Exploring Topic Maps
Margaret
Hilsbos,
Introduction
I
approached this project with a particular knowledge representation problem in
mind. From reading the literature
assigned in our class, I concluded that the most likely approach to solve the
problem would be “topic maps”. During my
exploration of topic map concepts and technologies, I added other perspectives
to my original problem. In this paper I describe the perspectives of my focus
problem, my ideas of what a solution might comprise, my experiences with a few
of the available technologies, and my own prototype approach to a topic map
creation facility.
The Problem
Problem 1a.
I
originally formulated my research problem as follows.
·
Wants not one, but multiple sources to answer the question
·
Wants to see, understand, and be able to navigate relationships between
sources
·
Needs visibility to basic attributes such as authority, scope,
currency, etcetera, for evaluating the appropriateness of the resource to their
question. These “evaluation-relevant” parameters are not always easy to find.
A
specific example of this problem is the case where a person needs to find out
some information about an organization, in order to a) decide whether or not to
support the organization, or b) identify the credibility and bias of
information received from that organization.
The facts that are really needed for such an assessment are usually not
very easy to find on the organization’s website. A careful potential donor will want to
“follow the money” of large contributors, identify the organization’s allies and
detractors, know the positions that the organization has taken on critical
issues, etcetera. The common aspect of
this information, besides being difficult and laborious to collect, is that it
all involves relationships between
various subject entities, and the need to view these relationships holistically.
In
the process of my research, I found (as usual) that my notes were disorganized,
I could not remember where I saw a particular item, or I had bookmarked a
resource but could not remember why, or how I had come to that resource in the
first place (which path is sometimes as useful as the resource itself). It occurred to me that this problem was
merely another view of my original problem. In my original problem, the user
would benefit from a “resource guide” where the relationships between resources
are already identified, and easy for the user to navigate to find a quick
answer to a question. In my own
research, I needed to be able to easily build
such a “resource guide” to map my references and notes as I went along, so that
I could easily refresh my learning later. Some might think that by now, as a
graduate student, I should have figured out how to take decent, organized notes
in some traditional way. But, even if I
had, it would be a system that was almost certainly unique to me, and would not
lend itself to being read (and made sense of) by others. In business environments, there is a need for
team members to be able to produce knowledge collaboratively from their
research on various business problems; a solution to my “research notes”
problem could potentially be useful in such a collaborative knowledge-building
environment as well.
In
summary, the problems are to:
1a)
present a user with a richly annotated resource guide, which shows the
relationships between resources and their attributes, and is easy to navigate;
1b)
provide a facility for a user to take notes in a “self-mapping” fashion, that
results in a resource such as described in 1a.
Envisioning the Solution
What are Topic Maps?
Topic
mapping is a conceptual approach to organizing information, which is described
in some detail in an introductory paper by Pepper (2000);
I will only review the basic concepts here. There are several defined, open
standard syntaxes for representing topic maps (the major ones are HyTM, LTM, XTM
1.0, and XTM 1.1), but as will be seen with the
commercial software I reviewed, companies freely use the term “topic mapping”
even though they do not conform to any of these syntaxes, or even incorporate
the basic elements of topic maps as defined by Pepper. Of the free software
tools that I reviewed, XTM seems to be the common interchange format (although
some have not caught up to XTM 1.1); I performed most of my work with the XTM 1.1 specification.
The
basic features of topic maps are topics, associations, and occurrences. The following is my own summary, based on the
XTM 1.1 specification (ISO/IEC, 2004), the Topic Map
Data Model (ISO/IEC, 2003), and readings in Park and
Hunting (2003):
Topics are used to refer to a
subject, whether conceptual or physical, and whether an online resource or some
other type of subject. Associations are used to describe
relationships between topics. Occurrences
are similar to the page numbers given for a subject in a back-of-the-book index
– an occurrence does not necessarily describe the topic, but does refer to it
in some way. The topic definition
should include a “subject identifier” – if the topic refers to an online information
resource, the subject identifier is the URL for the resource (a resourceRef in
the XTM schema); otherwise, one or more
“subject indicators” should be provided.
A “subject indicator” is an “xlink” uri that may point to a url or to
another topic.
