Off Course-On Target

The Future of Metadata

My recent presentation at the 2006 International Conference on Digital Archive Technologies (ICDAT 2006) in Taiwan, dealt with the connection between content, competencies, and context, and how these domains are based upon a foundation of supporting metadata.  Below is a SlideShare version of the slides I used and if you would like to reuse or repurpose any of the slides you can download ICDAT2006.ppt

Some of my key points for the future of metadata included:

• Automated metadata generation (AMG) for producing the vast majority of metadata
• Augmenting AMG with the manual metadata from both mass contribution (consumers) and experts such as librarians, indexers and the like
• Making metadata much more transparent to the end users and remember that metadata is a means to the end of finding (rather than searching!)the "right stuff"
• Metadata "beyond text" to include the vast majority of content which is in the form of video, audio, photos, etc.
• Getting rid of the limiting default definition for metadata as "data about data" and extending it to include such things as metadata on people (e.g. competencies, skills, knowledge), places (e.g. GIS, location based), and equipment
• The power of inferred and implicit metadata, that which can be inferred from existing metadata by their relationships, such as if we know that daughters = female then all instances of daughter are also instances of female
• Capturing and automating "Attention metadata" (patterns of preferences, practices)
• The critical role and need for capturing and using Contextual metadata

After covering some of the history and current myths and misconceptions, I addressed the topic of the future of Learning Objects and covered some of the following points:

• The mass contribution phenomena of consumers providing massive quantities of metadata. Examples include all the metadata for music such as that on CDDB and the metadata being generated by the millions using Flickr and del.icio.us
• The mass customization and personalization of Learning Objects, such that we are able to have "just the right" content at the just the right time, on just the right device or medium, etc.
• Automating the assembly of content in general and Learning Objects in particular
• The inclusion of contextual metadata to enable content assemblies that are ideally matched to the unique situation of each individual

Given that I managed to do the whole presentation without a voice and yet got a rousing applause and extended session of questions afterwards, it appeared that they were able to hear me, and the comments resonated strongly with this audience.

Till next time,

w
a
yne
=====

Meta fun with Metadata in Taipei

Recently, I was privileged to be an invited speaker at the 2006 International Conference on 2006 International Conference on Digital Archive Technologies (ICAT 2006) in Taiwan, which was held on the beautiful campus of Academia Sinica, the highest and largest academic institution in Taiwan. I joined a very illustrious group of local and international speakers.

The focus of these digital archives groups is on documenting and preserving the valuable assets of libraries, museums, and national archives. Metadata is their key focus. Taiwan is a world leader in metadata and this leadership was well illustrated by both the presentations and the demonstrations at this event.

Over the past seven or eight years, I've had the pleasure of working with many local and national project groups in Taiwan who have been working on metadata and Learning Objects at a very high level, so I know of their work firsthand. Their holistic approach and national scale is truly amazing. Part of my message to them was to become aware of the leadership they are unknowingly providing through example and to encourage them to extend their outreach so that more people can learn and benefit from their experiences.

Of particular note at this event was the degree to which there are some truly significant advancements both from Taiwan and from some EU projects in "getting beyond text" with metadata. There were some excellent examples of extracting and automating the metadata generation of metadata from video and audio. This included the growing efficiency and effectiveness of speech to text conversation to create fully searchable text transcripts of all audio (much like I do with my podcasts here on Off Course - On Target).

Even more impressive were the examples of being able to do not only automated face recognition of people within video, but also the recognition of "objects" such as cars and buildings. With the recent explosion of video on the Web, just start to imagine how much the value goes up when there is the accompanying metadata for all this video that enables you to find just what you need.

The conference organizers asked me to talk about the future of both metadata and Learning Objects, something that I had great fun doing, and I received a very enthusiastic response from the audience. I framed my presentation within my overarching theme of the convergence of content, competencies and context, and how each of these domains and their combinations are based upon a foundation of supporting metadata.

See the other postings to see or download the slides and learn more of my Taipei adventure.

Till next time,

w
a
yne
=====

C-ing the Future @ eLearning 2006 - Part 2

The convergence of Content, Competencies and Context, a topic that I presented at the AAEC eLearn 2006 event, enables us to take advantage of some of the real substance lurking behind what is often too much hype surrounding things such as Web 2.0. As Tim Berners-Lee noted about the web, most of what is talked about as Web 2.0 is really just getting to what he had in mind all along. His thoughts parallel my own on Learning Objects. We are only now starting to see the realization of the vision that I first had back in 1990 for truly personalized content that led me to create the original Learning Object model.

For example, more people are starting to recognize that “Reusable Learning Objects” are NOT what they want, because there is an inverse relationship between the amount of context and the amount of reusability. As we get better at creating highly-personalized learning objects or content that is “just right” for one person at any one time in one context, then its reuseability gets closer to zero. If the learning object or content is just right for only you and only for today, then this becomes the only time and situation that this exact assembly of content or Learning Object is of value. 

However, we also have to have the maximum amount of reusability if there is any hope that this can scale up to provide the massive amounts of content that we need to meet the needs of every person (all 6.4 billion of us and counting), every day. And we CAN have this, but NOT at the assembly level, but rather at the opposite end of the spectrum—the very small components or “raw assets” of content.

The often-used comparison to Lego blocks works very well here; it is the individual blocks that have the greatest reusable value, which makes Lego so enduring and enables the creation of any assembly imaginable from these individual blocks. The assemblies can be unique and “just right” yet you can disassemble them back to individual blocks which can in turn be used to create any number of other unique assemblies.

Check out the graphics in the presentations bleow or download my presentation slides (PPT format, 6.9 MB) for a clearer explanation of all this and some working examples.

Well, that's going to have to do me for today. Unfortunately the quake seems to have shaken my vocal cords rather badly. I just made it through the presentation today, but there’s not much left of my voice now so I’ll sign off and try to get it back in shape for the next stop in Taipei on Wednesday.

w
a
yne
=====

C-ing the Future @ eLearning 2006 - Part 1

C-ing the Future @ eLearning 2006 - Part 1

I was honored to be an invited speaker at this year's eLearn 2006  event, sponsored by the Association for the Advancement in Education Computing (AACE). I had a great time presenting my thoughts on "C-ing the Future: the convergence of Content, Competencies and Context" to a very receptive audience.

Of course, after the earthquake yesterday, the organizers had quite an interesting day trying to hold sessions in a conference center with no power! However, everyone seemed to take it in stride and the organizers are to be complimented for doing such a great job of minimizing the disruption. In fact several people suggested that there  should be a "no power day" at every event since it produced some great benefits, such as more genuine conversation and fewer pre-planned presentations. This is worth noting for all events, since it is all too easy for any of us to fall into the rut of the familiar and worry too much about staying “on course” and getting to far “off target” in the process.

Based on audience reactions and feedback this topic of "C-ing the Future" is proving to have lots of utility and value. The three large domains of Content, Competencies and Context pretty much cover the landscape and provide a simple framework that enables us to focus on each of these three "C’s" individually. Yet it also serves to remind us that learning is both a very human and very holistic experience. It needs to be worked on with this singularity of purpose, which requires all three to come together to create true learning and performance.

There's lots more to say about this model of 3 "C's", so I'll talk more about it next time.

w
a
yne
=====