On Topic

In the latest version of the Networked Planet TMCore EPiServer Module (henceforth TEM); we moved away from the model of using EPiServer page IDs to instead use a dynamically generated GUID.

EPiServer Module 1.x

In TEM 1.x each topic that represented in a page in EPiServer would have a subject identifier which was composed of a prefix (typically http://www.networkedplanet.com/2005/01/episerver/np-gen-psi/ followed by the integer ID of the page that this topic represented. The advantages of this system were:

• The ID was persistent in EPiServer, so there were no concerns with the ID changing within an EPiServer installation.
• The subject identifier of the topic could be easily derived from the page (just add the prefix), and vice versa.

Unfortunately this system also carried a penalty. If page data was exported from EPiServer and then migrated to a new EPiServer installation, a new ID would be assigned, usually different from the old id. This means that the link between the page and the topic was broken. In short, moving page data around between servers was very difficult and we felt we had to address this in the 2.0 release of the TEM.

EPiServer Module 2.0

In TEM 2.0 we require the definition a new dynamic property to EPiServer which is a GUID, or Globally Unique IDentifier. The GUID is then used, in combination with a prefix urn:guid:, to construct the subject identifier of the topic representing the page.

The result of this is that the GUID property of the page is unchanged throughout the life of the page, and is preserved through export operations. This therefore allows EPiServer administrators to export page data, export the XTM from TMCore and then import in to new systems and the system will work as expected.

The EPiServer page ID can still be derived from the a topic via an occurrence value on the topic, typed by the Page ID occurrence type topic, by default this a topic with the subject identifier http://www.networkedplanet.com/2005/01/episerver/pageid.

Pitfalls

Of course, nothing ever comes for free, there are few ill-considered things you can do to EPiServer which will cause problems with this new system.

The most serious of these is to delete the dynamic property from EPiServer. This causes all GUID values of pages to be destroyed. These values are unfortunately unrecoverable.

Manually setting the GUID is also unwise, not only do you risk setting the GUID to a value which will potentially clash with another page ID, but also you will put the page GUID out of sync with the topic GUID. This means that you must either:

1. Manually set the GUID subject identifier on the topic to match the changed page GUID; or
2. Run the TEM consistency checker which will create a new topic for the page, based on the new GUID. However, this will leave the old topic hanging around in the system, along with all its associations which may cause confusion.

A number of projects are using a Release Candidate or Beta release of the EPiServer Module 2.0. With the final release now public here are a few additional pieces of information:

Which files are added/deleted?

There have been additions to the documentation and the AJAX topic editing client. However, deploying over the top of an existing installation will work correctly; the obsolete files will simply not be used.

Is it necessary to run the EpiServerModule.sql when I am updating?

Yes, we've updated the SQL Stored Procedures. However, this does not change any of the data in the database.

Which web.config settings should be added/removed?

No settings must be removed or changed, however you should review your existing web.config configuration with the installation guide to ensure correct operation.

Adding a new Dynamic Property 'topicidentity' to the EPiServer Project. Should there be a default value?

No value is required as default. It will get filled in automatically on each page by the EPiServer Module, either during the page create post-processing, or by the consistency checker.

The first time I run the Consistency Checker I get a lot problems reported

page name has no GUID. Will create one: urn:guid:xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
page name has unwanted subject identifier http://www.networkedplanet.com/2005/01/episerver/np-gen-psi/pageid
page name requires new subject identifier: urn:guid:xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx

In TEM 2.0 we require the definition a new dynamic property to EPiServer which is a GUID, or Globally Unique IDentifier. The GUID is then used, in combination with a prefix urn:guid:, to construct the subject identifier of the topic representing the page.

The result of this is that the GUID property of the page is unchanged throughout the life of the page, and is preserved through export operations. This therefore allows EPiServer administrators to export page data, export the XTM from TMCore and then import in to new systems and the system will work as expected.

The EPiServer page ID can still be derived from the a topic via an occurrence value on the topic, typed by the Page ID occurrence type topic, by default this a topic with the subject identifier http://www.networkedplanet.com/2005/01/episerver/pageid.

The old subject identifier is no longer required and will be removed by the consistency checker.

We have noticed that when downloading the TMCore EPiServer Module it is sometimes the case that the help file, .chm file does not open correctly. This article explains why, and how to fix it.

The solution

Rather than tantalize and tease with the solution, to make the chm file readable on your machine open up the properties for the CHM file and click the Unblock button as shown here below:

The problem

The "problem" is caused by Windows protecting your system. Windows is aware that the CHM file originated from the internet. The help system, which is HTML based, contains code (ActiveX) which, if maliciously written, could damage your machine. Therefore, Microsoft issues are security patch (896358/890175) to disable this by default.

In order to allow the code within the CHM file to run, you must "unblock" the file.