I
identify the real power of the topic map model as the ability to “reify”
associations and occurrences as topics, so that something can be said about the
association or occurrence itself. For example, if the topic “Margaret Hilsbos”
is associated with the topic “Drexel University” with the association “student
at”, then it is useful to note the validity period of the association (i.e.
from 2002 to the present), or the student’s degree program (i.e. MSIS). Topic maps provide this capability via the
reification mechanism; this capability makes it possible (at least in theory)
to endlessly expand the knowledge represented in a topic map.
There
are several publicly available topic maps on the Internet that illustrate some
possible uses of the topic map approach. I've listed a few of these in Appendix A. Those
new to the concept of topic maps might find it helpful to explore these
examples.
How might Topic Maps be used
to solve the problem?
Early
in my research, I was particularly interested in developing some confidence
that the topic map paradigm was a good approach to the problem I had
formulated. First I discovered Omnigator, which includes some browseable online topic
maps; later I discovered several other topic maps available on the web. While
exploring these examples and experimenting with various software approaches, I
became convinced that topic maps could be used to address my problem. I then tried to envision what components would
be important to make such a system useful.
The following is the list I came up with. For a personal (single-user)
application, only the first two bullets are applicable. The subsequent bullets
address requirements for a collaborative environment.
Components of a Useful
Collaborative Topic Mapping Tool
·
Information Retrieval interface.
·
Construction interface.
·
Contribution interface.
o
Handling of “draft” and “submitted” versions
o
Handling of “public” and “private” data
o
Normalization and merging of public submissions from multiple
contributors
·
Basic infrastructure, such as efficient database storage, access, and
backup.
·
Maintenance and support interface.
In
the next section, I describe my evaluation of a few of the available tools to handle
the basic functions for the first two major bullets above – construction (i.e.
building topic maps) and retrieval (i.e. browsing topic maps).
Component Evaluation – or,
“My Journey in Topic Map-land.”
I
spent at least a little time experimenting with each of the tools listed in
Table 1 below. The tools are described
in more detail in the evaluations below; Table 1 provides a comparison of some
key features.
Table 1
|
Tool Name |
Process Area |
Pricing Model |
Import Formats |
Export Formats |
Edit
Capability |
|
Omnigator |
View
(tabular) / Edit |
Free for
non-commercial use |
XTM, LTM, HyTM |
XTM, HyTM,
RDF/XML, CXTM |
LTM (must
activate plug-in) |
|
TM4Jscript (aka JTMA) |
Build / View
(tabular) |
Free |
XTM, if first
transformed with provided XSLTs |
JavaScript
(for use by this program); XTM |
Tabular |
|
TMTab/Protégé |
Build, with
Ontology Rules |
Free |
n/a |
XTM |
Tabular |
|
PersonalBrain |
Build / View (graphical);
Spider |
Commercial (free trial) |
Other .brn
files; internet bookmarks; folder
structure |
XML (non-XTM);
loses types; requires plug-in |
Graphical |
|
Grokker |
View
(graphical); automatic web mapping
based on search criteria |
Commercial (free trial) |
Only open
other Grokker files, or create new file from web search or hard drive search. |
Outputs a
"GXML” file |
Graphical |
I
began my Topic Map exploration with Omnigator from
Ontopia. Omnigator is a topic map
browser with a tabular, text-based display.
It requires an installed Java Runtime Environment (JRE) of version 1.3
or higher, and it creates its own standalone Tomcat server installation. Running
the software requires two steps: starting the server, then accessing the start
page in a web browser. This feature is an inconvenience, but one of few
indications that one is running “free” software.
I
installed and explored Omnigator early in the project, and found it very
helpful for understanding the concepts and structures of topic maps. Ontopia
provides several example topic maps, as well as a tutorial for building a first
topic map. A good example topic map for
becoming familiar with Omnigator is the CIA World Factbook Topic Map
(“factbook.hytm”) which can be explored at the Ontopia website [TM-1],
without even installing the software.
The download includes some smaller example topic maps (World Factbook is
quite large, with 754 Topics, 2456 Associations, and 786 Occurrences).
