Using Bibio to Annotate Information about Web Site and Web Pages

[Pete DeVries]

I marked up some test files using the bibio ontology, these tie web
pages to taxon concepts.

You can see them at:

http://rdf.geospecies.org/sites/wbru_site.rdf

http://rdf.geospecies.org/sites/mosquito_catalog.rdf

http://rdf.geospecies.org/sites/iowa_mosquitoes.rdf

I was wondering what people thought about this use of the bibio
ontology and would appreciate feedback about whether I did
these correctly.

Thanks!

- Pete

[Rob Styles]

Hey Pete,

This is exactly the kind of thing I had in mind when I proposed the Web Site and Web Page classes. :-)

There are some minor details aside from the bibo stuff. In the first two, you state the dc:creator as ?.html. This is problematic as a html document can't be the creator of a web site. You may want to introduce a foaf:organisation to solve that.

Secondly, you could put all the taxa relationships inside the bibo:Website element, the seperate rdf:Description isn't necessary. Tough it may help you seperate your own thoughts, of course. The meaning is the same either way.

I wonder if the Website URI should refer to http://wrbu.si.edu rather than http://wrbu.si.edu/ in order that the site and its homepage can be distinguished? IIRC in the w3c specs the trailing slash is resource part for the root. That then raises the question of whether http://eco.bcb.iastate.edu/mosquito/ can be a website or if it is a webpage? I would have classified it a webpage within http://eco.bcb.iastate.edu but that may not reflect your, or the author's, understanding or intent.

Nice to see it being tried out.

rob

[kidehen]

Pete,

Very nice!

Could you please add a triple to the RDF information resource
(container of the RDF triples) that states something like this:

<http://rdf.geospecies.org/sites/wbru_site.rdf> a bibo:Document;
foaf:topic <for-each-about-uri-in-the-rdf-doc> .

Once done, I can then explore the data starting at the description of
the container/doc (information resource) itself.

Links:

1. http://linkeddata.uriburner.com/about/html/http://rdf.geospecies.org/sites/wbru_site.rdf
-- shows nothing meaningful until tweak above is made
2. http://linkeddata.uriburner.com/ode/?uri=http%3A%2F%2Frdf.geospecies.org%2Fsites%2Fwbru_site.rdf%23Dataset
-- see "Navigator Tab" which shows the graph hosted by information
resource.


Kingsley

[Pete DeVries]

Hi, I fixed the organization issue by creating a file for
organizations and adding those to the files.

I added Kingsley's changes to this file

http://rdf.geospecies.org/sites/wbru_site.rdf

but not the rest, I wanted to make sure I did what he wanted me to do.

Once know it is ok, I can do the other two, and more later...

Thanks!

- Pete

On Feb 11, 8:23 am, kidehen <kide...@openlinksw.com> wrote:
> Pete,
>
> Very nice!
>
> Could you please add a triple to the RDF information resource
> (container of the RDF triples) that states something like this:
>
> <http://rdf.geospecies.org/sites/wbru_site.rdf> a bibo:Document;
> foaf:topic <for-each-about-uri-in-the-rdf-doc> .
>
> Once done, I can then explore the data starting at the description of
> the container/doc (information resource) itself.
>
> Links:
>

> 1.http://linkeddata.uriburner.com/about/html/http://rdf.geospecies.org/...

> -- shows nothing meaningful until tweak above is made

> 2.  http://linkeddata.uriburner.com/ode/?uri=http%3A%2F%2Frdf.geospecies....

[Rob Styles]

That's great Peter, I still think it would be worth removing the trailing slash of the uris for the websites, so that the uri with the trailing slash can refer to the root page.

Does anyone else have thoughts on that?

rob

[Pete DeVries]

Thanks,

I think I got it all fixed. Also I added the bibo:Document blocks to
all of them.

At this point I want to keep the taxa info as a separate block. These
will get pretty long.

Also I don't have all the mosquito pages on these sites marked up, I
wanted to get the format
worked out so I can automate it.

Thanks for all the good suggestions!

- Pete

On Feb 11, 2:31 pm, Rob Styles <mmmmm...@gmail.com> wrote:
> That's great Peter, I still think it would be worth removing the trailing
> slash of the uris for the websites, so that the uri with the trailing slash
> can refer to the root page.
>
> Does anyone else have thoughts on that?
>
> rob
>

[kidehen]

Pete,

Awesome!

