BibTeX Mapping

5 views
Skip to first unread message

Pipian

unread,
Sep 25, 2007, 3:47:09 PM9/25/07
to bibliographic-ontolog...@googlegroups.com
I'm starting to map BibTeX to the ontology (the same mapping will in
theory be used to reverse from the ontology back to BibTeX) but there
are a few fields of BibTeX that I'm trying to figure out how they'll
map, namely:

SCHOOL (important for listing the school a thesis was written at)
HOWPUBLISHED (can't quite squeeze this into a dcterms:publisher)
NOTE (I think there's a mechanism for this, but I'm not sure)
KEY (May just shove this into a BibTeX extension vocab)
ADDRESS (particularly problematic for booklets, which lack a
publisher, Otherwise, I was thinking something along the lines of:

:whatever dcterms:publisher :publisher .
:publisher a foaf:Agent ;
foaf:isBasedNear [ a spacenamespace:Space ;
spacenamespace:name "BibTeX-Address" ] .

Perhaps with some sort of attempt at parsing out an address to get
more semantic information out (e.g. City, Province, country)

ORGANIZATION (I'm thinking bibo:organizer(?))
INSTITUTION (Used with tech reports. bibo:producer?)
CHAPTER (this is why we need bibo:chapter)
TYPE (used with tech report.)

I'm also not sure if I like how InBook/InCollection would have to be
mapped right now (to bibo:Article, when something like a bibo:Chapter
would be more appropriate)

Ian

P.S. Sorry for the disorganized nature of the content here. I'm
just writing out stuff as I generate it.

Bruce D'Arcus

unread,
Sep 25, 2007, 5:08:35 PM9/25/07
to bibliographic-ontolog...@googlegroups.com
I'm going to answer these, but I'm afraid you're not going to like what
I have to say (which is pretty vague). But perhaps we can push on
resolving these issues ...

Pipian wrote:

> I'm starting to map BibTeX to the ontology (the same mapping will in
> theory be used to reverse from the ontology back to BibTeX) but there
> are a few fields of BibTeX that I'm trying to figure out how they'll
> map, namely:
>
> SCHOOL (important for listing the school a thesis was written at)

We talked about this here before on the list. Honestly, I don't recall
what became of that. There are a lot of different ways to model things
like dissertations:

1) flat, a la BibTeX:

<http://ex.net/1> a bibo:Dissertation ;
bibo:school "Harvard" ;
bibo:degree "PhD" .

2) treat degree and institutions as resources:

<http://ex.net/1> a bibo:Dissertation ;
bibo:degree <http://ex.net/2> .

<http://ex.net/2> a edu:PhDDegree ;
bibo:awardingInstitution <http://ex.net/3> .

<http://ex.net/3> a foaf:Organization ;
foaf:name "Harvard" .

etc.

> HOWPUBLISHED (can't quite squeeze this into a dcterms:publisher)

No, we don't really have anything for that. Would be nice to see some
examples of actual data for this field.

> NOTE (I think there's a mechanism for this, but I'm not sure)

Another "thing we've discussed but not resolved."

> KEY (May just shove this into a BibTeX extension vocab)

Yes, I don't want this here, because it's really a user tag. It does not
really belong as a property of the resource per se.

> ADDRESS (particularly problematic for booklets, which lack a
> publisher, Otherwise, I was thinking something along the lines of:
>
> :whatever dcterms:publisher :publisher .
> :publisher a foaf:Agent ;
> foaf:isBasedNear [ a spacenamespace:Space ;
> spacenamespace:name "BibTeX-Address" ] .

We have the locality property for this. Another one of those details I'm
not entirely sure about/comfortable with.

> Perhaps with some sort of attempt at parsing out an address to get
> more semantic information out (e.g. City, Province, country)
>
> ORGANIZATION (I'm thinking bibo:organizer(?))

The organizer property is to link associated agents to an event. A
conference, for example, has an organizer.

<http://ex.net/1> a bibo:Paper ;
bibo:presentedAt <http://ex.net/2> .

<http://ex.net/2> a bibo:Conference ;
bibo:organizer <http://ex.net/3> .

<http://ex.net/3> a foaf:Organization ;
foaf:name "ABC Corp." .

> INSTITUTION (Used with tech reports. bibo:producer?)

Yeah, probably.

> CHAPTER (this is why we need bibo:chapter)

Huh? I don't think chapter in bibtex is used as Fred intended
bibo:chapter to be used. The former is the chapter *title*; isn't it?

The bibo:chapter property would be the chapter number.

> TYPE (used with tech report.)

Another TBD issue. We did talk about the need for a literal to describe
types (of letters, reports, etc.). But he hadn't resolved how. It seems
a little silly to call it "type."

