Saturday, June 16, 2007
Amazing!
Had to blog this.. it speaks for itself. Just watch, needs no introduction: http://www.ted.com/index.php/talks/view/id/129
Saturday, June 09, 2007
SWG - Semantic Web Gateway Project
Today a group of us (open source developers in sheffield) got together to discuss some components they all have, some problems that seem to exist with open data, and what might be done about it. Specifically, we were discussing issues with extracting the planning alerts data from Sheffield council web site in order to provide a feed into the planning alerts database (http://www.planningalerts.com/). This led onto wider ranging discussions about what other data could be scraped from council sites, about the management and preservation of that data, and about best mechanisms for sharing and alerting based on that data.
From there things got quite abstract, as we identified other domains and projects that could benefit from a drop in component or service that bridged the gap between existing (Lets say Web 1.0) data silos and newer semantic web applications.
The SWG or Semantic Web Gateway Project is an attempt to abstract out the functions of a wide range of screen scrapers and other data extraction tools out there and bring them all together in a common management environment. Having collected them together in one place, have those agents will spit out RDF into a jena repository fronted by OAI and SRW services.
SVN repository and other stuff to follow soon.
From there things got quite abstract, as we identified other domains and projects that could benefit from a drop in component or service that bridged the gap between existing (Lets say Web 1.0) data silos and newer semantic web applications.
The SWG or Semantic Web Gateway Project is an attempt to abstract out the functions of a wide range of screen scrapers and other data extraction tools out there and bring them all together in a common management environment. Having collected them together in one place, have those agents will spit out RDF into a jena repository fronted by OAI and SRW services.
SVN repository and other stuff to follow soon.
Monday, June 04, 2007
Upcoming changes to the LOM Repository / Online Learning Portal Project
The learning resource portal is being revamped! There's lots of meat and potato changes needed to support the reporting of eLearning credits and the authoring of learning object metadata and loads of other cool stuff (Including an ajax based alternative to the fat swing based tagging tool, and a whole new revision of the tagging tool itself). However..... we've also got our eye in on a load of not so mainline developments to try and re-engage the community around the sharing of learning resources generally.... Here's some of the things on the list for the upcoming iterations...
* the portal itself will become a branded application of a generic learning object repository system. This will allow the core system for sharing learning resources to be branded in different ways for different communities. The current portal will continue to support it's target audience of teachers and learners, but because the underlying learning object metadata is generally applicable it will be possible to upload things like sound recordings of lecture notes, e-reserve materials, lecture notes, useful images and animations, lesson plans.. basically anything that a particular learning community wants to share. We can arrange repositories vertically, for example by domain/subject/keystage or horizontally, for example by media/medium or other useful aggregations.
* User tagging: One of the most exciting aspects of the new generic Repository - end user tagging and similarity detection. If you tag a resource, we'll be able to detect similarities with other profiles using the SVD algorithms we've been playing with, sorta like last.fm for learning resources. This kinda goes hand in hand with the user reviews that are coming as a part of the next release. Learners, Teachers and pretty much anyone who wants to can tag and share info in this way. We may toy with automatically exporting resources based on national curriculum specifiers (More generically controlled vocabs) to delicious tags using those tags, to try and capture more users into the portal itself. One of the best things about this is that we hope to be able to take trivial tagging, ID3 for example, and compliance test it to be able to create minimal lom documents. This means you could take a whole bunch of ID3 tagged mp3 lecture recordings, and expose them as, for example, JISC IE Conformant learning objects (More on content packages later).
* Simple tagging interface.. Yeah lom is cool, but not everyone wants to create huge lom documents, so the new online tagging interfaces will mean that people can catalog and upload learning objects without needing a doctorate in metadata and semantic ontologies. For now we've left out auto extraction of id3 tags from mp3's but it's on the cards in the near future.
* More web services - The new col portal already exposes all it's search functions as SRW/SRU based web services, we'll be exposing the tagging and other metadata functions for people to intergate their own applications with... we'll be swallowing our own medicine, with version four of the stand-alone tagging tool being able to consume XML LOM objects downloaded from portal, and then re-upload them in a round trip.
* Cross searching... well.. actually the app already does this, but we'll be creating a federated portal that can cross aggregate specific learning object Repositories, so we have our schools repository, but it would be nice to cross search some FE and HE repos too. We've also had interest from some projects creating video footage who might be interested in creating LOM records for learning video resources they have created and uploaded to utube. Cool :) So there will be some content in the FE portal at least.
* Object storage.. maybe in for the release, ability to upload resources, docs, video, sound recordings, etc as well as metadata.
Well.. just a taster for whats to come over the next 6 months of learning portal redevelopment. Hopefully iteration 0 will kick off around the 18th of june with iterations following evey three weeks.
* the portal itself will become a branded application of a generic learning object repository system. This will allow the core system for sharing learning resources to be branded in different ways for different communities. The current portal will continue to support it's target audience of teachers and learners, but because the underlying learning object metadata is generally applicable it will be possible to upload things like sound recordings of lecture notes, e-reserve materials, lecture notes, useful images and animations, lesson plans.. basically anything that a particular learning community wants to share. We can arrange repositories vertically, for example by domain/subject/keystage or horizontally, for example by media/medium or other useful aggregations.
* User tagging: One of the most exciting aspects of the new generic Repository - end user tagging and similarity detection. If you tag a resource, we'll be able to detect similarities with other profiles using the SVD algorithms we've been playing with, sorta like last.fm for learning resources. This kinda goes hand in hand with the user reviews that are coming as a part of the next release. Learners, Teachers and pretty much anyone who wants to can tag and share info in this way. We may toy with automatically exporting resources based on national curriculum specifiers (More generically controlled vocabs) to delicious tags using those tags, to try and capture more users into the portal itself. One of the best things about this is that we hope to be able to take trivial tagging, ID3 for example, and compliance test it to be able to create minimal lom documents. This means you could take a whole bunch of ID3 tagged mp3 lecture recordings, and expose them as, for example, JISC IE Conformant learning objects (More on content packages later).
* Simple tagging interface.. Yeah lom is cool, but not everyone wants to create huge lom documents, so the new online tagging interfaces will mean that people can catalog and upload learning objects without needing a doctorate in metadata and semantic ontologies. For now we've left out auto extraction of id3 tags from mp3's but it's on the cards in the near future.
* More web services - The new col portal already exposes all it's search functions as SRW/SRU based web services, we'll be exposing the tagging and other metadata functions for people to intergate their own applications with... we'll be swallowing our own medicine, with version four of the stand-alone tagging tool being able to consume XML LOM objects downloaded from portal, and then re-upload them in a round trip.
* Cross searching... well.. actually the app already does this, but we'll be creating a federated portal that can cross aggregate specific learning object Repositories, so we have our schools repository, but it would be nice to cross search some FE and HE repos too. We've also had interest from some projects creating video footage who might be interested in creating LOM records for learning video resources they have created and uploaded to utube. Cool :) So there will be some content in the FE portal at least.
* Object storage.. maybe in for the release, ability to upload resources, docs, video, sound recordings, etc as well as metadata.
Well.. just a taster for whats to come over the next 6 months of learning portal redevelopment. Hopefully iteration 0 will kick off around the 18th of june with iterations following evey three weeks.
Subscribe to:
Posts (Atom)
