WebM Powered Semantic Video Demo

From the WebM project blog[1], a link to a JavaScript and WebM powered demo on what is possible with open video when it also carries rich metadata. It is actually just a taste of a larger project, Web Made Movies[2], looking to really push what is possible with open video and the latest generation of web technologies. The JavaScript library, Popcorn.js[3], that made it possible is downloadable though I cannot find an explicit license to figure out what obligations, if any, you incur by forking (as the github page openly invites everyone to do) and making changes.

I had to reload the demo once or twice to get the features to work, once they did, wow. It is a little overwhelming but an effective demo. I used the latest beta of Firefox 4.

The video in question is well worth watching in its own right. It and Web Made Videos a project made under the auspices of Mozilla’s Drumbeat initiative. Drumbeat specifically aims to pull in all kinds of people, not just techies, to build on and highlight what is possible with the open standards of the web. The film maker responsible for the demo page is Brett Gaylor, one of my favorites for his “RIP: A Remix Manifesto” documentary.


1. WebM Semantic Video Demo, WebM project blog
2. Web Made Movies
3. Mozilla’s Popcorn.js, Github

Wikipedia Experimenting with the Semantic Web

As Technology Review explains, some demonstrations at the 2010 Semantic Technology conference show directions the collaborative encyclopedia may pursue with future development. The idea behind the semantic web is that links and document structure aren’t just mechanical, tying topics together and allowing for attractive and readable presentation. In addition, meaning would be encoded into links, why two pieces of information are related and how. Adding that same layer to a document’s structure would clarify standalone content, what parts are the table of contents, preface matter, and so on.

The first targets for Möller and Parscal are the “infoboxes” that appear as summaries on many Wikipedia pages, and the tables in entries, such as this one showing the gross national product of all the countries in the world.

That’s Erik Möller, a deputy director at the WikiMedia foundation, and colleague Trevor Parscal, a user-experience developer working for the foundation. This low hanging fruit makes sense, it already has a certain organic structure that meshes well with making it more accessible to programs looking to parse, transform and otherwise make novel use of Wikipedia’s content.

On the whole. bringing semantically meaningful structure into such a dense information source makes more sense to me than just about any application of so-called Web 3.0 principles I’ve seen so far. Wikipedia also has the kind of human processing power that would be required to drive forward truly useful markup of their content, beyond the easy to markup infoboxes and tables. I would even suggest that adding a semantic layer could attract, or re-attract, a whole crop of contributors, enriching the more steady state work that has been going on, by all accounts, more recently.

Firefox Personas May Add Dynamic Update Capability

I’ll admit that I’ve never really seen a point in Mozilla’s sort of themes-light feature for Firefox, Personas. Sarah Perez at ReadWriteWeb captures exactly why I am not alone in feeling this way but goes on to suggest that may soon change.

Several ideas are under consideration for future development that are as equally meh-worthy as the idea of Personas itself. One stands out, adding the ability to incorporate ambient information into a Persona. What that would mean exactly is still somewhat vague. In other applications, ambient information is typically data that is aggregated and displayed in a highly visual, usually very simplistic way. Think about a widget that changed color from cooler shades to warmer as a stock you are tracking swings upwards in price. That would be a pretty good example.

I like this idea of using subtle aspects of Firefox’s appearance to enrich the information being viewed. I think it would do best for omnipresent information, like perhaps the response rate on your social message stream, or specific to what your are viewing. Imagine a Persona that could alter its appearance to cue you in to the ratio of ads to actual content in a page or the number of inbound links to help you develop a sense of how authority ebbs and flows over time.

In that vein, I see the ambient info suggestion for Personas as part of a very interesting cluster of features that have been explored by the lizard wranglers. It is very consistent with discussions around the semantic web, extracting, sensing and digesting information beyond the mere textual content that makes up even the most dynamic sites we are used to viewing today.