> I'm also not sure if I like how InBook/InCollection would have to be
> mapped right now (to bibo:Article, when something like a bibo:Chapter
> would be more appropriate)

You know, I could have sworn we had a bibo:Chapter (or BookSection)
class! Don't know what happened to that. We need to fix that.

Bruce

Pipian

unread,
Sep 25, 2007, 7:28:59 PM9/25/07
to bibliographic-ontolog...@googlegroups.com

I think the latter is a bit more accurate to the concepts, so it
might be better to go at it that way (and in the process, remove the
individuals of the degrees).

>> HOWPUBLISHED (can't quite squeeze this into a dcterms:publisher)
>
> No, we don't really have anything for that. Would be nice to see some
> examples of actual data for this field.

I haven't seen any either. If I find some I'll give an example.

>> NOTE (I think there's a mechanism for this, but I'm not sure)
>
> Another "thing we've discussed but not resolved."

I figured as much. (rdfs:comment?)

>> KEY (May just shove this into a BibTeX extension vocab)
>
> Yes, I don't want this here, because it's really a user tag. It
> does not
> really belong as a property of the resource per se.

You're most likely correct, but on the other hand, having a sort-by
tag (since that's essentially what it is) might be beneficial.

>> ADDRESS (particularly problematic for booklets, which lack a
>> publisher, Otherwise, I was thinking something along the lines of:
>>
>> :whatever dcterms:publisher :publisher .
>> :publisher a foaf:Agent ;
>> foaf:isBasedNear [ a spacenamespace:Space ;
>> spacenamespace:name "BibTeX-Address" ] .
>
> We have the locality property for this. Another one of those
> details I'm
> not entirely sure about/comfortable with.

That could be a use of it. I wasn't sure about how that property was
being used so...

>> ORGANIZATION (I'm thinking bibo:organizer(?))
>
> The organizer property is to link associated agents to an event. A
> conference, for example, has an organizer.
>
> <http://ex.net/1> a bibo:Paper ;
> bibo:presentedAt <http://ex.net/2> .
>
> <http://ex.net/2> a bibo:Conference ;
> bibo:organizer <http://ex.net/3> .
>
> <http://ex.net/3> a foaf:Organization ;
> foaf:name "ABC Corp." .

That's tricky then, because PROCEEDINGS doesn't actually have a link
to the actual conference, so the conference itself would have to be
(virtually) a bnode. Doable, but not as helpful as it could be. :)

>> INSTITUTION (Used with tech reports. bibo:producer?)
>
> Yeah, probably.
>
>> CHAPTER (this is why we need bibo:chapter)
>
> Huh? I don't think chapter in bibtex is used as Fred intended
> bibo:chapter to be used. The former is the chapter *title*; isn't it?
>
> The bibo:chapter property would be the chapter number.

The BibTeX CHAPTER property is supposed to be the chapter number. If
the chapter is identified by a name, it should probably be
INCOLLECTION instead of INBOOK.

>> TYPE (used with tech report.)
>
> Another TBD issue. We did talk about the need for a literal to
> describe
> types (of letters, reports, etc.). But he hadn't resolved how. It
> seems
> a little silly to call it "type."

It certainly does. I figured this was pretty vague and hadn't been
specified though.

>> I'm also not sure if I like how InBook/InCollection would have to be
>> mapped right now (to bibo:Article, when something like a bibo:Chapter
>> would be more appropriate)
>
> You know, I could have sworn we had a bibo:Chapter (or BookSection)
> class! Don't know what happened to that. We need to fix that.

Sounds good to me.

Ian

Frederick Giasson

unread,
Sep 26, 2007, 8:33:40 AM9/26/07
to bibliographic-ontolog...@googlegroups.com
Hi all,

>> SCHOOL (important for listing the school a thesis was written at)
>>
>
> We talked about this here before on the list. Honestly, I don't recall
> what became of that. There are a lot of different ways to model things
> like dissertations:
>
> 1) flat, a la BibTeX:
>
> <http://ex.net/1> a bibo:Dissertation ;
> bibo:school "Harvard" ;
> bibo:degree "PhD" .
>
> 2) treat degree and institutions as resources:
>
> <http://ex.net/1> a bibo:Dissertation ;
> bibo:degree <http://ex.net/2> .
>
> <http://ex.net/2> a edu:PhDDegree ;
> bibo:awardingInstitution <http://ex.net/3> .
>
> <http://ex.net/3> a foaf:Organization ;
> foaf:name "Harvard" .
>
> etc.
>
>
Would suggest to re-check the example on the wiki. We tried some things
with foaf:organization, etc. but a "besst practice" has to be agreed on.

