Introduction

It is widely recognized that modern document management systems must facilitate effective focused searches and that one of the most effective ways of doing so is by the systematic application of metadata, especially keywords, to documents collected in central document repositories. This document will serve as a place to collect policy and discussion regarding the evolving keyword set for the University of Toronto Department of Physics.

Some contextual disambiguation of keywords can be very useful and is easily introduced via a single-level prefix heirarchy. For example, it may be valuable to distinguish references to the Quantum Optics group in the Physics Department from references to Quantum Optics as a research field/activity. In principle, a paper by one group might have implications in an adjacent field and we would wish to highlight this fact by adding, for example, High Energy Physics, without implying a relationship with the local HEP group within the department. In this case we would use the prefix Org, as in Org: Quantum Optics and also add a keyword of High Energy Physics. (Note that this process will require some...)

Note that not all searching and identification challenges should be addressed with keywords. Excessive proliferation of keywords is a distinct worry, both from management and use perspectives. The establishment of local naming conventions will be an invaluable organization adjunct. The official departmental colloquium may or may not deserve a distinct keyword, but every new seminar series probably does not. Choose a good title for the series and ensure that every announcement instance contains it.

Notes

From where can we steal some pre-existing taxonomies and keyword sets?

Keyword Sets

Although the space of keywords is, by design, not heirarchical, it can be usefu l in the course of keyword development to do some coarse-grained classification to provide direction and focus to the keyword development.

Does Plone support infrastructure for keyword heirarchies? The idea is that the keyword space is heirarchical for presentation and organization of keywords, but the keyword space is flat with respect to the search mechanisms. In this case, the heirarchies are a useful tool for choosing effective keyword searches.

Departmental Organization Keywords (Org)

This group of keywords should be restricted to formally recognized parts of the organization. Individual groups will want to identify themselves, but that should probably be done by a combination of other keywords related to areas of activity. Authorship will also provide a partial means of identification. This could be a contentious issue, because there is no natural boundary that eliminates the absurd limit of one keyword per individual as each group identifies and sub-identifies its boundaries. (How can we use the group membership features of the system to enhance searching? It may be necessary to extend the extant Smart Folders.)

• Org: High Energy Physics
• Org: Quantum Optics and Condensed Matter Physics
• Org: Planetary Physics
• Org: Physics Teaching
• Org: Departmental Administration
  

Research Related Keywords

Issues around selection for this class probably should be resolved by adopting standards of the various research community journals, such as the scitation indexes. Imperfect standards are almost always better than attempts at imperfect alternative implementations, since this necessitates learning of a new system by the user.

• Experimental
• Theoretical
• ??
  

Departmental Activity Keywords

• seminar
• colloquium
• meeting
• 
  

Financial Management Keywords

Examples from this keyword set would be: invoice, purchase order, sale, asset retirement.

Academic Research Keywords

Information Technology Keywords

These are keywords related to the management and tracking of information technology resources such as computers, networks, security, etc.

Operating Systems

• Microsoft
• Microsoft Windows
• Windows XP

• Unix
• Linux
• SunOS
• HP-UX
• AIX

Linux Distributions

• Red Hat
• Debian
• Novell/Suse
• Gentoo
• Mandriva

Computer/OS Vendors

• Sun Microsystem
• Microsoft
• Red Hat
• Novell

General Computing Hardware/Software Vendors

• Compugen
• OpenStorage Solutions
• IBM