Community-Driven Language for Variability Models

The Universal Variability Language is a community effort towards a unified language for variability models. UVL is a direct result of the efforts within the MODEVAR initiative.

Want to tryout UVL yourself? Check our Playground

Do you have any questions on UVL? Do you want to contribute? We gladly receive your message (Contact).

Language Design

UVL specifies variability models with a tree-like structure to represent the hierarchical structure of variability models. Below you can see a simple example of a variability model in UVL. The tree structure is visualized on the right side in the FeatureIDE representation. The language core of UVL comes with several concepts to specify constraints:

• Mandatory: The feature needs to be selected if its parent is selected
• Optional: The feature can be selected if its parent is selected
• Or: At least one of the children need to be selected if its parent is selected
• Alternative: Exactly one of the children needs to be selected
• Cross-tree constraints: Arbitrary propositional constraints over the features

features
    Sandwich
        mandatory
            Bread
        optional
            Sauce
                alternative
                    Ketchup
                    Mustard
            Cheese
constraints
    Ketchup => Cheese

Language Levels

In addition to the language core described above, UVL provides several extensions providing additional language constructs. The following language levels are supported:

• Boolean Level
  • Boolean Core: Level without any extension as seen above
  • Group Cardinality: Specify selecting [n..m] (between n and m) children parent-children relationships D
• Arithmetic Level
  • Arithmetic Core: Supports constraints over feature attributes with standard arithmetic operations (+,-,*,/,=,!=,>,<)
  • Attribute aggregates: Simplifies specification of attribute constraints by allowing sum() and avg() aggregates
  • Feature Cardinality: A single feature can be selected between n and m times
• Type Level
  • Type Core: Supports String Features
  • Type Numeric-Constraints: Allows numeric operations like floor(..) and ceil(..) within constraints
  • Type String-Constraints: Allows string operations like len(..) and comparison (of String Features) within constraints

The following feature model in UVL syntax enriches the previous version with new language concepts. Further, feature attributes, such as the number of calories, are attached to features.

include
    Boolean.group-cardinality
    Arithmetic.aggregate-function
    Arithmetic.feature-cardinality
    Type

features
    Sandwich
        mandatory
            Bread {Calories 100, Sugar 20}
        optional
            Sauce
                or
                    Ketchup {Calories 40, Sugar 35}
                    Mustard {Calories 25, Sugar 5}
            Cheese
                [0..2] // Group cardinality
                    Cheddar {Calories 60}
                    Gouda {Calories 50}
                    Goat {Calories 35}
            Pickle cardinality [1..3] // Feature cardinality
                    
constraints
    Ketchup => Cheese
    Bread.Sugar + Ketchup.Sugar + Mustard.Sugar < 60 // Attribute constraints
    sum(Calories) < 160 // Attribute aggregate

Import Mechanism

UVL supports referencing submodels for composing larger UVL models. See here for more details

Examples

A set of further example UVL models is available here. We are happy to merge your examples if you want to contribute.

Grammar

Below you can find a simplified grammar for UVL in ANTLR notation. This simplified version is supposed to provide a quick overview on the structure of an UVL model but leaves out many details. A full version of this grammar can be found here.

featureModel: namespace? NEWLINE? includes? NEWLINE? imports? NEWLINE? features? NEWLINE? constraints? EOF;

includes: 'include' NEWLINE INDENT includeLine* DEDENT;
includeLine: LANGUAGELEVEL NEWLINE;

namespace: 'namespace' reference;

imports: 'imports' NEWLINE INDENT importLine* DEDENT;
importLine: ns=reference ('as' alias=reference)? NEWLINE;

features: 'features' NEWLINE INDENT feature DEDENT;

group
    : ORGROUP groupSpec          # OrGroup
    | ALTERNATIVE groupSpec # AlternativeGroup
    | OPTIONAL groupSpec    # OptionalGroup
    | MANDATORY groupSpec   # MandatoryGroup
    | CARDINALITY groupSpec    # CardinalityGroup
    ;

groupSpec: NEWLINE INDENT feature+ DEDENT;

feature: reference featureCardinality? attributes? NEWLINE (INDENT group+ DEDENT)?;

featureCardinality: 'cardinality' CARDINALITY;

attributes: OPEN_BRACE (attribute (COMMA attribute)*)? CLOSE_BRACE;

attribute
    : valueAttribute
    | constraintAttribute;

valueAttribute: key value?;

key: id;
value: BOOLEAN | FLOAT | INTEGER | string | attributes | vector;
vector: OPEN_BRACK (value (COMMA value)*)? CLOSE_BRACK;

constraint
    : equation                              # EquationConstraint
    | reference                             # LiteralConstraint
    | OPEN_PAREN constraint CLOSE_PAREN      # ParenthesisConstraint
    | NOT constraint                        # NotConstraint
    | constraint AND constraint             # AndConstraint
    | constraint OR constraint              # OrConstraint
    | constraint IMPLICATION constraint     # ImplicationConstraint
    | constraint EQUIVALENCE constraint     # EquivalenceConstraint
	;

Software

Any tool support for UVL or tool integrating UVL missing? We are happy to embed your suggestion.

