Amazing plenoptic camera

Plenoptic camera imageHands up who’s ever heard of a plenoptic camera? No, me neither. A group at Stanford University have created a plenoptic camera that takes a single still image but uses software to allow the focus to be set after the picture is taken. Yes you heard right, the focus can be set and indeed changed after the picture is taken. According to the Stanford team this wizardry is achieved by “inserting a microlens array between the sensor and main lens, creating a plenoptic camera. Each microlens measures not just the total amount of light deposited at that location, but how much light arrives along each ray. By re-sorting the measured rays of light to where they would have terminated in slightly different, synthetic cameras, we can compute sharp photographs focused at different depths.” So that’s clear then.

Check out the gallery of images that show how pictures can be refocussed after the event. Amazing.

Technorati Tags: ,

There is no shelf

Clay Shirky posted an interesting piece a while back declaring “Ontology is Overrated”. I’ve only just seen it (thanks Ben) but it struck a number of chords. Clay suggests that many of the present classification systems such as the Library of Congress system were under some degree of physical constraints, for example issues of shelving reflecting books as physical objects. In the digital domain there is no shelf, no physical constraint, and indeed with hyperlinks there’s the possibility of expressing a rich level of interconnectedness, or intertwingularity that would not have been impossible with pre-digital semi-manual classification systems. Clay goes on to discuss how tagging hyperlinks creates a “market logic” dynamic classification system that frees the content creator attempting to grapple with descriptive metadata from having to decide which pre-defined category box his/her object fits in.

This richly dynamic informal classification based upon every individual’s unique perspective (and therefore context) is at the same time its greatest strength and its greatest weakness. The impermanence, fluidity and highly contextualized nature of folksonomies can create new links or routes of inquiry for the seeker of information yet it can also shield from view the exact object being searched for as a result of nobody using the same tags that you would have used because their context is different (though of course there’s always the brute force method of resource discovery by just searching largely unclassified data using Google). Where folksonomies gain however is in their scalability. Folksonomies benefit from a kind of wisdom of crowds effect when large numbers of individuals are tagging objects such as links.

What I’ve been increasingly thinking of attempting to implement in our learning content management system is a resource discovery mechanism that uses formal classification for each object as part of its metadata but also allowing learners to tag resources for their own purposes. It would be interesting to see how informal tagging complements formal classification and how learners will use both method to find resources. Could we expect the formal classification to define the nature of an object, by stating its intended use as declared by the teacher for example, but find that the tags describe how objects get used in practice?

Technorati Tags: , , , ,

Now using Ecto

Wagamama ramenI’m now using Ecto to post to this weblog. Since moving to WordPress I’ve found it much easier to manage my weblog even though I’ve been using the WP web forms to do so. After reading many recommendations I’ve decided to give Ecto a try as a desktop weblog editing tool. It seems to integrate quite well with my desktop media such as pictures and music stored in iPhoto and iTunes respectively. If this works you should see a picture from my iPhoto library (a recent trip to Wagamama) and the track I’m currently listening to (Good Fortune from the album ‘Stories From The City, Stories’ by PJ Harvey). Nice.

[Update] Hey cool, Ecto even has a built-in method of creating Technorati tags.

Technorati Tags: ,

Learning Resource Type vocabularies – a case for free-form tagging

The National Science Digital Library maintain a wiki on how they manage metadata. Here’s list of the types of learning resources that can be defined in IEEE LOM 5.2:Educational.LearningResourceType (and presumably if you’re using Dublin Core, DC.Type).

Now on one level this is a good thing. When creating a metadata record for a learning resource, knowing in advance how other people describe their learning resources helps you to decide how to describe yours. For example, if you upload a learning resource to a repository and declare that it’s of type ‘video’ by adding ‘video’ to the object’s 5.2:Educational.LearningResourceType metadata, then if everyone else uses ‘video’ rather than ‘movie’ to for all similar objects, pretty soon we’d get a lot of consistency and it’d be easy to search for all the videos in the repository. That’s the benefit of controlled vocabulary metadata, we select keywords from a predefined list. Selecting from a limited list of predefined types of objects works well when there are clearly distinct types of objects.

Where it gets tricky however is with resources whose type is not obvious. For example, from the NSDL’s own list, two adjacent terms, ‘project’ and ‘proposal’, two aspects of the same thing. A proposal defines a project, but is the proposal always distinct from the project itself? Plus more fundamentally if you disagree with definition of either of these types of object, which type, if any, should you use? It’s not easy to decide.

Another more fundamental problem with such a list of types of resources is that they often mix the form or format of a resource with its educational purpose. For example, ‘video’ tells me something about the data format, but ‘assessment’ tells me nothing about the data format but tells me a lot about what educational context it might be used for. Can an assessment contain a video for example? Probably. So there’s yet another problem. Aggregations of content can be very difficult to define when they contain multiple data formats. For example multimedia learning resources, should they be described in terms of their technical format, or their educational format, or both?

So, is there merit in creating a controlled vocabulary of types of learning resources or should the user be allowed to define their own type? Well, before you decide one way or another, consider the mine-field presented by the different subject domains. Any list that you can come up with someone will always come up with a type of resource that you hadn’t though of. So rather than attempt to define the empirical list of learning resource types, let the user enter whatever keywords best describe their resource. Like will attract like and soon enough people working in the same domain will be able to find resources of interest. Perhaps too much of what makes metadata difficult for many are the enforced choices a user has to make when describing their resource. Let the content creator decide based on their own professional judgement and maybe we’ll get more descriptive metadata being created voluntarily.

Technorati Tags: ,