After
exploring some example topic maps, I started out to create a topic map of my
own. Omnigator is not a topic map
creation tool itself, so it was
necessary for me to learn the XTM syntax and create my topic map in an XML
editor. Not having an XML editor when I started, I began creating my file in
Notepad and reviewing it in Internet Explorer (which displays XML files with a
default stylesheet that makes them easier to read than in Notepad). However I soon
realized the necessity of a better tool, and was lucky enough to find Cooktop (Pavlov, 2003), which is a free XML editor that
proved a welcome addition to my hard drive.
With
my first problem in mind, I created a very simple “topic map” (well, actually
only 3 topics, and no associations). I
successfully loaded this mini-map into Omnigator, but I was disappointed by how
much work I’d had to do to create such a miniscule result. Thinking “there must be a better way”, I
looked for a tool which billed itself as a “topic map creation tool” (which
Omnigator does not). I located TM4Jscript, and
digressed from my Omnigator evaluation briefly to determine if TM4Jscript
improved the task of topic map creation. The results of that test are described
below. For now suffice it to say, after trying
TM4Jscript I still felt the need for a better way. I had discovered several
other tools in my search (see Appendix B), but they
all seemed to require either technical know-how or software environments with
which I was inexperienced (such as Perl or Python).
At
this time I found that I was having the usual trouble keeping my notes and
sources organized (problem 1b above), so I thought it
would be interesting to try to start capturing my research notes as a topic
map. However, without an easy-to-use
topic map creation facility, this could
have been a very inefficient way to conduct research! I considered that it might be relatively easy
to construct formulas to generate the XTM code from elements entered in a
spreadsheet. So I started listing my notes in an Excel worksheet, which I
thought might lend itself to generating the XTM. The code generation turned out
to be more complicated than I anticipated, but it was perhaps the most useful
exercise for me to learn the nuances of XTM (at least, the nuances of the
elements I implemented). I maintained and developed the spreadsheet
approach throughout the project, ultimately resulting in a prototype topic
map of reasonable size (77 topics, 13 associations, 13 occurrences).
Another
concern I had developed by this time was the browser interface style used by
both Omnigator and TM4Jscript
– tabular and text-based. I had
originally envisioned a more graphical system, perhaps with a hyperbolic tree
structure. I located two commercial
software packages (Grokker and PersonalBrain)
and one open-source software package (Hypergraph)
that claimed to provide visualization of topic maps. Due to time limitations, I only reviewed
Grokker and Personal Brain, which both offer free 30-day trial versions. I chose these over Hypergraph because I hoped
that, being commercial packages, they would represent the “high end” of
capability, and also because I found no installation or user guide
documentation available for Hypergraph.
As
I worked on my own Excel-based “XTM generator”, I found that between sittings I
would forget my “implicit ontology” – whatever meaning I had in mind for various
terms that I used to type topics and associations – or I would struggle with
choosing a correct term for a new type.
This experience motivated me to lean towards Vatant’s argument (2004) that topic mapping is more effective when the
builder commits to an explicit ontology, either one defined for the specific
topic map under construction or one chosen from existing published
ontologies. As Garshol (2004) notes, the term “ontology” can have many
meanings. I see value for topic maps in ontologies as Garshol and Vatant
describe them, which I think of as “taxonomy + vocabulary + validation”;
"validation" meaning, for instance, that I can define a rule that
says one cannot assign a topic which is an instance of “bookshelf” to the role “author”.
A
brief foray into ontology-land led me to TMTab, which is a
tool for creating an ontology (or a topic map with ontology control) which can be exported as a topic map in XTM
syntax. After exploring TMTab, in the
process of wrapping up the project, I revisited Omnigator
and discovered some features which I had not noticed (or recognized) earlier:
At
the conclusion of this project I have several versions of “prototype” topic
maps, created by various tools, which can be viewed in Omnigator;
or in theory, in TM4Jscript (by applying the provided
XSLT’s to convert them to JavaScript first). I also have a large “PersonalBrain” file representing a significant chunk
of my research (which apparently cannot be ported easily to any other tool – I may
have to buy a license). The most
relevant files are included in an archive file which may be downloaded from http://www.pages.drexel.edu/~mah55/info653/termpaper_supplement.zip.
