Mimris

Documentation · History and concepts

From Metis to Mimris

A pioneering meta-modeling platform and the ideas that continue in an AI-native workspace.

The Metis modeling tool was one of the pioneering enterprise modeling environments developed in Norway during the 1990s. Unlike traditional CASE tools or UML-only modeling tools, Metis was designed as a meta-modeling platform: a system for defining modeling languages, methods, and analysis capabilities rather than being restricted to a single notation.

Product history

The product evolved through several owners:

PeriodCompanyRole
Early 1990sMetis ASOriginal development
Mid 1990sDigital Equipment Corporation (DEC)Acquisition and further development
LaterAT&T / NCREnterprise product
2001Computas ASAcquired the Metis product and development team
2005Troux TechnologiesAcquired Computas Technology and Metis
LaterTroux, acquired by PlanviewEnterprise Architecture portfolio evolution

The philosophy behind Metis

Everything is a model

Even the modeling language itself is a model.

Instead of hardcoding UML or BPMN, Metis separated the meta-model, model, and instances. This meant users could create entirely new modeling languages.

Meta-model↓Business process language↓Business process model↓Actual business process
Meta-model↓Electrical engineering language↓Circuit model↓Physical installation

This flexibility made Metis very different from tools such as Rational Rose or Visio.

Core concepts

A repository of reusable meaning

1. Meta-modeling

Meta-modeling was Metis’s strongest feature. Users could define object types, relationship types, properties, constraints, behaviors, graphical symbols, and modeling rules—effectively building their own modeling language.

The capability is similar to what is now found in Eclipse EMF, JetBrains MPS, Microsoft DSL Tools, MetaEdit+, and Sirius, although Metis predates many of them.

2. Everything is an object

Almost everything inside Metis was represented as an object: processes, documents, people, organizations, requirements, goals, applications, databases, and roles. Relationships were also first-class objects with their own properties.

EmployeeWorks InDepartment

The Works In relationship could itself contain a start date, responsibility, percentage, and comments.

3. Multiple views

A single object could appear in many diagrams. A Customer, for example, could appear in an organization model, CRM model, process model, and information model. Every reference pointed to the same underlying object, so changing it updated every view.

This idea has become common in modern digital twins and knowledge graphs.

4. Repository-centric

Unlike drawing tools, Metis stored objects, relationships, and attributes in a shared repository. Diagrams were merely different visualizations of the same underlying information.

RepositoryDiagram ADiagram BDiagram CReportsQueries

Modeling methods and enterprise domains

Metis was intentionally methodology-independent. It shipped with templates for methods including UML, EEML (Extended Enterprise Modeling Language), GEM, MEML, i*, and Misuse Cases. Users could also combine methods into hybrid modeling languages.

It was particularly strong in enterprise architecture, supporting domains such as:

  • Business processes
  • Organization
  • Information
  • Applications
  • Infrastructure
  • Requirements
  • Goals
  • Strategies
  • Risks
  • Products
  • Regulations

Many Norwegian government organizations used Metis for enterprise modeling.

Beyond diagramming

Analysis, navigation, and collaboration

Impact and dependency analysis

Models could be queried and analyzed. A change to one requirement could identify affected business processes, applications, databases, and servers. Dependencies could likewise be traced from applications through interfaces and services to business processes.

Requirement↓Business process↓Application↓Database↓Server
Application↓Interfaces↓Services↓Business processes

Completeness checking

Rules could identify missing documents, orphan processes, missing owners, and invalid relationships.

Report generation

Reports could automatically generate documentation, HTML, specifications, inventories, and model catalogs.

Hierarchical navigation

Objects could contain sub-models. Double-clicking an object navigated deeper into the model—for example, from Company to Sales, Order Processing, Order Approval, and Invoice.

Team collaboration

Long before cloud collaboration became common, Metis supported shared repositories, versioning, distributed modeling, and sub-model merging, allowing multiple teams to work on different parts of a large enterprise model.

Technical architecture

Internally, Metis was based on a graph model: objects connected to other objects through relationships. Every object had a unique identity, properties, methods, graphical representations, and relationships.

ObjectRelationshipObjectRelationshipObject

This architecture closely resembles today’s property graph databases such as Neo4j.

Ahead of its time

Metis and its modern equivalents

CapabilityMetisModern equivalent
Repository modeling✓Sparx EA, Cameo
Meta-modeling✓EMF, MPS
Multiple diagram views✓Cameo, Enterprise Architect
Traceability✓Jama, DOORS
Knowledge graph concepts✓Neo4j, RDF
Digital twin concepts✓Azure Digital Twins
User-defined modeling languages✓Sirius, MetaEdit+
Impact analysis✓Modern EA platforms

Metis significantly influenced Scandinavian enterprise architecture research, particularly Enterprise Engineering, EEML, UEML, collaborative enterprise modeling, and knowledge management. Many of these initiatives involved collaboration between industry—including Computas—universities, and European research projects.

Conceptual lineage

How the vision evolves in Mimris

There is a clear conceptual lineage from Metis to Mimris, but the current vision extends it substantially.

MetisMimris vision
Meta-modelingAI-generated domain models
RepositoryKnowledge graph plus vector knowledge
Multiple viewsDomain, POPS, IRTV, TYPE, and BPMN layers
Object repositoryLiving AI workspace
Manual modelingAI-assisted collaborative modeling
ReportsExecutable workspaces and generated artifacts
Static model analysisAI reasoning and autonomous generation
Enterprise modelingDomain engineering and execution platform

Mimris can be viewed not merely as a successor to Metis, but as an evolution from a meta-modeling platform into an AI-native knowledge engineering environment that combines modeling, reasoning, generation, and execution.