Towards a Uniform User Interface for Editing Data Shapes Ben De - - PowerPoint PPT Presentation

towards a uniform user interface for editing data shapes
SMART_READER_LITE
LIVE PREVIEW

Towards a Uniform User Interface for Editing Data Shapes Ben De - - PowerPoint PPT Presentation

Towards a Uniform User Interface for Editing Data Shapes Ben De Meester , ben.demeester@ugent.be, @Ben__DM Pieter Heyvaert, Anastasia Dimou, and Ruben Verborgh IDLab, Department of Electronics and Information Systems, Ghent University


slide-1
SLIDE 1

Towards a Uniform User Interface for Editing Data Shapes

Ben De Meester,  ben.demeester@ugent.be,  @Ben__DM Pieter Heyvaert, Anastasia Dimou, and Ruben Verborgh IDLab, Department of Electronics and Information Systems, Ghent University – imec, Ghent, Belgium

slide-2
SLIDE 2

Garbage In, Garbage Out?

Data shapes

2

slide-3
SLIDE 3

Easy editing is important

“Fitness for use” Constraint languages: declaration and implementation are decoupled

SHACL (W3C Recommendation) ShEx …

Machine-processability in mind 3

slide-4
SLIDE 4

What are the necessary features for visually editing data shapes?

4

slide-5
SLIDE 5

Outline

SOTA Features PoC: unSHACLed

UI Features

Conclusions 1 2 3 4 5

slide-6
SLIDE 6

Outline

SOTA Features PoC: unSHACLed

UI Features

Conclusions 1 2 3 4 6

slide-7
SLIDE 7

State of the Art

Data shapes

Validation based on: OWL | SPARQL | SHACL, ShEx

Data shape editors

Depend on the (constraint) language or enforce a linear workflow

Editors

Data editors: text-based, form-based, use-case specific Ontology editors: graph-based, indented-tree-based, UML-based SPARQL editors: text-based Linked Data generation rule editors: form-based, graph-based

7

slide-8
SLIDE 8

Outline

SOTA Features PoC: unSHACLed

UI Features

Conclusions 1 2 3 4 8

slide-9
SLIDE 9

Desired Features for Data Shape Editing

Independence of constraint language Support multiple data sources Support different serializations Support multiple ontologies Multiple alternative editing approaches Non-linear workflows Independence of execution 1 2 3 4 5 6 7 9

slide-10
SLIDE 10

Outline

SOTA Features PoC: unSHACLed

UI Features

Conclusions 1 2 3 4 10

slide-11
SLIDE 11

unSHACLed

Visual Data Shapes editor as Web application Drag-and-drop loaded data graphs and data shapes Add data shapes using templates Get visual feedback on conformance Export shape https://w3id.org/imec/unshacled/app 11

slide-12
SLIDE 12

The unSHACLED UI, consisting of an Overview Sidebar (left), an Action Toolbar (top), and an Editing Area (middle-right)

12

slide-13
SLIDE 13

Different elements enable the different features

Features UI Element Overview Sidebar Action Toolbar Editing Area

  • F1. Independence of constraint language

✓ ✓

  • F2. Support multiple data sources

✓ ✓

  • F3. Support different serializations

  • F4. Support multiple ontologies

  • F5. Multiple alternative modeling approaches

  • F6. Non-linear workflows

✓ ✓ ✓

  • F7. Independence of execution

13

slide-14
SLIDE 14

Outline

SOTA Features PoC: unSHACLed

UI Features

Conclusions 1 2 3 4 14

slide-15
SLIDE 15

Conclusions

No need to write RDF / SHACL / ShEx Use-case independent Open Issues

(Map)VOWL or UML or … ? User evaluation graphical representation Representation large data shapes

Workspaces Detail levels

Features as starting point for visual data shape editors 15

slide-16
SLIDE 16

Towards a Uniform User Interface for Editing Data Shapes

Ben De Meester,  ben.demeester@ugent.be,  @Ben__DM Pieter Heyvaert, Anastasia Dimou, and Ruben Verborgh IDLab, Department of Electronics and Information Systems, Ghent University – imec, Ghent, Belgium