Industry 4.0 and the Internet of Things (IoT) have changed technical documentation. Users now read small chunks of information that match their context, task, or role. Instead of classical documentation, we now need information that is modular, neutral with respect to format, and accessible via metadata and full-text search.
We call this information: intelligent information.
Getting Started with Metadata: The Data About Data
Metadata is structured data that contains information about characteristics of other data. As a result, large amounts of data can be summarized in a clear and automated manner – for example, when describing videos, documents, objects, sitemaps, or metatags.
To develop metadata, we can use the PI classification, developed by Prof. Dr Wolfgang Ziegler*:
- Metadata from the P classification describes the relation of the documentation modules to assemblies, components, technical modules, and product configurations.
- Metadata from the I classification describes the information types (descriptive, instructive, etc.) and their relation to work activities, processes, target groups, and document types.
How to Derive metadata from Your requirements
The PI classification offers an excellent structure for developing a metadata model. But consider it a method rather than an out-of-the-box solution. To develop corporate and user-specific metadata, you need to check if the potential metadata matches with your actual requirements. The following describes a procedure for developing a metadata concept that is theoretically sound and practical.
1. Target Groups
First, take a look at the target groups for the intelligent information. The target groups tell you what types of information you need for your documentation. You could ask the following questions:
- Who works with the product?
- What are the tasks of the different target groups? Tip: Create a who-does-what matrix.
See also the article about the procedural model for creating technical documentation.
- In which work environments do the users work?
You also need to consider the existing knowledge of the target group members and what they expect from intelligent information:
- What type of content supports users best: video clips, training material, or technical data and a list of spare parts?
- Which media and devices are useful?
- Are there different access rights for the different target groups?
- Do you need to protect information from unauthorized access?
From this analysis, you derive the required information types (I classification). You also get first clues about an authorization concept for your documentation application and a media plan for publishing the information.
2. Customer Journey
Take your customers and users on a journey! The concept comes from marketing but can also be used to learn more about the different stations the users pass along the life cycle of a product. Where do customers and users stop, and which information do they need at which stop? How do they feel when they search for information at the stops?
Consider the following tracks and stops:
- Commissioning phase. Stops: Assemble the engine, configure software
Required information: Assembly instructions, reference of configuration parameters
- Operation phase. Stops: Retooling for specific scenarios, regular maintenance by operator. Required information: Retooling instructions, maintenance instructions for operators.
Your company’s requirements determine where you start the journey. You can begin with the product specification, or later, when the product has been manufactured. At the end of the journey, you will have use cases or user stories. Analyze these to find out which metadata you need for finding, filtering, and requesting intelligent information for each use case. You will also get a fourth class of metadata – the functional metadata. Functional metadata supports specific use cases for retrieving information, for example:
- An error code triggers the display of repair instructions.
- A QR code on a component opens a video with instructions on how to assemble the different parts of a machine.
3. Technical Implementation
To utilize intelligent information, we need applications, such as content delivery portals or service applications. They aggregate, evaluate, and display the information. The functions of these applications can also affect the required metadata. Consider the following questions:
- Which search filters should be available?
- Should the application respond to certain events?
- Do you need roles and access rights for using the application?
- How do users search?
- How do users navigate through large quantities of information?
- Where does the information come from (documentation, training, service, spare part lists, etc.)?
- In which context will the application be used?
Some of these requirements might not be supported by metadata but instead could be realized by the application that processes intelligent information. Other requirements need metadata, for example, information about error codes in instructions.
Standards Help to Deal with Complexity
Developing a metadata concept for intelligent information is a complex task. Are there other models besides the PI classification that we could use as a guideline? The good news is – yes.
tekom, the German association for technical communication, founded an intelligent information initiative, called IN³, and a working group that develops the delivery standard iiRDS. iiRDS stands for “intelligent information request and delivery standard”. It offers a standardized metadata model for intelligent information and a package format for the delivery of intelligent information to portals and applications. With iiRDS, we will be able to exchange documentation between different vendor systems and also make it available to other parties.
iiRDS cannot standardize all metadata. There are too many individual products and tasks along a product’s life cycle. But we can standardize metadata that describes technical documentation in general as well as the standard life cycle phases. And that’s what iiRDS intends to do. Apart from iiRDS, there are publicly available metadata models, such as Dublin Core or schemas on schema.org. These can also be useful for your metadata model.
* The PI classifikation was developed by Prof. Dr. Wolfgang Ziegler (University of Applied Sciences, Karlsruhe and I4ICM, Karlsruhe, Germany). As of May 2015, the method has been registered as a trademark: PI-Class®. Read more.
- The Core of Intelligent Information – Metadata - 3 July 2017