November - December 2004
Education and Career Opportunities

Metadata and Project Management

Rainer Volz

Metadata and project management? There are connections, although they might not be obvious at first. Like any other modern business practice project management deals a lot with the management and processing of information. Technology made it possible to create and store a lot of information that would have been transitory previously. And we use this possibility a lot, for various reasons. These include the need to be able to show how we came to decisions, where all the money went, and that we complied with regulations.

But the amount of information stored makes it also more difficult to find relevant documents, when we need them. Another issue is that information ages quickly. Not only in the sense that it becomes obsolete because of new developments, but due to the fact that it quickly loses its connections to its context. Project documentation is one example for this. Typically only a few documents of a project are still usable for outsiders a while after the project has been finished. Not that there weren't enough documents available. But it is often not easy to find the relevant ones, relevant to our current interests. Because the "soft data", the meaning behind the documents is lost over time.

Project-centric organisations, like consulting firms, know this and invest in intranet knowledge management systems, to store all kinds of project related information that could be reused later. The better of these KM systems don't store only the artifacts but also relate them to some context, i.e. to some process or organisational principle, client information, metrics, thereby adding meta-information, metadata. Context information that makes it easier to find and apply the resources again.

These systems exist in corporate intranets, with a uniform data format, and central management. In an Internet environment the situation is slightly different. Information sources, documents appear in different formats, versions, and languages. To make sense of the data available there, to be able to find the documents and services we are searching for, we need a lot of metadata, preferably in a machine-readable way. Fortunately, some people are already working on standards and applications to help us there.

An interesting example for the application of metadata are PM-oriented web sites, like the Project Management World Today. The PM World Today site contains a lot of information, much of it useful beyond the lifetime of the edition it was published in. Until now there were only two ways to find something again, if I didn't store the link. I can use some of the category overviews and indexes, hand-made by the Editor of PM World Today David Curling, or use the search engine provided.

Manually created overviews and indexes are a lot of work for the publisher, and are necessarily biased. With search-engines the user has all the work, finding the best keyword combination matching his/her interests and manually browsing through the results. Adding metadata to published documents could help to reduce the efforts of both the publisher and the reader, and could improve the search experience. David Curling explained his metadata approach for the Project Management World Today in his recent article The Semantic PM Web.

How does it work? Simply by making the metadata, the meaningful extra data related to a document machine-readable. This is done with RDF. XML (Extended Markup Language) brings data into a standardised format, so that different applications can process the data, irrespective of their own, proprietary data storage format. RDF does the same to data describing other data, metadata.

RDF allows to create statements about things in the form ";subject";";predicate"; ";object";. In the article mentioned above David Curling gave an example of the metadata he uses (the subject is the web page itself, its URL; predicates = name; object = content) :

< meta name="DC.title" content="State of the Art of Project Management ..." />
< meta name="DC.creator" content="Russell D. Archibald" />
< meta name="DC.subject" content="project management, project schedule, ..." />
< meta name="DC.description" content="The practice of project management ..." />
< meta name="DC.publisher" content="PM World Today January-February 2004 ISSN 1492-5354 " />
< meta name="DC.contributor" content="Max Wideman, David Curing,Allan Harpham" />
< meta name="DC.date" scheme="DCTERMS.W3CDTF" content="2004-01-01" />
< meta name="DC.type" scheme="DCTERMS.DCMIType" content="Text" />
< meta name="DC.format" content="text/html 39259 bytes" />
< meta name="DC.language" content="english" />

In this example David used the possibility to embed a metadata vocabulary, Dublin Core, directly into the HTML page. It would also have been possible to put this metadata into a separate RDF file and link to it. But the main point here is that the information shown above produces meaning for a machine. While we human readers could have gathered these details from the text, search-engines need the extra formatting. So there is a little extra effort in creating the metadata in a certain format, but as a result the metadata allows readers do qualified searching with bots. Instead of searching merely for some keywords in the text I could now execute a database-like search: give me all texts (DC.type), by Russell Archibald (DC.creator), published in January 2004 (DC.date).

I didn't mention databases accidentally. Creating RDF metadata for resources results quasi automatically in a database. It might not look like the typical SQL database we are used to, but it is one. There are even SQL-like languages to query them. With the metadata example above David
created a database record, describing a resource, in this case the web page with the Russell Archibald article. All metadata entries of a PMWT edition collected would result in database describing the content of this edition, which could then be used to search the edition, or for further processing e.g., to create category overviews and indexes.

The search for names and dates is easy, what about the rest? Aren't we back to keyword hunting with fields like DC.subject? Not necessarily. As already mentioned, there are vocabularies. The one used here, Dublin Core, defines DC.subject as:

"Typically, a Subject will be expressed as keywords, key phrases or classification codes that describe a topic of the resource. Recommended best practice is to select a value from a controlled vocabulary or formal classification scheme."

So there should be no arbitrary content in a DC.Subject statement. We shouldn't need to brainstorm for possible keywords -- if we know the vocabulary used, we can simply look them up, and decide in which we are really interested. Much faster than brainstorming. David decided in his article to use Max Wideman's PM Glossary as a controlled vocabulary for the Project Management World Today.

With these two aspects taken together metadata technologies like RDF provide us with a way to describe arbitrary resource (e.g., HTML pages, images, word processor documents, spreadsheets, web services, programs) in a machine-processable structure, and with defined content vocabularies. This allows us to provide context to these resources, so that it will be easier to understand the relevance of one resource in relation to others.

As shown previously the collection of all metadata entries of a web site provide us with a database-like view on it. But we don't need to stop there, with only one web site. If more PM web sites would follow the example of David Curling, users interested in project management (or better: their search agents) could collect their metadata and search qualified across web sites or organisational boundaries. Even if the individual web sites used different controlled vocabularies for their descriptions. That would make it easier to find information relevant for PM professionals on the Internet.

Of course, the same mechanisms can be used to provide access to information, services and other resources across organisational boundaries. But that is another story....

copyright® 2003 Rainer Volz

Follow Rainer Volz's Web-based PM Groupware discussion through his previous PM World Today Viewpoints Columns.

About the Author:

Rainer VolzRainer Volz has been an IT architect and manager for over 10 years, mainly in the airline and financial services industries. He specializes in technology and management consulting for virtual organizations, and edits the "Virtual Projects" website (www.vrtprj.com / www.vrtprj.de) to provide more information about collaboration and distributed project management.

Global research project orientation [international]

The PROJEKTMANAGEMENT GROUP of the Vienna University of Economics and Business Administration has launched a global research: project orientation [international]. The objectives of this research are:

The research is a further development of the project “POS BM” (Project-oriented Society Benchmarking), which we conducted between 2000-2002 in cooperation with IPMA. Cooperation partners were: Denmark, Hungary, Ireland, Latvia, Norway, Romania, Sweden and Austria. In 2003 Moscow as a project-oriented region has been analyzed.

Results of POS BM can be found at www.wu-wien.ac.at/pmg/pos

Now we have further developed the project orientation maturity model and will choose a different approach. In project orientation [international] we concentrate on the analysis of the project-oriented organizations and project-oriented industries. Are you interested in how mature your project-oriented organization /industry /region or nation is? Then join this global research project !

Further information on the research project and the description of the project orientation maturity model see: www.wu-wien.ac.at/pmg/poi or contact the project manager

Top of Page