- From: Martin Hepp \(DERI\) <martin.hepp@deri.org>
- Date: Wed, 14 Dec 2005 15:40:05 +0100
- To: "'Shashi \(MIT\)'" <shashi.mit@gmail.com>, "'Asankhaya Sharma'" <asankhaya@yahoo.com>
- Cc: <semantic-web@w3.org>
Dear all: I agree with Shashi and Sandro but would like to add a comment. First of all, I think there are two types of content that can be annotated: a) Content from communities in which there is a proper incentive to annotate your data correctly in the sense that the authors want to contribute to not only a high recall BUT ALSO A HIGH PRECISION. I think that e.g. academic research is such a field - you do not want your papers to appear on top of Google for any query. "Annotation Spamming" does not make too much sense in such areas. There is significant overlap with what Shashi calls "the most information hungry of people", but the explanation is different. b) Content for which there is some incentive for incorrect annotation (e.g. including keywords that are far-fetched or even obviously incorrect). This is an area in which without additional mechanisms, "Annotation Spamming" will likely make annotations without additional mechanisms useless, since malicious content providers will with intention add incorrect annotations to their content. Folksonomies and lightweight approaches for type a) content can basically lift off immediately - we only need search engines that make use of this, which will appear as soon as there is enough such content to become famous or rich with such an engine (I believe in Adam Smith that the improvement of one's own fate is an inevitable incentive for the provision of goods or services ;-)). But as soon as we go to business scenarios or other type b) content, we need ranking algorithms or other type of functionality that helps us filter out the noise and bias created by economic incentives for providing false or far-fetched annotation; and, to my knowledge, such is not yet available in a mature form. Just my two cents :-) Martin --------------------------- martin.hepp@deri.org, phone: +43 512 507 6465 http://www.heppnetz.de / http://www.deri.org -----Original Message----- From: semantic-web-request@w3.org [mailto:semantic-web-request@w3.org] On Behalf Of Shashi (MIT) Sent: Dienstag, 13. Dezember 2005 18:42 To: Asankhaya Sharma Cc: semantic-web@w3.org Subject: Re: Re: How will the semantic web emerge I am with Sandro on this one. The Semantic Web is likely to be orthogonal to the current www - in the sense that it will complement the human web. My guess is HTML and current web-base scripting languages ( a al Qwerty keyboards) are so well established that getting them to change to include Ontological data seems unlikely. Ontologies and RDF etc will run in the background as 'enhancements' by people exporting their databases, possibly via semantic webservices and what have you. So mixing in with HTML seems less and less a possibility I have also one more thought on this, that the initial wave of the Semantic Web will be used by the most information hungry of people e.g. researchers, scientists, business analysts etc - they will be the "lead users". Your average Mom and pop store Website may little use for Semantics for a good time to come.
Received on Wednesday, 14 December 2005 14:41:09 UTC