Tuesday, March 01, 2005
Semantic Blogging
Crayzer, Steve. "Semantic blogging and decentralized knowledge management." Communications of the ACM, 47 (12), December 2004, 47-52.
Steve Crayzer discusses the needs of the HP Laboratories' Semantic Web research group (Bristol, UK) to set up "a system capable of aggregating, annotating, indexing, and searching a community's snippets." Snippets being small pieces of information that might be relevant for retrieval later.
Crayzer points out that blogs alone do not meet the group's requirements. While blogs allow easy capture of information, decentralized data-gathering, and easy methods to add to the collected information (through comments and annotations), Crayzer wants his system to have a flexible data model, be extensible, and make it possible to infer new metadata from existing metadata.
Perhaps not surprisingly, Crayzer uses Semantic Web framework tools to add the necessary functionality to the weblog framework. Specifically, RSS1.0 and RDF.
What I find most interesting about this article is Crayzer addressing the other models for ontology creation, albeit briefly and as a contrast to the Semantic Web model. Decentralized ontologies, such as those created through Flickr, del.icio.us, and Topic Exchange enable a user community to set up its own ontology by creating a list of tags or bookmarks for content. "However, in all of these systems the ontology lacks semantics and are both centralized and universal. The Semantic Web may be better served by precise, local, domain-specific vocabularies that are loosely coupled, rather than by a one-size-fits-all central ontology, no matter how collaborative." (p 52)
This argument parallels the debate between controlled vocabulary searching and full-text searching, and human-created indexing versus automated indexing (relating specifically to exhaustivity). The Semantic Web approach allows for individual disciplines and groups to take more control over access to their information. While this may enable more precise search and retrieval for knowledgeable experts, the ability of John Q. Public (or the sales guy in the field) to obtain rapid access to the same information may be limited by a lack of familiarity with the thesaurus.
I think within an individual business enterprise, the Semantic Web approach recommended in this article could be the most effective approach. However, if you conceive that the information may be used by a wider audience (up to and including the average consumer or general public), being able to search full-text (or through a publicly available and generalized tag set) will better promote access.
Both approaches have merit. As I've said before, if you only learn one tool, and it's a hammer, every problem looks like a nail. The more tools and approaches you are willing to examine, the more creative you can be in proposing solutions.
There are several links at the end of this article I will be looking into, including Planet RDF for communal blogs, Platypus for wikis, the semblog platform for aggregators, and Compendium as an authoring tool utilizing Semantic Web blogging ideas.
This is a good article, and thought provoking! I hope you have a chance to read it.
|
Steve Crayzer discusses the needs of the HP Laboratories' Semantic Web research group (Bristol, UK) to set up "a system capable of aggregating, annotating, indexing, and searching a community's snippets." Snippets being small pieces of information that might be relevant for retrieval later.
Crayzer points out that blogs alone do not meet the group's requirements. While blogs allow easy capture of information, decentralized data-gathering, and easy methods to add to the collected information (through comments and annotations), Crayzer wants his system to have a flexible data model, be extensible, and make it possible to infer new metadata from existing metadata.
Perhaps not surprisingly, Crayzer uses Semantic Web framework tools to add the necessary functionality to the weblog framework. Specifically, RSS1.0 and RDF.
What I find most interesting about this article is Crayzer addressing the other models for ontology creation, albeit briefly and as a contrast to the Semantic Web model. Decentralized ontologies, such as those created through Flickr, del.icio.us, and Topic Exchange enable a user community to set up its own ontology by creating a list of tags or bookmarks for content. "However, in all of these systems the ontology lacks semantics and are both centralized and universal. The Semantic Web may be better served by precise, local, domain-specific vocabularies that are loosely coupled, rather than by a one-size-fits-all central ontology, no matter how collaborative." (p 52)
This argument parallels the debate between controlled vocabulary searching and full-text searching, and human-created indexing versus automated indexing (relating specifically to exhaustivity). The Semantic Web approach allows for individual disciplines and groups to take more control over access to their information. While this may enable more precise search and retrieval for knowledgeable experts, the ability of John Q. Public (or the sales guy in the field) to obtain rapid access to the same information may be limited by a lack of familiarity with the thesaurus.
I think within an individual business enterprise, the Semantic Web approach recommended in this article could be the most effective approach. However, if you conceive that the information may be used by a wider audience (up to and including the average consumer or general public), being able to search full-text (or through a publicly available and generalized tag set) will better promote access.
Both approaches have merit. As I've said before, if you only learn one tool, and it's a hammer, every problem looks like a nail. The more tools and approaches you are willing to examine, the more creative you can be in proposing solutions.
There are several links at the end of this article I will be looking into, including Planet RDF for communal blogs, Platypus for wikis, the semblog platform for aggregators, and Compendium as an authoring tool utilizing Semantic Web blogging ideas.
This is a good article, and thought provoking! I hope you have a chance to read it.
