Making a Contribution: Modularity, Integration and Collaboration Between Tools in Pliny

Computing tools have been an issue since the foundation
of Humanities Computing, and building modular tools
that work together has been recognised as important since at
least the CETH meetings in 1995. Geoffrey Rockwell and I
first raised issues of modularity in our paper given at the
Canadian Learned Societies Conference in June 1992 entitled
“Towards new Research Tools in Computer-Assisted Text
Analysis”. Our proposed tool framework combined data-flow
tool modularity with the ability to create pages that mixed
scholarly writing with interactive elements like in a
Mathematica “Notebook”. Of course the WWW was sweeping
all this away by 1995 and when we took a paper (Bradley,
Rockwell 1995) to the CETH “tools” meeting we were merging
our modular view with the WWW as it was then emerging.
In our community “modularity” often comes down to the
sharing of file formats. Separate tools that can all read the same
file format can all operate on common data, and therefore can
contribute their particular facilities to the task at hand. File- or
Unix-style modularity is a part of what I called the
“transformation model” of computing (Bradley 2005), and
provides a powerful approach to manipulating data. TuStep
(described in Ott 2000) is a splendid example. The model holds
a strong attraction within the Digital Humanities community –
even extraordinarily creative projects as TAPoR’s text analysis
portal (TAPoR 2006) or the Nora project (Nora 2006) are, in
fact, primarily based on this. However, thinking of modularity
primarily in those terms limits our views to those of computing
about 25 years ago – before the advent of the graphical user
interface (GUI).
The GUI radically changed the way we think about computing,
and even today, more than 25 years later, its significance
continues to reverberate. It is surely true to say that most
humanists have the GUI as a model driving how they think of
their computer. The impact has been twofold:
• First, users now expect that the computer screen will allow
them to directly manipulate materials that interest them.
Think of the words in a word processor.
• Second, interaction between tools needs to be possible not
only between files, but also on the screen as well. Users
expect to be able to incorporate parts of spreadsheets inside
word documents, for example.
To reflect this broader sense of collaboration between tools,
we need some different language. In this paper I will use the
word integration for that aspect of tool collaboration that
focuses on GUI issues. The use of integration might make us
also think of the integrative nature of humanities scholarship
– an intentional parallel.
Integration in the GUI presents challenges. First, the
development of tools allowing for direct manipulation of objects
on the screen is more complex and costly than resolving
file-sharing issues. Furthermore, if independent tools are to
interact on the screen – elements maintained by potentially
different tools sharing screen space – then they must operate
in a computing framework that makes this kind of thing
practical.
Pliny (Pliny 2006) has been developed precisely to draw
attention to these two issues and to encourage some thinking
about cooperating tools beyond file-oriented “modularity”. It
both significantly broadens how computers can support
humanities research and suggests a much richer and more
acceptable interface to those tools that might well attract a larger
number of scholars. Pliny tackles this in two ways. First, it is
supports annotation and note taking – functions central to
several aspects of humanities scholarship and which I believe
have been largely neglected by the DH community – and,
second, it is built using a framework called Eclipse (Eclipse
2006) which is deliberatively created to support GUI-level
interaction between tools of the kind I mentioned above. Eclipse
also supports rich collaboration between tools developed by
independent developers.
Personal annotation and note-taking compels us to think about
tool integration because it is widespread in humanities
scholarship and runs across all kinds of scholarly work. Scholars
write notes to record their reactions to not only books they read,
but also web pages they view. Furthermore, if they have tools
that do (say) text analysis, they would probably want to record
notes about that as well. The provision of note-taking within a
particular website is insufficient for personal note-taking
(although it might well fulfil a useful need supporting public
comment about the material the website offers) because most
scholars work across a broad range of materials, and their note
taking capability must reflect this. Notes from a book will at
some point need to be brought in contact with notes about, say,
an online archive or a journal, or about findings from the text analysis tool. Personal scholarly note taking integrates by its
very nature.
Personal annotation also draws our attention back to the
software application as the context in which tools can and
should be built, rather than thinking of the browser as the
context for tool delivery. It is surprising how difficult it is to
make this point to those in the Digital Humanities. Of course
part of the reason is that XML, one of the key technologies that
fuels some portion of the DH community, has been developed
specifically to work in the context of the WWW. However, if
we wish to develop a more broadly based humanities
community who use technology in more sophisticated ways to
support their research we need to broaden our focus beyond
the WWW. Providing scholarly resources on the WWW
supports scholars, of course, and there remains some excitement
that they can get materials readily right to their desktop.
However, as I argued in Bradley 2005, once on that desktop
all they can do with them is read them on the screen or print
them out – a webpage, even one designed with all the clever
parts of CSS and AJAX, allows only manipulation within its
own page context. It contributes little to the integrative aspect
of personal scholarship which must, by its very nature, often
bring materials together from disparate sources.