References

The links below explain the exact nature of the problem and alternative workarounds to this problem.

• Microsoft Knowledge Base Article 902225
• Microsoft Knowledge Base Article 892675

Us Topic Mappers know the power of the paradigm - we can envision how the generic data model of topics maps combined with its robust treatment of identity can be used as the basis for powerful information integration solutions. However, thus far most people have seen topic maps exposed as part of a web content management system or as part of a stand alone portal.

This however, was all part of a very cunning plan. For once people see the value topic maps can bring to web site navigation and general content findability they begin to realise and think about how this paradigm, these benefits, can be applied across more information systems. They realise that Topic Maps is not just about related content links, but it is about harmonising and contextualising the access to information. Information that is distributed and heterogenous in nature. Topic Maps with one CMS application is the thin end of the wedge, the value and benefits of using Topic Maps across an organisation, across applications increase exponentially.

There is an obvious value proposition for topic maps, it is the glue that binds information together. The value of a single piece of information is only as useful as the means for people to find and utilise it. Thus the value of a CMS or a Web CMS is fundamentally limited unless the mechanism by which users can find content is effective. Topic Maps, this semantic binding, adds meaning to information, provides context to access and enables a semantic layer that people can interact with before delving into the depths of one of many information repositories.

Let's consider an example of how Topic Maps is the glue that binds information together. TMCore already integrates with SharePoint and EPiServer. One common problem is that people want to collaborate and create content in SharePoint but then hook it into the intranet or internet site that is published via EPiServer. In addition, they want to provide a unified Web Service for accessing all content to enable other systems to integrate.

TMCore and Topic Maps provide a topic based information glue that can hold this web of information together in a way that allows it to be exploited in many different applications and forms.

The basic metaphor is that there are two kinds of topics, there are concept topics, people, places, departments, skills, projects etc and there are content topics. Content topics are topics that are proxies for a document or piece of content in some content repository.

When content is created in any system that is hooked into the topic map technology a new topic is created. This topic is uniquely identified and can unambiguously locate, fetch or more commonly, address the piece of content that it represents.

When content is created or modified users or auto-classificaiton features can connect the topic that represents the content item to topics that are in the concept space (the people, projects, skills etc).

Now, in either EPiServer or SharePoint it is possible to query the topic map to find topics and content that are located nearby in the semantic space. The key thing here is that the amount of classification to achieve a high degree of interconnected topics is very low (4 or 5 connections between a content topic and concept topic is enough). All of this information can be rendered as clickable links that either display a page for another topic or can go directly to a content item stored in a content repository.

Furthermore, as all the binding information is in the topic map, it can be used to support the implementation of web services that provide unified access to all information is a given set of systems.

These are just a few ideas on how Topic Maps and TMCore can be used as way to deliver seamless enterprise information.

The TMCore SharePoint Module is an integration between Microsoft SharePoint 2007 and the TMCore Information Server that allows SharePoint users to quickly and easily organise content in a SharePoint site using a topic map ontology defined and managed separately by a librarian or systems administrator.

Features of this first beta of the integration include:

• Classify any content against the topic map - any SharePoint site, list or list item can be classified.
• Ontology-driven classification - the classification UI uses the topic map ontology to make it easy for users to connect items to other topics and to provide only occurrence data that is relevant to the item.
• Map SharePoint Content Types to the ontology - this makes the classification of SharePoint items almost completely automated. The administrator sets up a content type to topic type mapping that defines how content items should be classified.
• Browse and edit associations as SharePoint meta-data - a custom field type allows administrators to add fields to SharePoint lists that reflect association infomation from the topic map. Users see these fields as hyperlinks taking them to the related item and can edit them on the normal SharePoint meta-data editing screens.
• Suite of WebParts - a selection of web parts allow users to search or browse the topic map as well as to display or edit the details of topics.

The beta is now available for download. Please note that this software is in beta form and should be considered UNSTABLE. To perform the installation you will need to have first installed SharePoint 2007 Server and TMCore07 SP1. If you have a license for TMCore, please contact us for download details for TMCore07 SP1. If you do not have a license, you can request an evaluation copy of TMCore.

Download the TMCore SharePoint Module 2.0 Beta 1.

The TMCore Research Service is an implementation of the Microsoft Office Research web service interface that allows Office (2003 and 2007) and Internet Explorer (6 or 7) users search and browse a topic map using the Research panel.

This initial beta is packaged as an ASP.NET 2.0 web application which you can install on any IIS server that can access a TMCore database through the Topic Map Web Service (TMWS). The package contains full installation instructions and a developers guide describing ways in which you can configure the output of the service.

This beta version has been built and tested against TMCore07 SP1. If you have a license for TMCore, you can contact us for download details. If you do not currently have a license for TMCore but want to try out this beta, please request an evaluation copy of TMCore.

Download the TMCore Research Service BETA.