A list of the included files along with a brief description is provided in Appendix C.
In
the sections below, I briefly describe my findings relative to each tool.
Omnigator
Installed
version: 007, v2.0.3, Build 2004-03-15 #1349
Available
from: Ontopia at http://www.ontopia.net/download/freedownload.html
I
found Omnigator to be the best of the free tools, especially for learning. Except for a decent interface for building a
topic map, Omnigator has it all: import and export capability for a variety of
open-standard formats, a reasonable user interface for browsing, and the capability
to build ontology rules as a schema definition (although I didn’t test that
feature). There are several other useful
features, such as a statistics page. On
the export page, files can be previewed in the various formats, which is an
interesting way to compare the different syntaxes. It is possible to add topics
and associations using Ontopia's LTM syntax. LTM is a
simple, easy-to-use syntax, but it does not support all elements of XTM 1.1.
Also, if non-semantic identifiers are used (as seems essential in a large map,
to ensure uniqueness), then it is difficult to construct an LTM statement and
be sure of correctly referencing the intended identifiers. Still, it is a nice
feature for quickly demonstrating the addition of elements to the topic map.
Overall,
I would definitely recommend Omnigator to anyone with even the slightest
interest in exploring topic maps.
TM4Jscript (also called
JTMA – Javascript Topic Map Application)
Installed
version: 0.9
Available
from: http://tm4jscript.sourceforge.net.
Once
installed, TM4Jscript can be started from a browser shortcut. TM4Jscript provides a nice tabular layout for
building a topic map, as well as a basic tabular viewer to view the loaded
map. It will export a map in XTM
format, but it cannot directly import an XTM file. I found the documentation
somewhat hard to follow, such as when to use the “Add new topic as subtype”
button vs. “Add new topic as instance”. I ran into trouble on my first small
map attempt, because in the XTM export, TM4Jscript inserted the “instanceOf”
clause in the wrong place for each occurrence element. As this was early in my study, I could not
identify the source of the problem at the time. I only knew, at that time, that
I could not load the exported XTM file from TM4Jscript into Omnigator, until I
commented out the problematic clause on the occurrences – which had the effect
of removing any typing from the occurrences.
I also disliked that TM4Jscript insists on a top-level topic named
“ur-type”. Between these two “features” I was quickly disenchanted with
TM4Jscript. However, revisiting the
software after more experience, I think it may be possible to overcome these
issues, or at least live with them; the interface is otherwise easy to use,
compared with the other (free) alternatives I explored.
TMTab/Protégé
Installed
version: Protégé – 2.1 (beta, build 197); TMTab – 0.50
Available
from: Protégé - http://protege.stanford.edu/ ;
Techquila -http://www.techquila.com/tmtab.html
As
mentioned above, I reviewed the TMTab/Protégé approach because of my interest
in how an ontology might be applied while constructing a topic map. TMTab runs
as a plug-in for the Protégé Ontology Editor, so it was necessary to install
Protégé first, do the tutorial, and practice with the interface a little before
attempting to add TMTab. I found Protégé
quite easy to use on its own, but I found TMTab a little confusing. Once I successfully installed TMTab (after
realizing I had installed the wrong version of Protégé and correcting the
situation – pay attention to TMTab’s release notes), my next issue was with the sparseness of the
user guide. Given that this is an early
(v 0.5!) version of a free software, I can’t complain too much, but I did
struggle with understanding the intended mode of usage. The plug-in provides classes for associations
and occurrences, which are described briefly in the first part of the user
guide; but then the user guide details an easier and more useful way to build
associations and occurrences (supporting a hierarchy), where those provided
classes are not used. Eventually I was
able to create and export a reasonable-size prototype XTM with 44 topics, 17
associations, and 2 occurrences. The XTM
file displayed with no errors in Omnigator, but there were some quirks. For
example, all superclasses in the hierarchy are displayed for a topic, rather
than just the immediate parent.
