Although metadata has been around forever, IT began to identify and recognize it not very long ago. Progress has been paramount. Outside of IT, however, heavily focused metadata interests, such as library science, have used metadata as their backbone for a very long time. This article is a means to evaluate and perhaps redirect IT metadata paths.
The Metadata Timeline
It is not easy to identify metadata's birth in IT. After something is born and as life continues, various milestones occur, with variations dependent upon how the life is lived and what life-impacting events are experienced along the way. Metadata is no exception. Its major milestones have been viewed in IT in terms of its application. Those milestones are illustrated in Figure 1.
Figure 1: The Metadata Lifecycle
- Identification of the need. The inability to locate information is often a driver for organizations. If we can't locate something we are looking for, we have a problem. Once we identify the need, we can deliver a simple metadata solution.
- Analysis of the problems. If the information world itself is not functioning as designed, individual research into key causes for failure begins. The evaluation of these causes will generally involve process analysis, emphasizing the creation and storage of key information. In many situations, the metadata timeline stops here. A solid solution will almost always require some business process re-engineering as well as heavy cost/benefit analysis. Unfortunately, the performed cost/benefit analysis is often restrictive and siloed. A more liberal and open view realizes that the originally identified need is in and of itself often too narrow. Information access and usage issues span the entire organization.
- Development of a plan. At this step in the metadata timeline, we see diverse strategies, focuses and execution plans. It may be too early to tell if success is imminent. Based upon achieved successes, some of us can move to the next metadata solution level, the advanced solution.
Advanced Metadata Solution Role Models
Contributing to metadata's rebirth requires a movement to the next level - that of the "meta-metamodel." In this scenario, metadata becomes the same as the information many used as motivation to begin the metadata lifeline.
Meta-metamodels are often considered abstract, but they are clearly the backbone of advanced metadata solutions in that they treat all metadata, metadata storage and metadata usage with the same level and amount of software-based processing. Those industries and information-focused specialties with advanced treatment of metadata have recognized this requirement since their initial dependence upon metadata. These industries have always been admired for their immediate ability to organize in a repeatable, standard fashion.
One of the disciplines most dependent upon metadata, library science, quickly advanced to use all Internet benefits. By standardizing a means to search and retrieve, the discovery and selection of the acquired information becomes predictable and accurate. Items for recovery are classified with recognized and common qualifications - tags, keywords and a solid and standard digital library resulted. The library's existence and commonality is based upon the "meta meta" consistency. Figure 2 shows the digital library's underpinnings, otherwise known as "meta meta" depiction. Notice how everything is a "thing" and how every "thing" falls into the categories that are necessary for the library ontology model. These "thing" types can vary across our numerous situations.
Figure 2: A Digital Library's Ontology
The geographic industry, with potential for immense metadata conflict, has localized its metadata, with a connection to an "enterprise standard" via regulated and fixed identifiers. Consider maps, for example, and the fact that maps of individual localities, such as towns and counties, may contain detailed block and lot numbers. This lower level of identification is not referenced directly by larger scoped maps such as country maps. However, to get from a town map to a county map to a state map to a country map requires consistent connectors. Consider ZIP codes, state names, telephone exchanges, latitudes and longitudes, country names and spatial identifiers. Figure 3 shows the meta-meta consistency that keeps all of these instances connected.
Figure 3: The ISO Standardization View
In this industry, the metadata begins with the clarification of notes - a music scale names the notes. Notes are all part of a measure; within the measure, the rhythm is determined via notes (half notes, quarter notes, etc.). When all measures are combined into songs, they are named and associated with tracks, which are associated with a music medium. Because of this meta-meta view, everyone knows how to locate and understand music.
Now that digital recordings are common, music has moved to the world of metadata. Figure 4 shows groups of notes required to represent vocals or any other combination of notes with or without a singing voice and lyrics. Figure 5 combines the notes into projects. These metamodels standardize and clarify all components of music.
Figure 4: A Meta-Meta View of a Vocal Recording
Figure 5: Combining Notes into Projects
At this point, it should be clear that IT cannot be considered as a metadata or metamodel-based practice. IT objects, our building blocks and completed projects are not standard in any way. There are no required components, no required component types and no universal way of putting the pieces together. Now it is time for metadata's new clothes.
Many attempts at creating the consistent IT building blocks and plans have come and gone and were then revisited. Each was successful within its own window, but when the windows were joined or combined, the success evaporated. The biggest issue within IT is still the inability to put things together. This issue has been referred to technically as integration.
Metadata was never considered a piece of the potential solution to this problem. Traditionally, it was viewed as an after-the-fact expansion and explanation of the components as well as the resulting project. Metadata at its best covers all aspects of the objects being described, regardless of their place in the project development lifecycle. The metamodel combines the metadata to some degree by connecting and attempting to standardize those descriptive details that are tracked.
A simple metamodel is illustrated in Figure 6. This example of a data warehouse and its main physical components helps us look at all data warehouses from a similar perspective when we are evaluating the database structure. But when we want to delve outside of the database structural perspective - perhaps into the specifics of how the database is populated (where does it come from, what happens to it along the way, what programs and/or software products are used, etc.), there are so many options as to what is tracked and how it is named and related that we must admit that our metamodeling hopes stop here. We only keep the metadata that we want within the realm of what we want it for. Today's awareness of metamodeling limitations keeps IT professionals looking for an interproject, interscope, interbuilding-block metamodel. The solution can only be a meta-metamodel.
Figure 6: A Simple Metamodel
Meta-metamodels are used to connect metamodels by keeping track of each metamodel and its given components. They are rarely considered practical; they are considered theoretical. They are here today, but we just don't see them. It is time now to focus our metadata solutions on the meta-metamodel.
Just as we use data models to think of our data irrespective of its implementation, we use meta-metamodels to represent the many renditions and instances of the metamodel irrespective of their implementations. Metadata's new clothes will reveal metadata's point and position within the metadata enterprise. Figure 7 shows the ultimate objective - an enterprise meta-metamodel. It is very easy to draw, but a bit harder to implement.
Figure 7: The Enterprise Meta-Metamodel
As with the successful metadata-focused industries (libraries, geography, music), IT's focus on metadata solutions begins at the object or building-block level. With the wide variety of implementation options that exist today within IT, the difficulty of standardizing the components is obvious. Object orientation, master data management, service-oriented architectures and enterprise data models are all focused on the same objective. Again, each methodology's objective is within a clearly delimited perspective.
The New Solution
Integration and standardization within IT begins with a new mode of delivery. This new delivery focus, whether it is for information, documentation, software or processing, must consider the delivery of generic building blocks that can be reused again and again. Standardized building blocks require external and internal collaboration; perhaps IT should follow the roadmaps of model industries. Success will only result when objects are defined, categorized and named as people use them and these definitions become shared and required across the industry.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access