Tool Support for UVL

• Parser
  • ANTLR Base Parser (currently support for Python and Java): UVL Parser
  • Java realization of ANTLR parser with feature model representation: Java Meta Model
• Rust-based Language Server Protocol for integrating UVL in your favorite IDE: UVLS
• Web-based playground based on UVLS UVL Playground
• Visual Studio Code extension based on UVLS

Tools Integrating UVL

• IDE for feature-oriented software development: FeatureIDE
• Python-based analysis frame for variability models: Flama
• Transformations between different variability modelling approaches: TraVarT

Publications

Here, we only present a small selection of papers. Any interesting work on UVL missing? We are happy to embed your suggestion.

Work on UVL

• Chico Sundermann, Kevin Feichtinger, Dominik Engelhardt, Rick Rabiser, and Thomas Thüm. 2021. Yet another textual variability language? a community effort towards a unified language. SPLC '21. https://doi.org/10.1145/3461001.3471145
• Chico Sundermann, Stefan Vill, Thomas Thüm, Kevin Feichtinger, Prankur Agarwal, Rick Rabiser, José A. Galindo, and David Benavides. 2023. UVLParser: Extending UVL with Language Levels and Conversion Strategies. SPLC '23 - Tool Track. https://doi.org/10.1145/3579028.3609013
• Chico Sundermann, Tobias Heß, Dominik Engelhardt, Rahel Arens, Johannes Herschel, Kevin Jedelhauser, Benedikt Jutz, Sebastian Krieter, and Ina Schaefer. 2021. Integration of UVL in FeatureIDE. MODEVAR@SPLC '21. https://doi.org/10.1145/3461002.3473940
• Jacob Loth, Chico Sundermann, Tobias Schrull, Thilo Brugger, Felix Rieg, Thomas Thüm. 2023. UVLS: A Language Server Protocol For UVL. SPLC '23 - Tool Track. https://doi.org/10.1145/3579028.3609014
• Kevin Feichtinger, Johann Stöbich, Dario Romano, and Rick Rabiser. 2021. TRAVART: An Approach for Transforming Variability Models. VaMoS '21. https://doi.org/10.1145/3442391.3442400
• Dario Romano, Kevin Feichtinger, Danilo Beuche, Uwe Ryssel, and Rick Rabiser. 2022. Bridging the gap between academia and industry: transforming the universal variability language to pure::variants and back. MODEVAR@SPLC '22. https://doi.org/10.1145/3503229.3547056
• Prankur Agarwal, Kevin Feichtinger, Klaus Schmid, Holger Eichelberger, and Rick Rabiser. 2024. On the Challenges of Transforming UVL to IVML. MODEVAR@VaMoS'24. https://doi.org/10.48550/arXiv.2403.01952
• José A. Galindo and David Benavides. 2020. A Python framework for the automated analysis of feature models: A first step to integrate community efforts. MODEVAR@SPLC '20. https://doi.org/10.1145/3382026.3425773
• José A. Galindo, Jose-Miguel Horcas, Alexander Felferning, David Fernandez-Amoros, and David Benavides. 2023. FLAMA: A collaborative effort to build a new framework for the automated analysis of feature models. SPLC' 23 - Tool Track. https://doi.org/10.1145/3579028.3609008
• Jose M. Horcas, Jose A. Galindo, Mónica Pinto, Lidia Fuentes, and David Benavides. 2022. FM fact label: a configurable and interactive visualization of feature model characterizations. SPLC '22. https://doi.org/10.1145/3503229.3547025
• David Romero, José Á. Galindo, Jose-Miguel Horcas, and David Benavides. 2021. A first prototype of a new repository for feature model exchange and knowledge sharing. SPLC '21. https://doi.org/10.1145/3461002.3473949
• David Romero, José A. Galindo, Chico Sundermann, Jose-Miguel Horcas, and David Benavides. 2024. UVLHub: A feature model data repository using UVL and open science principles. JSS https://doi.org/10.1016/j.jss.2024.112150Y

Work Relevant for the Design of UVL

This is only a small selection of work that was vital for the design of UVL.

• Maurice H. ter Beek, Klaus Schmid, and Holger Eichelberger. 2019. Textual Variability Modeling Languages: An Overview and Considerations. MODEVAR@SPLC '19. https://doi.org/10.1145/3307630.3342398
• Thorsten Berger and Philippe Collet. 2019. Usage Scenarios for a Common Feature Modeling Language. MODEVAR@SPLC '19. https://doi.org/10.1145/3307630.3342403
• Thomas Thüm, Christoph Seidl, and Ina Schaefer. 2019. On Language Levels for Feature Modeling Notations. MODEVAR@SPLC '19 https://doi.org/10.1145/3307630.3342404

Contributors

Think you should be on this list as a past and/or future contributor? Don't hesitate to contact us.

Core Team

• University of Seville
  • David Benavides
  • José A. Galindo
• Johannes Kepler University Linz
  • Rick Rabiser
• Karlsruhe Institute of Technology
  • Kevin Feichtinger
• University of Paderborn
  • Thomas Thüm
• University of Ulm
  • Chico Sundermann

Further and Former Contributors

• Dominik Engelhardt (TU Braunschweig)
• Dario Romano (JKU Linz)
• Stefan Vill (University of Ulm)
• Jacob Loth (University of Ulm)
• Prankur Agarwal (JKU Linz)
• David Romero-Organvidez (University of Seville)