Wikipedia-based online resource declaration and retrieval

[Yao Ziyuan]

Could develop a program or browser extension that helps put such a
"resource declaration" on a web page, blog post, etc. to declare a
resource to facilitate its retrieval by people interested in the same
topic and resource type.

//////////////////////////////////////////////////////////////////////////////////////////////
// Topic-ID: http://en.wikipedia.org/wiki/Mobile_phone
// Resource-Type: product
// Title: Galaxy Nexus
// Description: Touchscreen slate Android smartphone by Samsung and
Google
//////////////////////////////////////////////////////////////////////////////////////////////

This will be "the Semantic Web done right", I think.

For one thing, Wikipedia has the largest, most comprehensive ontology
on the Web (even larger than academia-produced WordNet and Cyc). For
another thing, people really like Wikipedia and use it every day.

[Yao Ziyuan]

Post by Yao Ziyuan
Could develop a program or browser extension that helps put such a
"resource declaration" on a web page, blog post, etc. to declare a
resource to facilitate its retrieval by people interested in the same
topic and resource type.
/////////////////////////////////////////////////////////////////////////// ///////////////////
// Topic-ID:http://en.wikipedia.org/wiki/Mobile_phone
// Resource-Type: product
// Title: Galaxy Nexus
// Description: Touchscreen slate Android smartphone by Samsung and
Google
/////////////////////////////////////////////////////////////////////////// ///////////////////
This will be "the Semantic Web done right", I think.
For one thing, Wikipedia has the largest, most comprehensive ontology
on the Web (even larger than academia-produced WordNet and Cyc). For
another thing, people really like Wikipedia and use it every day.

Let me make a more complex example. Suppose I'm a professor interested
in hiring a PhD student to research "father-bother merger". I could
post such a resource manifest on my blog:

//////////////////////////////////////////////////////////////////////////////////////////////
// Topic-ID: http://en.wikipedia.org/wiki/Phonological_history_of_English_low_back_vowels#Father.E2.80.93bother_merger
// Resource-Type: job
// Title: PhD Student
// Description: Looking for a PhD student to research the phonological
phenomenon of "father-bother merger".
//////////////////////////////////////////////////////////////////////////////////////////////

Perspective PhD students interested in this specific topic ("father-
bother merger") will use a program to automatically search the web or
the blogsphere for any new resource manifests whose Topic-ID is
"http://en.wikipedia.org/wiki/
Phonological_history_of_English_low_back_vowels#Father.E2.80.93bother_merger"
and whose Resource-Type is "job". Of course they can further filter
manifests whose Title contains "PhD".

[Yao Ziyuan]

Post by Yao Ziyuan

Post by Yao Ziyuan
Could develop a program or browser extension that helps put such a
"resource declaration" on a web page, blog post, etc. to declare a
resource to facilitate its retrieval by people interested in the same
topic and resource type.
/////////////////////////////////////////////////////////////////////////// ///////////////////
// Topic-ID:http://en.wikipedia.org/wiki/Mobile_phone
// Resource-Type: product
// Title: Galaxy Nexus
// Description: Touchscreen slate Android smartphone by Samsung and
Google
/////////////////////////////////////////////////////////////////////////// ///////////////////
This will be "the Semantic Web done right", I think.
For one thing, Wikipedia has the largest, most comprehensive ontology
on the Web (even larger than academia-produced WordNet and Cyc). For
another thing, people really like Wikipedia and use it every day.

Let me make a more complex example. Suppose I'm a professor interested
in hiring a PhD student to research "father-bother merger". I could
/////////////////////////////////////////////////////////////////////////// ///////////////////
// Topic-ID:http://en.wikipedia.org/wiki/Phonological_history_of_English_low_back...
// Resource-Type: job
// Title: PhD Student
// Description: Looking for a PhD student to research the phonological
phenomenon of "father-bother merger".
/////////////////////////////////////////////////////////////////////////// ///////////////////
Perspective PhD students interested in this specific topic ("father-
bother merger") will use a program to automatically search the web or
the blogsphere for any new resource manifests whose Topic-ID is
"http://en.wikipedia.org/wiki/
Phonological_history_of_English_low_back_vowels#Father.E2.80.93bother_merge r"
and whose Resource-Type is "job". Of course they can further filter
manifests whose Title contains "PhD".

Note that a manifest can have more than one Topic-ID. For example, in
the above example, we can add two Topic-IDs:

//////////////////////////////////////////////////////////////////////////////////////////////
// Topic-ID: http://en.wikipedia.org/wiki/Phonological_history_of_English_low_back_vowels#Father.E2.80.93bother_merger
// Topic-ID: http://en.wikipedia.org/wiki/English_phonology
// Topic-ID: http://en.wikipedia.org/wiki/Phonology
// Resource-Type: job
// Title: PhD Student
// Description: Looking for a PhD student to research the phonological
phenomenon of "father-bother merger".
//////////////////////////////////////////////////////////////////////////////////////////////

[Ivan Shmakov]

[…]

Post by Yao Ziyuan
Perspective PhD students interested in this specific topic ("father-
bother merger") will use a program to automatically search the web or
the blogsphere for any new resource manifests whose Topic-ID is
"http://en.wikipedia.org/wiki/
Phonological_history_of_English_low_back_vowels#Father.E2.80.93bother_merger"
and whose Resource-Type is "job". Of course they can further filter
manifests whose Title contains "PhD".

Well, sort of, but the question is: will there be a powerful
query language for this kind of metadata, or will it be some
kind of a simplistic keyword search engine instead?

FWIW, one can use SPARQL against structured data extracted from
Wikipedia, like, right now. Consider checking the DBpedia
project at http://dbpedia.org/.

E. g., one can search for all the fantasy genre fiction,
authored by persons born in the XIX century with a SPARQL query
like:

SELECT DISTINCT ?author ?born ?work
WHERE {
?author
a <http://dbpedia.org/ontology/Person> ;
<http://dbpedia.org/ontology/birthDate> ?born .
FILTER (?born < "1901-01-01"^^<http://www.w3.org/2001/XMLSchema#date>)
?work
<http://dbpedia.org/property/author> ?author ;
<http://dbpedia.org/ontology/genre> ?genre .
OPTIONAL {
?genre
a <http://dbpedia.org/class/yago/FantasyGenres> .
}
OPTIONAL {
?genre
<http://dbpedia.org/property/redirect>
[ a <http://dbpedia.org/class/yago/FantasyGenres> ] .
}
FILTER (bound (?genre)) .
}

Thanks to RDFa, such descriptions could readily be added to,
e. g., XHTML-based Web pages, and then processed by a variety of
tools, such as those associated with the Redland RDF library.