Overall,
I found TMTab more difficult to use than TM4Jscript. However, once the ontology
is built, adding instances is quite easy. The benefit of the ontology structure
is apparent, as the possible associations for a selected topic type are
specified in the instance window. Of
course, the ontology control is only provided within the TMTab/Protégé
application, so once the XTM file is exported, it should not be edited
externally.
Personal Brain
Installed
version: 3.02 (
Available
from: The Brain Technologies Corp.,
at http://www.thebrain.com/.
PersonalBrain
is a commercial information mapping and visualization software. Once I became
accustomed to the interface and made some crucial customizations (for example,
lightening the background so that I could actually read the topic
names), I found it quite easy to use for data entry, and very helpful in
information retrieval. It became my tool of choice for actually recording
research notes quickly. PersonalBrain
has a good graphical navigation interface, so that once I recorded information,
I was confident not only that I could retrieve that item quickly, but also see
its context of related items at the same time – something that Omnigator does
not do well, and probably is not even possible with a tabular, text-based style
interface.
PersonalBrain
seems to be targeted at average personal computer users, so it does not
emphasize the topic map paradigm; the user guide does not talk about topic map
constructs such as topics, associations, and occurrences. Instead it uses the
terms "thoughts" and "links" for topics and associations.
Occurrences are not really supported, except as entries in the
"notes" for a "thought".
A new "brain" is created with a top-level "thought",
and every "thought" added must be linked to an existing thought. It
is possible to assign types to thoughts and links, which is similar to typing
of topics with an "instanceOf" element in XTM. Types can be color-coded, and there are a few
implicit types, such as for url’s. Web url’s can easily be added as new
thoughts, or added to existing thoughts.
A
PersonalBrain file can be exported as XML[1] (not
XTM), but unfortunately it appears that association types and thought types are
not exported. Still, it may be possible to devise a transformation to convert a
PersonalBrain file to XTM, in order to share the map with users who don't have
a PersonalBrain license.
Grokker
Installed
version: 2.01
Available
from: Groxis, Inc. at http://www.groxis.com/service/grok/
Of
the software I reviewed, Grokker was the least useful for my purpose. I
selected Grokker to review for two reasons: my initial title for my project was
“Grokking Topic Maps”, before I heard of
the Grokker software; then I read the
review by
Grokker
essentially provides visualization and filtering of meta-search results. It
uses plug-ins (several of which come pre-installed; I did not add any) to cull
results from several Internet search engines.
The results are categorized and grouped hierarchically by Grokker’s own
logic initially; then the user may remove items, or move items to different
categories, or create their own categories.
Grokker does indeed provide a different visualization technique than the
other software I reviewed. However,
while Stirling refers to it as a topic-mapping
software, I find that to be a stretch. The only manipulations I was able to do
involved hierarchical categorization; there did not appear to be any other ways
to create associations between “topics”.
As I progressed in my research it occurred to me
to attempt to “topic map” the resources I was locating in my research. A main reason for this was my dissatisfaction
with my previous methods of keeping track of references – essentially a list of
citations, with (sometimes) notes regarding how I located each resource and the
relationships I identified between resources.
Since these lists and notes are never as organized as a topic map would
be, I thought it would be interesting to prototype a process for recording
notes in a form that could easily be added to a topic map.
I
chose Excel for my prototype only because it is a format with which I am fairly
experienced, and I anticipated that using a spreadsheet might avoid the need to
program an input form. My general approach was to establish simple tables for
entering data, with a set of background sheets with formulas to generate the
XTM code. After entering several topics, occurrences, and associations, I
manually transferred (by copy and paste) the XTM code from Excel to my XML editor (Cooktop 2.5).
Once the XTM file successfully validated in the editor, I loaded it into
Omnigator and reviewed the results. With some
tweaking, I became fairly satisfied with the ability of my prototype to handle
the basic XTM elements.
As
an application, my prototype is not quite "ready for primetime" –
some manual manipulation is needed in order to export the XTM from new entries,
and only the basic XTM 1.1 elements are supported. It does not support multiple
baseNames, variants, or scopes except as used for association roles. In the
time for this project I could not do any macro programming at all – all
calculations are based on embedded formulas, and depend on manually extending
the appropriate spreadsheet ranges before making entries (a simple exercise for
me, but difficult to explain to someone else).