See: http://linkeddata.uriburner.com/about/html/http://rdf.geospecies.org/sites/wbru_site.rdf

Failing to describe the data container is an all too common oversight
on the part of RDF data publishers :-)

Kingsley

[kidehen]

Pete,

On additional request, could up please add a triple that provides a
literal value to a property that is a sub-property of rdfs:label,
should you not want to use rdfs:label itself. Once this is done, the
"About" label in <http://linkeddata.uriburner.com/about/html/http://
rdf.geospecies.org/sites/wbru_site.rdf> will contain a more human
friendly value.

Kingsley

On Feb 11, 12:33 pm, Pete DeVries <pete.devr...@gmail.com> wrote:

[Bruce D'Arcus]

On Wed, Feb 11, 2009 at 7:58 PM, kidehen <kid...@openlinksw.com> wrote:

> On additional request, could up please add a triple that provides a
> literal value to a property that is a sub-property of rdfs:label,
> should you not want to use rdfs:label itself.

Does dcterms:title qualify? If not, is it worth suggesting this to the DCMI?

Bruce

[Peter DeVries]

Done,

Seems to work :-)

- Pete

--
---------------------------------------------------------------
Pete DeVries
Department of Entomology
University of Wisconsin - Madison
445 Russell Laboratories
1630 Linden Drive
Madison, WI 53706
------------------------------------------------------------

[Frederick Giasson]

Hi Pete!

Thanks for using BIBO for this kind of work :0

However, here are a couple of issues I noted:

1- The namespace you are currently using for bibo is not good; it should
be "http://purl.org/ontology/bibo/" instead of
"http://purl.org/ontology/biblio/"
2- In this document
"http://rdf.geospecies.org/sites/iowa_mosquitoes.rdf" you wrote:

==========
<bibo:Website rdf:about="http://eco.bcb.iastate.edu/mosquito">
<bibo:uri>http://eco.bcb.iastate.edu/mosquito/</bibo:uri>
<dcterms:title>Iowa-Mosquito.net</dcterms:title>
<dcterms:creator
rdf:resource="http://people.geospecies.org/people/Lyric_C_Bartholomay/"/>

<doap:Project
rdf:resource="http://people.geospecies.org/projects/Iowa_Mosquito.net/"/>
<bibo:AcademicArticle
rdf:resource="http://rdf.geospecies.org/refs/sucaet2008wbr/" />
<foaf:Organization
rdf:resource="http://rdf.geospecies.org/org/index#iastate_ent" />
</bibo:Website>
==========

The problem here is with:

==========
<doap:Project
rdf:resource="http://people.geospecies.org/projects/Iowa_Mosquito.net/"/>
<bibo:AcademicArticle
rdf:resource="http://rdf.geospecies.org/refs/sucaet2008wbr/" />
<foaf:Organization
rdf:resource="http://rdf.geospecies.org/org/index#iastate_ent" />
==========

You are using classes where you should use properties.

You should think about something like:

==========

==========
<bibo:Website rdf:about="http://eco.bcb.iastate.edu/mosquito">
<bibo:uri>http://eco.bcb.iastate.edu/mosquito/</bibo:uri>
<dcterms:title>Iowa-Mosquito.net</dcterms:title>
<dcterms:creator
rdf:resource="http://people.geospecies.org/people/Lyric_C_Bartholomay/"/>

<dcterms:relation
rdf:resource="http://people.geospecies.org/projects/Iowa_Mosquito.net/"/>
<dcterms:relation
rdf:resource="http://rdf.geospecies.org/refs/sucaet2008wbr/" />
<dcterms:relation
rdf:resource="http://rdf.geospecies.org/org/index#iastate_ent" />

</bibo:Website>

<doap:Project
rdf:about="http://people.geospecies.org/projects/Iowa_Mosquito.net/">
.....
</doap:Project>

<bibo:AcademicArticle
rdf:about="http://rdf.geospecies.org/refs/sucaet2008wbr/" />
....
</bibo:AcademicArticle>

<foaf:Organization
rdf:about="http://rdf.geospecies.org/org/index#iastate_ent" />
......
</foaf:Organization>

==========

You could create subProperties of dcterms:relation to describe the kind
of relation that holds between this website document and the other
resources (2 articles and a project).

Tell me what you think.

Thanks!

Take care,

Fred

[kidehen]

Pete,

See:
http://linkeddata.uriburner.com/about/html/http://species.geospecies.org/family_concept_uuid/1e0e9bfe-f1ee-4b14-9511-cb896e8ebf97/

In seeAlso you point to: <http://dbpedia.org/page/Mosquito> instead of
<http://dbpedia.org/resource/Mosquito> .
I would even go one further and use an "owl:sameAs" link instead :-)

Kingsley

On Feb 11, 2:16 am, Pete DeVries <pete.devr...@gmail.com> wrote:

[kidehen]

Pete,

Also look at the content encoding for some of the LSID URN style
links. We de-ref URNs also, courtesy of wire protocol proxies in our
RDFizer middleware (aka. Sponger).

