Machine-based semantic annotation of content in Confluence Enterprise Wiki

Annotation of content in any CMS, including traditional wikis, to provide contextual reference to any word or phrase is a common practice.  I am doing that in my blogs as I write.  When I reference specific topics or items, I go through the steps of creating links to them. 

Wiki annotation, therefore, also typically involves highlighting words or phrases and associating each to some reference - an external link or another page within the wiki.   It is a tedious process that involves marking up each word or phrase using specific wiki tags.  It not only requires considerable effort but also a meticulous nature and some dedication towards creating meaningful content.

What if it were automatic?  All you had to do is to add a macro at the beginning of your paragraph or page and everything meaningful within that text would automatically be annotated.

What if it were dynamic, so that the wiki content is still in its original form but all annotations are done when the page is rendered.  The annotation would automatically reflect any changes to the reference.  It would be more reliable and not prone to going stale.

What if the annotation was more contextual to your specific domain, to your business processes and functions, rather than highlighting of people, countries, organizations -- which may have less relevance when the text if referring to some technical requirements for a software component.

And finally, what if the annotation not only provided a link to some reference, but also some information about its type (is it a reference to a Project, a Requirement, a Task, a Bug, a Customer, a Software Component, ...)

Well, that's the capability that we will soon release for Atlassian's Confluence Enterprise Wiki in zAgile Teamwork and Wikidsmart products.

A simple, yet powerful capability in our upcoming release of Wikidsmart (Release 2.0) is the machine-based semantic annotation of wiki content.  It is easier to play with this feature than to describe it to really understand the possibilities that it opens up in making content more meaningful and accessible in enterprise wikis.  But here is an explanation:

By machine-based, I mean all you need to do to enable this capability is to surround the wiki paragraph with a macro (zannotate).

By semantic annotation, I mean that words or phrases in that paragraph that match any reference in our ontology-based semantic repository will be highlighted and linkable to additional information about the references.  Since zAgile Teamwork ships with ontologies for software engineering, you will see annotations of words and phrases that reference instances of Requirements, Projects (JIRA Projects), JIRA Issues, Tasks, Bugs, Software Components, Test Cases, etc.

You can test drive this capability in the zTeamwork sandbox.

Here is what it would look like:

WRITE A PARAGRAPH IN A WIKI AND BRACKET IT WITH ANNOTATION MACRO

SAVE THE PAGE AND THE PARAGRAPH WILL BE AUTOMATICALLY ANNOTATED.  CLICKING ON ANY TAGGED PHRASE WILL POPUP ITS REFERENCE CONTEXT

FOR EXAMPLE, REFERENCES TO JIRA ISSUES NOW ARE AUTOMATICALLY ANNOTATED WITH TYPE, DESCRIPTIONS AND LINKS TO THE INDIVIDUAL ISSUES

AND FINALLY, THE SAME APPLIED TO MEDICAL TEXT - APPLIED AGAINST AN  ONTOLOGY DERIVED FROM MeSH (Medical Subject Headings) THESAURUS, ALSO IN CONFLUENCE - SHOWING HIGHLIGHTED TERM AND ITS SYNONYMS

- Sanjiva

To test drive this feature, visit our sandbox

Comments

[Dan]

Not only does this address a problem that I've been noodling for a while, but the idea of using a tag to denote what receives the semantic attention is so simple and elegant that is deserves attention and flattery. Congrats!