Once the ranges are extended, XTM fragments are generated on hidden
worksheets as new entries are made. When I am ready to update the actual topic
map XTM files, I copy and paste columns of code fragments into an XML editor
and validate the result. If there are no
new types, only instances, it is necessary to
1) combine occurrences with
topics (since there is a 1:n relationship of topics to occurrences, this would
require a macro to perform automatically);
2) copy and paste the new topic
and association fragments into the editor.
New topic types are simple, as I made no provision
for occurrences on types – so it is only necessary to copy and paste the topic
type fragments as automatically generated. New association types require new
role types to be copied as well (I tended to forget this, and the roles then
showed up as "no name" topics in Omnigator).
While this may sound like it requires a fair amount
of effort to create the XTM file, when I simply could have typed the code in
the editor, there are some advantages, even in the current state of the
prototype:
1) It provides a place to
record information on-the-fly, without worrying about the code behind it;
2) Element identifiers (id
attribute) are generated and referenced automatically. Uniqueness is guaranteed
and the user doesn’t need to conjure up identifiers with semantic content.
An example of my final "Excel-generated"
XTM file is available in the archive file provided as a supplement to this
paper. I have also included the Excel
file itself. To view the generated XTM fragments, unhide the sheets labeled
"Tagged-…" (from the Excel menu, Format > Sheet > Unhide). I
make no claim that you will be able to actually use this file to generate XTM! However, the XTM file can be browsed in Omnigator, which can be downloaded and installed free of
charge, and the Excel formulas may be of interest.
Conclusion
In
the course of my research I have identified an array of topic mapping tools and
conceptualized several useful applications of topic maps with these tools.
However with the exception of the commercial packages, all would require some
implementation effort before being applied to a real-world problem. Of the two
commercial packages, only PersonalBrain allows the
user to construct a topic map from scratch, based on the thoughts relevant to
the user's current focus. This
capability seems more useful to the problems which I had in mind, than Grokker's reverse approach, which is to provide an initial
map of existing web resources and then let the user select those of interest.
Of
the free tools, I consider Omnigator an indispensable "first
step" for understanding topic maps, and for browsing topic map
prototypes. Also, the publicly available
topic maps listed in Appendix A can provide the novice
with some sense of the power and potential of topic maps.
Regardless
of the tool, I am enthusiastic about the potential for topic map technology,
especially when coupled with appropriate visualization techniques, to
revolutionize knowledge harvesting and knowledge representation. The technology is still young, and there is
plenty of room for growth. It will be interesting to see, in five years or so,
whether topic map technology becomes embedded in common tools, or if it remains
the purview only of knowledge specialists.
Appendix
A: Example Topic Maps on the Web
[TM-1] CIA World Factbook at
Ontopia http://www.ontopia.net/omnigator/models/topicmap_complete.jsp?tm=factbook.hytm
[TM-2] Wilde's WWW Online
Glossary (This is an extremely useful resource, as well as a good example of an
application of topic map technology) http://dret.net/glossary/
[TM-3] Techquila’s Topic Map
World. http://www.techquila.com/topicmaps/tmworld/.
[TM-4] Proceedings of the
Third GCA International HyTime Conference,
[TM-5] A listing of topic
maps on the web. Many of these are XTM
files and would need to be loaded into a browser such as Omnigator. http://www.topicmapping.com/registry.html.
[TM-6] "Long Island Sea
Creatures" – a website generated from an XTM file using an XSL file. The construction of this topic map is
described in Park and Hunting (2003).
Website: http://www.cogx.com/ctw/seacr
XTM file: http://www.cogx.com/ctw/seacr.xtm
XSL file: http://www.cogx.com/ctw/book.xsl
Appendix B: Lists of Free Topic Map Tools
EasyTopicMaps (2004, April
20). Topicmap Free Tools. http://www.easytopicmaps.com/wakka.php?wakka=TopicmapFreeTools&v=91h
Woodman,
Garshol, Lars Marius (2004, February 20). Free XML Tools. http://www.garshol.priv.no/download/xmltools/cat_ix.html#SC_TMEngine
Appendix C:
Contents of the Supplement File
|
File Name |
Description |
|
Excel_XTM_Gen.xls |
Excel file used to
generate XTM from spreadsheet entries. The file was saved in a format which
displays only the pages and columns for inputting topics, associations, and occurrences.