Examples with issues:

1. http://linkeddata.uriburner.com/about/html/http://lsid.tdwg.org/urn:lsid:ubio.org:namebank:380629
2. http://linkeddata.uriburner.com/about/html/urn:lsid:ubio.org:namebank:380629

Example that works:

1. http://linkeddata.uriburner.com/about/html/urn:lsid:ubio.org:namebank:423551

Kingsley

Kingsley
On Feb 11, 2:16 am, Pete DeVries <pete.devr...@gmail.com> wrote:

[Peter DeVries]

Thanks everyone, Also found there are other problems with my ontology. :-(

Fixed the dbpedia/pages/ /resource issue. Fortunately it seems to be just one record

with that error. 

Still wrong if you go to today's live uri, but fixed in the data and the rdf dump.

I added the following to my ontology. Maybe these would be worth adding to the bibo ontology?

    <!-- http://rdf.geospecies.org/ont/gsontology.owl#hasProject -->

    <owl:ObjectProperty rdf:about="#hasProject">

        <rdfs:label>hasProject</rdfs:label>

        <rdfs:subPropertyOf rdf:resource="&terms;relation"/>

    </owl:ObjectProperty>

    <!-- http://rdf.geospecies.org/ont/gsontology.owl#hasArticle -->

    <owl:ObjectProperty rdf:about="#hasArticle">

        <rdfs:label>hasArticle</rdfs:label>

        <rdfs:subPropertyOf rdf:resource="&terms;relation"/>

    </owl:ObjectProperty>

 

Then modified the rdf file with this replacement

      <geospecies:hasProject rdf:resource="http://people.geospecies.org/projects/Iowa_Mosquito.net/"/>

      <geospecies:hasArticle rdf:resource="http://rdf.geospecies.org/refs/sucaet2008wbr/" />

Will look into the LSID resolving issue later today.

Thanks Again!

- Pete

[Pete DeVries]

Hi Kingsley,

The LSID appears to be correct. It matches what is at uBio.

http://names.ubio.org/browser/details.php?names=on&authors=on&sci=on&vern=on&namebankID=380629

Also this resolver works for that LSID

http://lsid.tdwg.org/urn:lsid:ubio.org:namebank:380629

They may have some sort of rate limiting system. The Scorpion LSID,
the example that works, was just added a few days ago, while the
Mosquito LSID has been
in my database for a long time.

Also, I was not sure about using SameAs for levels above species.
These different groups like dbpedia etc. have different sets of sub
and superclasses and
I wonder if it is a good idea for me to SameAs them. It is probably
not a good idea to do it at the species level either for my work, but
I am still experimenting.

Also, what is the best way to build ontologies using the bibo
ontology? Currently I have problems if I import the bibo ontology into
protege 4.0 v108, with my
current ontology. If I try to start a new one with imports of bibo and
geonames, I also get errors. I need the geonames in some form since
that it what I
use to determine if a species is in a particular state or county.

I also fixed some errors in the rdf files, I was being a doap with
doap:Project :-)

Thanks!

- Pete

On Feb 12, 9:48 am, kidehen <kide...@openlinksw.com> wrote:
> Pete,
>
> Also look at the content encoding for some of the LSID URN style
> links. We de-ref URNs also, courtesy of wire protocol proxies in our
> RDFizer middleware (aka. Sponger).
>
> Examples with issues:
>

> 1.http://linkeddata.uriburner.com/about/html/http://lsid.tdwg.org/urn:l...
> 2.http://linkeddata.uriburner.com/about/html/urn:lsid:ubio.org:namebank...
>
> Example that works:
>
> 1.http://linkeddata.uriburner.com/about/html/urn:lsid:ubio.org:namebank...

[kidehen]

Pete,

Issue showed up via:

curl http://lsid.tdwg.org/urn:lsid:ubio.org:namebank:380629 | xmllint
-

document states utf-8 , but in fact it is western charset.

I'll try again if this is fixed.

Kingsey

On Feb 13, 2:37 am, Pete DeVries <pete.devr...@gmail.com> wrote:
> Hi Kingsley,
>
> The LSID appears to be correct. It matches what is at uBio.
>

> http://names.ubio.org/browser/details.php?names=on&authors=on&sci=on&...