Web Taxonomy Creation

An important contribution to Public Understanding ofamong scientists. This builds on research we have
Science and for enabling of new insights would beundertaken to enable end-user programming. We will
creation of an online systematic representation ofvisualise scientific information and make this editable
scientific information that gives a holistic view ofonline. The system created must be judged
related knowledge. This would reuse information andaccessible and user friendly by users. In addition, it
ideas provided by other researchers. These systemsmust share information successfully with other
organise and visualise information. Letondal [1] of thesoftware and people. So interoperability is extremely
Pasteur Institute provides a tool to generate webimportant, examples of interoperability research we
interfaces for molecular biology programs using anwill make use of are INTEROP [21] and MOMOCS
end-user programming approach. TreeBASE [2] is a[22], and combine this with a model driven approach.
relational database of phylogenetic information hostedIn order to achieve interoperability, the utilisation of
by the Yale Peabody museum. Hughes [3] of theopen formats should be favoured in order to
Computational Biology Unit (CBU) University ofmaximise chances of forward compatibility with
Bergen has investigated visualisation of phylogeneticevolving scientific models. We will also create or
trees and written an application that converts thecustomise editing tools such as wikis blogs, and
NCBI (National Center for Biotechnology Information)Semantic Web editors [23] to document the models
[4] taxonomic data into a format that makes itand collect user's feedback. We will visualise scientific
possible to visualise the whole of the NCBI tree ofinformation and make this available online.
life. Klyne [5] of the Image Bioinformatics ResearchReferences
Group, Zoology Department, Oxford University[1] Letondal C. - - Pasteur Institute, Computing
makes use of Semantic Web technology to build anCenter.
image publication repository. UK Moths [6] online guide[2] TreeBASE - a Database of Phylogenetic
to the moths of Great Britain and Ireland illustratesKnowledge.
the role of amateurs who send in pictures to be[3] Hughes T. Phylogenetic tree visualisation - Tim's
categorised. Mammals Family tree [7] provides anHome Page - [
innovative way of visualising complex trees as a[4] Wheeler DL, Chappey C, Lash AE, Leipe DD,
taxonomy on a wheel, this is a technique which couldMadden TL, Schuler GD, Tatusova TA, Rapp BA
be developed as an interactive visualization. Fishbase(2000). Database resources of the National Center
[8] is a web accessible and searchable resource forfor Biotechnology Information. Nucleic Acids Res 2000
accessing fish information by various categories. IHOPJan 1;28(1):10-4.
(Information hyperlinked over Proteins) by Hoffman[5] Klyne G, 2007, Building a Semantic Web accessible
[9] provides a gene network for navigating moreimage publication repository - Image Bioinformatics
than ten million abstracts in PubMed [4]. By usingResearch Group - Zoology Department - Oxford
genes and proteins as hyperlinks between sentencesUniversity.
and abstracts, this converts the information in[6] UKMoths - Welcome to UKMoths, your online
PubMed into one navigable resource. Proviision of 3Dguide to the moths of Great Britain and Ireland.
hyperbolic representations of trees is investigates by[7] Mammals family tree - Use zoom tool to see
Hughes et al [10]. This capability could be provided onmammal details on the circle's edge - - BBC News
the web using Semantic Web Technologies [11] [12]Science - Mammal rise 'not linked' to dinos - PDF
and this would enable others to edit and interact withFamily tree visualisation.
these represenatations.[8] Fishbase - Taxonomy of Fish Species.
To achieve the above a collaborative modelling[9] Hoffman R, Information hyperlinked over Proteins
approach is required. The literature on collaborative- A gene network for navigating the literature.
modelling is extensive. Huhns [13] and Paternò[10] Hughes T, Hyun Y and Liberles, 2004. Visualising
[14] both explain that alternatives to currentvery large phylogenetic trees in three dimensional
software development approaches are required forhyperbolic space. DA. BMC Bioinformatics. BMC
ease of model creation. The need is to translateBioinformatics Apr 29;5:48.
from a model-based visual representation understood[11] Eklund P, Roberts N, Green S, 2002. OntoRama:
by users to software. This makes it possible toBrowsing RDF Ontologies using a Hyperbolic-style
engage with end-users and non-specialists in general.Browser, The First International Symposium on Cyber
Johnson [15] explains that successful interactionWorlds, CW02, Theory and Practices, IEEE Press.
requires mapping between levels of abstractions and(2002) pp 405-411.
that translation between these abstraction levels[12] Talis Engage Community information at your
required by people and computers is difficult. Johnsonfingertips - an online community information solution
explains that this problem often means systems arethat allows citizens to organise, create and publish
created that make users cope with the problems ofdetails of their events, organisations and groups.
mis-translation. The representation of rules and[13] Huhns M, 2001, Interaction-Oriented Software
information can be illustrated diagrammatically and it isDevelopment, Journal of Software Engineering and
possible to describe algorithms through concreteKnowledge Engineering.
examples rather than abstractly. Models must be[14] Paternò F, 2005, Model-based tools for
designed and visualised so that they convey to userspervasive usability, Interacting with Computers Vol
a representation of a problem that assists with their17(3), pp 291-315.
vision of it. This modelling approach is explored by[15] Johnson P, 2004, Interactions, Collaborations and
Crapo et al [16] and is the basis of our visualisationbreakdowns, ACM International Conference
techniques allowing the creation and understanding ofProceeding Series, 3rd annual conference on Task
taxonomies and models.models and diagrams Vol 86.
Scaffidi et al [17] show that most people who[16] Crapo A W, Waisel L B, Wallace W A, Willemain
develop software are de-facto programmers lackingT R, 2002, Visualization and Modelling for Intelligent
a formal computing background, this will often be theSystems, Intelligent Systems: Technology and
case for scientists and the public contributing toApplications, 1, pp 53-85.
taxonomies. End-user programming is particularly[17] Scaffidi C, Shaw M, Myers B, 2005. Estimating
important in this research as we are making softwarethe Numbers of End Users and End User
development accessible to non-experts. Research byProgrammers. IEEE Symposium on Visual Languages
Ko [18] explains the need for engagement ofand Human-Centric Computing, 21-24 September,
end-users, including non-specialists by providing themDallas, USA.
with the capability to interact and amend software.[18] Ko A J, 2007. Barriers to Successful End-User
An e-science interactive environment is ideal forProgramming. End-User Software Engineering Dagstuhl
involving anyone interested in science to amend orSeminar.
produce personal content. The environment will[19] Stutt A, Motta E, 2004. Semantic Learning Webs.
benefit from an interactive e-learning approachJournal of Interactive Media in Education, 2004(10),
influenced by 'Semantic Learning Webs' [19]. TheSpecial Issue on the Educational Semantic Web.
capabilities we will provide over the web are similar,[20] Fischer G, 2007. Meta-Design: A Conceptual
but more collaborative and advanced than providedFramework for End-User Software Engineering.
by spreadsheets for modelling, and web editors forEnd-User Software Engineering Dagstuhl Seminar.
knowledge sharing. We will produce an alternative[21] INTEROP
methodology for scientific modelling that hides the[22] MOMOCS - MOdel driven MOdernisation of
complexity of low-level programming code fromComplex Systems.
users. This is a kind of meta-design, as explained by[23] Quint V, Vatton I, 2005. Towards Active Web
Fischer [20] that can be standardised to create aClients. DocEng Bristol United.
collaborative environment for sharing information