Some questions about modeling journal article sections
Larry Hunter
Folks,
As a result of some conversations at ICBO, I have decided to take a
crack at modeling journal articles and related phenomena for IAO.
Before I go too far down the path, I wanted to get a sense from you
all about how things got to be where they are now.
1. There's a confusion about what is a "report," a "report part" and
"journal article content" that is so bad I put it up on the tracker as
an individual issue. The report comment says "Examples of reports are
gene lists and investigation reports. These are not published
(journal) articles, but may be included in a journal article." The
text definition of report is "A narrative object assembled by an
author for the purpose of providing information to an audience" which
would seem to me to include a journal article. My question is what
the motivation was for the apparent desire to exclude journal articles
from this class?
2. There appear to be only three direct subclasses of narrative object
(report, study interpretation and report element), but there are
references to "articles" and "patents submissions" (sic) in the
textual comments in narrative object term. Were these intended to be
subclasses of narrative object? Was the problem that they include
images, etc. which are not compatible with the narrative object
definition? Or was there some other reason?
3. Report element is defined as "a narrative object in which
information is presented and consumed by a human being" and report is
defined as "A narrative object assembled by an author for the purpose
of providing information to an audience" which seem indistinguishable
to me. The commentary around report element suggests that the intent
is that they be non-textual, like figures. That would be clearly
distinguishable from report, but would perhaps best be called
something like a "figure" -- report element connotes text to me. But
figures need not be part of a report (even though they often are).
Furthermore, there is a sibling to narrative object called "image".
Some of the commentary around narrative object asks explicitly whether
(but not answering) whether we mean only text.
To get discussion going, I propose the following rough sketch of a
rework of the ICE hierarchy here (material in parens is approximately
the differentia from the parent):
Information Content Entity (artifact intended to be understood by
humans and that is about something; define author as a creator of this
sort of artifact)
Narrative Object (a textual expression of a set of propositions;
perhaps rename to "text" or "textual entity" -- narration implies
description of an occurrence, cf exposition)
Citation (identifies a specific publication)
Author identification (identifies a specific author)
Institutional identification (identifies a specific institution)
Caption (describes a figure, as defined below)
Title (names a document, as defined below)
Table (a two dimensional arrangement of texts repeated at regular
intervals across a spatial range, where the spatial relationship among
the texts expresses propositions)
Table of abbreviations (texts are abbreviations and their
expansions)
Figure (a two dimensional arrangement of information content entities
where the arrangement itself is about something)
Diagram (Figure that expresses one or more propositions)
Image (as currently defined)
[Table would be inferred to be a figure, although the asserted is-a
should probably be a narrative object]
Document (a collection of other information content entities intended
to be understood together as a whole)
Publication (a document that has been accepted by a publisher; need
to define publisher outside of IAO)
Scientific publication (about an investigation)
Patent (a document that has been accepted by a patent authority;
need to define patent authority outside of IAO)
Document part (part of a document)
Abstract (summary of a document, substantially smaller than the
document it summarizes)
Introduction (part of a scientific publication that is about the
research objective and hypothesis of the investigation)
Methods section (part of a scientific publication that is about the
study design of the investigation)
Results section (part of a scientific publication that is about one
or more data collection or study result)
Discussion section (part of a scientific publication that is about
study interpretation)
References section (part of a scientific publication that enumerates
citations)
Author list (part of a scientific publication that enumerates its
authors)
Institution list (part of a scientific publication that enumerates
the institutional identifications associated with each author)
Author contributions section (part of a scientific publication that
is about the specific contributions of each author to the investigation)
Acknowledgements section (part of a scientific publication that is
about contributions to the investigation by people or institutions
other than authors)
Footnote (part of a scientific publication that is about a specific
other part of the same document)
Supplementary material (part of a scientific publication that is
segregated from the rest of the document due to its size; synonym for
appendix)
Your comments most welcome. When I figure out how to bend Protege 4
to my will sufficiently, I will produce a formal version of the above
(taking into account any discussion that happens in the mean time).
I am also thinking about whether it is worth adding things like
sentence and paragraph, and where they would go, and whether the
current definition of symbol should be a subclass of narrative object,
but that isn't crucial at the moment.
Larry
Larry Hunter
On Jul 28, 2009, at 9:23 AM, Kevin B. Cohen wrote:
> Should this either have subclasses for figures and tables, or be
> redefined to describe figures and tables?
Trivial to subclass to Figure caption and Table captions.
L
Michel Dumontier
sections... we've been using [1] and the overlap is nearly exact, so
i'll be happy to switch to IAO when they become available there.
Your list looks great, i might suggest the following for inclusion
* running title
* keyword section
* copyright section
Review articles may also contain
* table of contents
* table of figures
May I also suggest that the section descriptions be written generally,
so that they may apply to unpublished documents, as opposed to
scientific publications - although we may want to think about stating
that scientific publications *should* contain a set of those parts.
----
I would also suggest that we further attempt to detail document types
- and Alan identified during his tutorial, there are many existing
vocabularies (*bibtex*, Bibo, Cito, SWAN, MARC, Medline) - here's a
really rough first pass.
* document
* booklet (A document that is printed and bound, but without a named
publisher or sponsoring institution)
* manuscript (An unpublished document)
* note (A brief document.)
* publication (A document that is made available by a publisher)
** blog (A collection of web pages that acts as an online
journal.)
** book (A set or collection of written, printed, illustrated, or
blank sheets, made of paper, parchment, or other material, usually
fastened together to hinge at one side.)
** article (A publication by one or more authors in a journal,
magazine, newspaper, or any other periodical publication)
** manual (A publication that instructs on the usage of a
particular device)
** technical report (A document that describes the process,
progress, and or results of technical or scientific research or the
state of a technical or scientific research problem.)
** thesis (A written research component of a post-secondary
institution that contains a hypothesis supported by arguments)
* periodical (A publication that appears in a new edition on a
regular schedule).
** journal (A periodical in which scholarship relating to a
particular academic discipline is published.)
*** scientific journal (A journal that is subject to peer
review).
** magazine (A periodic publication containing pictures and
stories and articles)
[1] http://semanticscience.googlecode.com/svn/trunk/ontology/publication.owl
Cheers!
-=Michel=-
Barry Smith
A two-day training course in ontology, to be held in Buffalo, NY on
October 10-11, 2009. The course will consist of lectures given by
Barry Smith accompanied by a series of highly interactive
discussions. It is designed to be of interest to both philosophers
and those with a background in computer and information science. No
prior knowledge of ontology is presupposed.
Details are provided here:
http://ncorwiki.buffalo.edu/index.php/Ontology_Class_2009