>> HOWPUBLISHED (can't quite squeeze this into a dcterms:publisher)
>>
>
> No, we don't really have anything for that. Would be nice to see some
> examples of actual data for this field.
>
>

What is that?


>> NOTE (I think there's a mechanism for this, but I'm not sure)
>>
>
> Another "thing we've discussed but not resolved."
>
>

Well, I thought we did.

Please read this blog post to see how I use Notes:

http://fgiasson.com/blog/index.php/2007/08/15/describing-documents-articles-series-volumes-and-conferences-using-the-bibliographic-ontology/


>> KEY (May just shove this into a BibTeX extension vocab)
>>
>
> Yes, I don't want this here, because it's really a user tag. It does not
> really belong as a property of the resource per se.
>
>
>> ADDRESS (particularly problematic for booklets, which lack a
>> publisher, Otherwise, I was thinking something along the lines of:
>>
>> :whatever dcterms:publisher :publisher .
>> :publisher a foaf:Agent ;
>> foaf:isBasedNear [ a spacenamespace:Space ;
>> spacenamespace:name "BibTeX-Address" ] .
>>
>
> We have the locality property for this. Another one of those details I'm
> not entirely sure about/comfortable with.
>
>

you should describe the place using

wgs84_pos:SpatialThing

then you can plugin a bibo:locality to write the string describing it.

And you have to plug this place description from your resource, to the place's resource using something like bibo:place (event:place) or whatever.

Bruce: we did we choose to delete bibo:place?

>> Perhaps with some sort of attempt at parsing out an address to get
>> more semantic information out (e.g. City, Province, country)
>>
>> ORGANIZATION (I'm thinking bibo:organizer(?))
>>
>
> The organizer property is to link associated agents to an event. A
> conference, for example, has an organizer.
>
> <http://ex.net/1> a bibo:Paper ;
> bibo:presentedAt <http://ex.net/2> .
>
> <http://ex.net/2> a bibo:Conference ;
> bibo:organizer <http://ex.net/3> .
>
> <http://ex.net/3> a foaf:Organization ;
> foaf:name "ABC Corp." .
>
>
>> INSTITUTION (Used with tech reports. bibo:producer?)
>>
>
> Yeah, probably.
>
>
>> CHAPTER (this is why we need bibo:chapter)
>>
>
> Huh? I don't think chapter in bibtex is used as Fred intended
> bibo:chapter to be used. The former is the chapter *title*; isn't it?
>
> The bibo:chapter property would be the chapter number.
>
>
>> TYPE (used with tech report.)
>>
>
> Another TBD issue. We did talk about the need for a literal to describe
> types (of letters, reports, etc.). But he hadn't resolved how. It seems
> a little silly to call it "type."
>
>
>> I'm also not sure if I like how InBook/InCollection would have to be
>> mapped right now (to bibo:Article, when something like a bibo:Chapter
>> would be more appropriate)
>>

Take care,

Fred

Frederick Giasson

unread,
Sep 26, 2007, 8:35:24 AM9/26/07
to bibliographic-ontolog...@googlegroups.com
Hi,

>
>>> NOTE (I think there's a mechanism for this, but I'm not sure)
>>>
>> Another "thing we've discussed but not resolved."
>>
>
> I figured as much. (rdfs:comment?)
>
>


No :)


As I said previously, check the blog post.

I reuses Annotea, etc.


Fred

Bruce D'Arcus

unread,
Sep 26, 2007, 8:51:11 AM9/26/07
to bibliographic-ontolog...@googlegroups.com
Pipian wrote:

...

>>> CHAPTER (this is why we need bibo:chapter)
>> Huh? I don't think chapter in bibtex is used as Fred intended
>> bibo:chapter to be used. The former is the chapter *title*; isn't it?
>>
>> The bibo:chapter property would be the chapter number.
>
> The BibTeX CHAPTER property is supposed to be the chapter number. If
> the chapter is identified by a name, it should probably be
> INCOLLECTION instead of INBOOK.

Ah, OK. It appears a little vague though:

"A chapter (or section or whatever) number."

I guess in that case, we might consider adding it back.

Bruce

Bruce D'Arcus

unread,
Sep 26, 2007, 9:14:17 AM9/26/07
to bibliographic-ontolog...@googlegroups.com
Frederick Giasson wrote:

...

>>> HOWPUBLISHED (can't quite squeeze this into a dcterms:publisher)
>>>
>> No, we don't really have anything for that. Would be nice to see some
>> examples of actual data for this field.
>>
>>
> What is that?

See:

<http://artis.imag.fr/~Nicolas.Holzschuch/bibtex.html>

>>> NOTE (I think there's a mechanism for this, but I'm not sure)
>>>
>> Another "thing we've discussed but not resolved."
>
> Well, I thought we did.
>
> Please read this blog post to see how I use Notes:

No, the BibTeX "note" is a simple literal that always gets added to a
citation. Again, see link above.

