information Architecture for PIMs?

Chandler is the Open Source Applications Foundation's flagship "Personal Information Manager (PIM) Application and Platform" project. This presentation attempts to explain some of the research we've done in the realm of end-user Information Architecture and how it has been applied in the design.

Why should we care about Information Architecture in the PIM domain?

The explosion in the quantity (if not quality) of personal information we produce, receive and interact with on a daily basis has made library scientists out of us all, whether we want to be or not. Armed with seemingly bottomless folder hierarchies, Categories expressed in every color of the rainbow and Tags as far as the eye can see, how is the modern information worker faring under the burden of their Incredible-Ever-Expanding-Inbox? Not very well. http://www.nytimes.com/2005/11/10/fashion/thursdaystyles/10EMAIL.html?8hpib

Our new-found "wealth" in information compels us to revisit age-old library science problems with an eye towards more flexible, personalized approaches to managing large data sets.

Some questions we've asked ourselves include:

A. How does our relationship to information change when:

1. The granularity of information goes from something on the scale of a book, a journal article or a letter to something that's a three-sentence email or a two-word task?
2. As a direct result of 1., the sheer quantity of information we interact with has skyrocketed to a few hundred discrete items of information per day, all related to each other in myriad ways.
3. When information is no longer "just reference material" that "helps us" do our work but becomes the substance of our work instead. (What is an information worker anyway?)

B. How do our conceptions of "what needs to be done with information" change when the significance and meaning of our information changes all the time?

C. How do "Information Management" and "Information Processing" differ from "Organization of Information" and what are their respective roles in this new information universe?

In Chandler, we have done a use-case based "end-state" analysis of the scope and substance of information in the PIM domain in an attempt to re-taxonomize personal information into a loose, human-centered, conceptual framework. However, we are also re-examining PIMs from a "changing-state" perspective, looking at how personal information and personal organizations of information evolve over time.

In particular:

1. How does the meaning and significance of personal information change for people over time?
2. How do people extract and generate more information from the information they already have?

More generally:

1. What is the life cycle of information as understood through the limitless ways human users interact with their personal information?
2. What are people trying to accomplish with their information at various stages of the information life cycle?

In sum, how can we come to terms with the overwhelming gnarly-ness of personal information through a linear, workflow-based view of the problem. Can a more adaptive organizational paradigm that changes with information as it passes through the stages of its lifecycle smooth out the flow of information through our PIMs?

And finally, with respect to specific organizational affordances:

1. In the context of end-user goals, workflows and information life cycles what are Tags, Faceted Systems and Hierarchies good for respectively? What are they bad for respectively?
2. Should Tags, Facets and Hierarchy be offered to users as independent organizational affordances? OR Are they really phases in the lifecyle a single affordance?