Software Applications work with personal materials. Surely
this is a central element of humanities scholarship. A word
processor creates materials that belong to the researcher. Until
web browsers can be used to create materials that are stored as
personal data – on the user’s own machine – they cannot replace
applications (see a similar assertion from a technical perspective
in Charland 2005). Even the XML folk – part of the community
driving a WWW view of humanities computing because of
XML’s compatibility with website creation – use an application
such as Oxygen to create XML materials in the first place. For
the same reason that the creation of an XML file involves an
editor rather than a web browser, the creation of personal notes
– especially those gathered from reactions across a great range
of sources – requires a personal application. Furthermore, more
than one study into computing and humanities scholarship (see
both Brockman et al and Siemens et al 2004) has found that
scholars have tried to apply inappropriate applications such as
word processors to support this need, and have not been very
satisfied with the result.
Applications can be built in several different frameworks. Pliny
is written in Java using the Eclipse framework rather than the
one provided by Sun to create desktop tools. I didn’t choose
Eclipse because it was easy for me to use it – indeed I had to
learn it from the beginning during the past year or so. I choose
Eclipse because it is designed specifically to support the
integration of tools developed by separate developers. Eclipse
provides a way to share out screen space between windows
managed by different tools. Furthermore, it also provides
mechanisms for elements from separate tools to share the same
window (called “making a contribution” in Eclipse). Tools built
with the Eclipse plug-in model need not operate in isolation
from other related tools. A text analysis tool built in Eclipse
could use Pliny elements to allow users to record notes while
using the TA tool, and the notes that were recorded in this way
would also be visible in the context of notes created with other
materials – say from reading a book, or annotating a web page.
So, in this presentation I have an impossible task. First, I am
promoting the idea that we focus more on application
development than web development, a technically more
demanding activity, and one that goes against the grain of much
work in DH over the last decade. Second, for those who might
subscribe to this idea, I am promoting that we build our tools
according to the Eclipse model rather than the more widely
understood Sun/Java framework. Perhaps I won’t convince
anyone here. However, I believe that unless we start to think
of tools in the context of applications, as our scholarly
community does, and unless we start to think more seriously
of tool building in the context of GUI integration in addition
to data sharing, we will never get the attention of most scholars
in the humanities.
Bibliography
Bradley, John. "What You (Fore)see is What You Get: Thinking
about Usage Paradigms for Computer Assisted Text Analysis."
Text Techology 14.2 (2005): 1-19. Accessed 2006-09-01. <ht
tp://texttechnology.mcmaster.ca/pdf/vol14
_2/bradley14-2.pdf>
Bradley, John, and Geoffrey Rockwell. "Towards New Research
Tools in Computer-Assisted Text Analysis." Paper presented
at The Canadian Learned Societies Conference, June 1992.
1992. <http://www.cch.kcl.ac.uk/legacy/staf
f/jdb/papers/learneds.html>
Bradley, John, and Geoffrey Rockwell. "The Components of
a System for Computer Assisted Text Analysis." Paper
presented at the CETH Workshop on Future Text Analysis
Tools, October 1995. 1995. <http://www.cch.kcl.ac
.uk/legacy/staff/jdb/papers/ceth95.html>
Brockman, William S., Laura Neumann, Carole L. Palmer, and
Tonyia J. Tidline. Scholarly Work in the Humanities and the
Evolving Information Environment. Washington, D.C.: Digital
Libary Federation and Council on Library and Information
Resources, 2001.
Charland, Andre. "Will Ajax Replace the Desktop?"
developer.com. 2005. Accessed 2006-10-01. <http://www.developer.com/design/article.php/10925_3
574116_1>
Eclipse. 2006. Accessed 2006-10-01. <http://www.ecli
pse.org/>
Nora. 2006. Accessed 2006-10-01. <http://www.norap
roject.org/>
Ott, Wilhelm. "Strategies and Tools for Textual Scholarship:
the Tübingen System of Text Processing Programs (TUSTEP)."
Literary & Linguistic Computing 15.1 (2000): 93-108.
Pliny. 2006. Accessed 2006-10-01. <http://pliny.cch
.kcl.ac.uk/>
Siemens, Ray, Elaine Toms, Stéfan Sinclair, Geoffrey Rockwell,
and Lynne Siemens. "The Humanities Scholar in the
Twenty-first Century: How Research is Done and What Support
is Needed." ALLC/ACH 2004 Conference Abstracts. Göteborg:
Göteborg University, 2004.
TAPoR: Text Analysis Portal for Research. 2006. <http://
test-tapor.mcmaster.ca/portal/portal>