I'm looking for some really good examples of BiBTeX files that use these
fields. Here's one that at least qualifies as OK:

<http://www.tug.org/pracjourn/2006-4/fenn/ptj-bibtex.bib>


...

> you should describe the place using
>
> wgs84_pos:SpatialThing

We went through this before. I don't want to use that vocabulary
directly. It's hopelessly abstract/ugly. I think I recall preferring to
reuse frbr:Place.

> then you can plugin a bibo:locality to write the string describing it.
>
> And you have to plug this place description from your resource, to the place's resource using something like bibo:place (event:place) or whatever.
>
> Bruce: we did we choose to delete bibo:place?

I don't remember. I've a cold, so life is a little fuzzy now, but ...

I think this part is frankly a mess. We have a variety of different ways
to represent the same thing, each from different vocabularies.

We're dealing with organizations and events, and the places and/or
addresses where they are (respectively) located or held.

So what are the examples we need to handle:

1) theses. Ian's impulse is to use bibo:producer, and describe the
location of the department using FOAF. But I'm not sure a department
quite produces a thesis. It is the institutional body to which one
submits the document for approval, and which in turn grants approval and
the subsequent degree.

It might be that we just need a property like bibo:submittedTo and to
describe the department and school using FOAF.

BTW, Ian, we decided we needed degrees as individuals because there are
quite a diversity of degrees.

2) other unpublished or informally publisher resources where you might
want to include where it came from. I believe Elena gave the example of
a letter. See this post and subsequent discussion:

<http://groups.google.com/group/bibliographic-ontology-specification-group/browse_frm/thread/e245575791704c38/9720d298cc2f9c51?lnk=gst&q=elena+letter+new+york&rnum=1#9720d298cc2f9c51>

3) events like conferences and hearings; I'm fine with using the event
properties for this I guess. I'm uncomfortable with the confusion in all
this.

Bruce

Frederick Giasson

unread,
Sep 27, 2007, 8:22:38 AM9/27/07
to bibliographic-ontolog...@googlegroups.com
Hi Bruce!

> No, the BibTeX "note" is a simple literal that always gets added to a
> citation. Again, see link above.
>
> I'm looking for some really good examples of BiBTeX files that use these
> fields. Here's one that at least qualifies as OK:
>
> <http://www.tug.org/pracjourn/2006-4/fenn/ptj-bibtex.bib>
>
>

Well, you can use the bibo:Note as do within my foaf profile

You can even detect if the note is a URL only, then handling it like any
other web pages related to a document (see examples)

> We went through this before. I don't want to use that vocabulary
> directly. It's hopelessly abstract/ugly. I think I recall preferring to
> reuse frbr:Place.
>
>

yeah, but remember that it is the root :) So from there, you can use
anything else.
>> then you can plugin a bibo:locality to write the string decribing it.


>>
>> And you have to plug this place description from your resource, to the place's resource using something like bibo:place (event:place) or whatever.
>>
>> Bruce: we did we choose to delete bibo:place?
>>
>
> I don't remember. I've a cold, so life is a little fuzzy now, but ...
>
> I think this part is frankly a mess. We have a variety of different ways
> to represent the same thing, each from different vocabularies.
>
>

Yup :)

And it is why we can't do anything against it. What we can do is to
propose something and to add properties if needed.


It is why I suggested to reuse the Event ontology for events
(conferences), etc.


> We're dealing with organizations and events, and the places and/or
> addresses where they are (respectively) located or held.
>
> So what are the examples we need to handle:
>
> 1) theses. Ian's impulse is to use bibo:producer, and describe the
> location of the department using FOAF. But I'm not sure a department
> quite produces a thesis. It is the institutional body to which one
> submits the document for approval, and which in turn grants approval and
> the subsequent degree.
>
> It might be that we just need a property like bibo:submittedTo and to
> describe the department and school using FOAF.
>
> BTW, Ian, we decided we needed degrees as individuals because there are
> quite a diversity of degrees.
>
> 2) other unpublished or informally publisher resources where you might
> want to include where it came from. I believe Elena gave the example of
> a letter. See this post and subsequent discussion:
>
> <http://groups.google.com/group/bibliographic-ontology-specification-group/browse_frm/thread/e245575791704c38/9720d298cc2f9c51?lnk=gst&q=elena+letter+new+york&rnum=1#9720d298cc2f9c51>
>
> 3) events like conferences and hearings; I'm fine with using the event
> properties for this I guess. I'm uncomfortable with the confusion in all
> this.
>
>

Yeah, we should document it properly to make things clearer.


Take care,


Fred

Reply all
Reply to author
Forward
0 new messages