To see the code, unhide the hidden sheets and columns (Format > Sheets
> Unhide). |
|
TMresources_XLS.xtm |
XTM file based on the
Excel file. |
|
TMResources_TMTab.xtm |
TMTab XTM output |
|
uia_2_JTMA.xtm |
TM4Jscript XTM output
of first “mini-map, with offending “instanceOf” clauses commented – easy to
see if you open the file in IE 6. Refer to the XTM 1.1
spec to verify that the “instanceOf” clause was placed out of sequence by
TM4Jscript. |
|
TMresources_JTMA.xtm |
Another TM4Jscript XTM
file with some of the TM Resources topics used in the other maps. |
|
TM resources_PB.xml personalbrain.dtd |
The PersonalBrain XML
file and associated dtd. This file seems
to contain all of the information in the PB file, except topic and
association types. |
|
Topic Map
Resources.brn Topic Map
Resources_brn (folder with .rtf files) |
The PersonalBrain .brn
files. If you have PersonalBrain or
install the free trial, you should be able to import these files. |
References
Brain
Technologies Corp. (2004, March 1). PersonalBrain (Version 3.02) [Computer software]. Available at http://www.thebrain.com/.
Garshol,
Lars Marius (March 2004) “Metadata? Thesauri? Taxonomies? Topic Maps! Making
sense of it all.” Retrieved
Gibbs, Mark (2004). Grokker
groks searching. Network World,
Groxis,
Inc. (2004). Grokker (Version 2.01) [Computer software]. Available at http://www.groxis.com/service/grok/.
Hypergraph.
[Computer software]. Available at http://hypergraph.sourceforge.net/example_tm.html.
ISO/IEC JTC1/SC34. (2002, May 19). Topic Maps.
ISO/IEC 13250 [HyTM specification].
Retrieved
ISO/IEC JTC1/SC34. (2003,
November 2). Topic Maps — Data Model. ISO/IEC 13250-2 [TMDM].
Retrieved
ISO/IEC JTC1/SC34. (2004,
January 25). Topic Maps — XML Syntax. ISO/IEC 13250-3 [XTM 1.1 specification]. Retrieved
Ontopia (2002, March 15).
Omnigator 007 (Version 2.0.3, Build 2004-03-15 #1349). [Computer
software]. Available from: Ontopia at
http://www.ontopia.net/download/freedownload.html
Ontopia (n.d.). LTM: The
Linear Topic Map Notation. [Website]. http://www.ontopia.net/topicmaps/ltm.html.
Park, Jack, and Hunting, Sam (Eds.). (2003). XML Topic Maps: Creating and Using Topic Maps for the Web.
Passin,
Thomas and Johannesen, Alexander. (2003). TM4Jscript: Topic Maps for Javascript
(Version 0.9). [Computer software]. Available at http://tm4jscript.sourceforge.net.
Pavlov, Victor. (2003). XML
Cooktop (Version 2.5) [Computer software].
Available at http://www.xmlcooktop.com/.
Pepper, Steve. (2000) The Tao of Topic Maps (June 2000). Retrieved
Protégé Ontology Editor (Version 2.1 (beta), Build
197). [Computer software]. Available at
http://protege.stanford.edu/
Techquila. (2004, May 10). TMTab (Version
0.5.0). [Computer software]. Available at
http://www.techquila.com/tmtab.html. [Release Notes]
available at http://www.techquila.com/tmtab-status.html.
TopicMaps.Org.
(2001, August 6). XML Topic Maps (XTM)
1.0: TopicMaps.Org Specification. Retrieved
Vatant, Bernard. (2004)
Ontology-driven Topic Maps. Conference
Proceedings, XML Europe 2004.
[Paper] Retrieved
[1] Although I must have seen it at some point, I could not locate the place in the documentation that instructs how to do this. The key is to rename the output file of the exporter to “.zip” (from .pba) and use any unzip program to extract the files.