Text Analysis Portal for Research, Using the Public Release

Introduction
In April of 2007 the TAPoR project, after extended testing
and two public beta versions, released Version 1.0 of the
Text Analysis Portal for Research.1 This portal gathers text
analysis tools configured as web services into an environment
where users can define the texts they want to work with, select
the tools that best work with different text formats, and log
their research. The portal is, on the one hand a broker for piping
electronic texts to tools and then managing the results, and on
the other hand is a study environment where projects can be
managed over time. Our poster will do three things:
• Show an overview of how the portal works with annotated
screen shots.
• Show the life of the project including the user testing phases.
• We will have a laptop in order to demonstrate a live version
of the portal to interested.
Demonstration
Provided we can have a live internet connection we will be
prepared to demonstrate:
• How the portal can be used even without an account using
the Try It feature or myLinks pages where users can publish
sets of texts for analysis.
• How to get a free account from TAPoR.
• How to define texts that may be elsewhere on the internet
or uploaded. How to organize these texts using tags and
how to publish a set of searchable texts for students or
colleagues.
• How to use the Workbench to organize a set of texts and
specific tools for a project. How to run tools on texts and
how to save results. • How to log the progress of a research project and share it
through TAPoR.
If a live connection is not available we will have a local version
running based on the TAPoR Live CD which is a bootable
version of linux and the portal that works locally.
User Testing and the Life Cycle of a
Tool Project
Version 1.0 of the portal is the result of extensive user
testing which will be visually represented in the poster
as will ways new users can provide us feedback as we develop
the priorities for further releases. 2
For those familiar with earlier versions of the portal there are
a number of improvements including:
• A new type of news channel called a Research Log. Users
can save results from tools to this log along with comments.
They can also save the text/tool configuration to run again.
• A new feature called Analyze Text that will open a window
with the text on the right and suggested tools on the left for
close study of one text.
• The myLinks feature has been adapted so users can publish
their public texts for others to search across and for other
to analyze. This can be used, for example, for providing
students with a study set of texts.
• The interface has been simplified to make it easy for users
to get projects going.
Background
The TAPor Portal is one of the outcomes of a 5 year project
funded by the Canada Foundation for Innovation (CFI)
that involves 6 universities across Canada, including the
University of Victoria, the University of Alberta, McMaster
University, the University of Toronto, the Université de
Montréal and the University of New Brunswick. Provincial
funding bodies and our respective universities have also
supported the TAPoR project. The portal was developed by
Open Sky Solutions with a team primarily at McMaster
University. The poster will recognize the supporting institutions
but will focus on the portal and not present the wealth of other
projects that are part of TAPoR. 3
This poster will show the current version of a major text analysis
tools development effort that developed a model for sustainable
and modular text analysis for a broad community. This is a
challenge that has concerned computing humanists since the
1980s. As Susan Hockey put it in a post to HUMANIST in
1996 that reported on a meeting she convened at CETH, "For
some time, those of us active in humanities computing have
felt the need for better and/or more widely accessible text
analysis software tools for the humanities. There have been
informal discussions about this at a number of meetings, but
so far no substantial long-term plan has emerged to clarify
exactly what those needs are and to identify what could to be
done to ensure that humanities scholars have readily-available
text analysis tools to serve their computing needs into the next
century."4
The TAPoR model meets a number of the objectives described
for text analysis tools in the 1990s by Hockey and others.5 It
is widely accessible. It provides a long-term model using web
services to bring tools together where they can be used in a
study environment. It allows new tools to be added and existing
tools to be improved or replaced over time as new needs
emerge. It allows research to be recorded and shared.
Future Plans
The TAPoR Portal also has limitations. Depending on tools
provided as web services by others increases the chances
of simple operations not working. Many of the tools are not
multi-lingual. The tool broker model where the portal gets a
remote text and passes it to a remote tool is not as efficient as
an all-in-one model where texts can be preindexed. And, as
with any large software project there are still inconsistencies,
awkward interface panels, and bugs.
The poster will conclude by outlining future plans. TAPoR
hopes to be funded for a second phase through the CFI Leading
Edge Fund in order to add large-scale text datamining and
aggregation tools. Currently the portal model does not pre-index
large collections (except in limited cases) and provides limited
aggregation or crawling tools. TAPoR 2, if funded, will continue
improving the interface and will add large-scale functionality
by adapting crawling, scraping and data mining tools to the
study space.
1. For more about the TAPoR project see <http://www.tap
or.ca>
2. TAPoR interface at <http://www.tapor.ca/interf
ace> details a persona oriented investigation conducted by
Audrey Carr and Joanna Dacko. Dr. Wendy Duff at the University
of Toronto also conducted extensive interviews that have not been
published, but which were absracted for the portal development
team. Finally, we have a layered testing process combined with
the gathering of tool statistics that is being conducted on the
release.
3. Mind Technologies edited by Raymond Siemens and David
Moorman (U of Calgary Press, 2006) includes a number of articles
about the variety of projects supported by TAPoR. 4. Susan Hockey, Humanist Discussion Group 10.54 (23 May 1996).
See <http://lists.village.virginia.edu/l
ists_archive/Humanist/v10/0054.html>. There
is also a note from Michael Sperberg-McQueen pointing to a trip
report available at <http://tigger.uic.edu/~cmsm
cq/trips/ceth9505.html>
5. Geoffrey Rockwell and John Bradley, "Eye-ContTact: Towards
a New Design for Text-Analysis Tools", CHWP A.4. (1998). See
<http://www.chass.utoronto.ca/epc/chwp/
rockwell/>