=Paper=
{{Paper
|id=Vol-2542/MOD20-TuD4
|storemode=property
|title=CoreWEB - Semantic Expressions in Conceptual Models for Generation of Information Systems
|pdfUrl=https://ceur-ws.org/Vol-2542/MOD20-TuD4.pdf
|volume=Vol-2542
|authors=Tomas Jonsson,Matts Rimfors
|dblpUrl=https://dblp.org/rec/conf/modellierung/JonssonR20
}}
==CoreWEB - Semantic Expressions in Conceptual Models for Generation of Information Systems==
https://ceur-ws.org/Vol-2542/MOD20-TuD4.pdf
Joint Proceedings of Modellierung 2020 Short, Workshop and Tools & Demo Papers
208 Modellierung 2020: Tools & Demo Papers
CoreWEB - Semantic Expressions in Conceptual Models for
Generation of Information Systems
Tomas Jonsson,1 Matts Rimfors2
Abstract: Information systems are partially defined by conceptual models. Conceptual models
declare structures of concepts which need to be augmented with semantic logic in some form in
order to get towards fully functional information systems. CoreWEB is an environment with a one
button transformation of conceptual models with expressions declaring semantics, into executing
information system. In CoreWEB, semantic logic is added to declared concepts by defining expressions
(parameterless functions without side effects). In the executing information system, consistency of
data as stated by expressions is maintained by an event driven data-flow mechanism.
Keywords: Model Driven Systems; Information Systems; Conceptual Models; Semantic Models;
Model Execution; Data-flow
1 Introduction
Our research and development focus is on seamless life cycle management of Information
Systems (is) which are semantically consistent with their users. As part of this effort,
CoreWEB was developed as a tool for education, research and prototyping.
Model driven development and execution of is has a long history [JE15; Pa04; PG92; Re77].
Automated transformation of is systems, from requirements and conceptual models, is
available from some independent suppliers and has proven to work for full scale enterprise
information systems. Documented benefits of automatic transformation of domain models,
a category which includes conceptual information systems models, into software systems
include lead time, cost and quality in an order of magnitude [MD08].
We consider pure conceptual models as declarations of structures of concepts, whereas
semantic models also need definitions of the meaning of each concept. For information
systems construction, we consider conceptual models describing structures of data to be
suitable. For the purpose of automatic transformation of these models into information
systems, formal machine readable definitions of concepts are required. I.e. rules regarding
the concept and if the concept is a property, what is the rule for the value of this property.
In many approaches, these kinds of rules are embedded in processing concepts separate
from the data concept they affect, such as in object methods, object constraints, processes
1 Genicore AB, R&D, Östra Hamngatan 52, S-41109 Göteborg, Sweden, tomas@genicore.se
2 Genicore AB, R&D, Östra Hamngatan 52, S-41109 Göteborg, Sweden, matts@genicore.se
Copyright © 2020 for this paper by its authors.
Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CoreWEB and Semantic Expressions 209
or transactions. This allows for multiple and conflicting definitions of concept rules. In our
demonstrated modelling tool it is only possible to annotate declared data concepts with its
logic definitions. Thus, if a data value is incorrect, conflicting definitions can be ruled out
and the location of the fault is obvious.
CoreWEB, previously demonstrated [JE17] as an object oriented entity relational conceptual
modelling and execution environment, has been extended with the ability to define formal
semantics as expressions on object attributes. Expression execution is at runtime controlled
by a generated data flow engine, working with two phases in the expression dependency
Directed Acyclic Graph (DAG). First phase is a bottom up mark (properties to be recalculated)
and second phase is top down DAG traversal to recalculate values. This approach ensure
transactional integrity and avoids multiple recalculations of the same value for single data
change events.
In order to simplify implementation and providing generality of expression language,
expression syntax and semantics is simply javascript, executed by the WEB browser.
2 Calculations in Sales Order Model
To exemplify concept definition using expressions, a model of an order is used and the
expression of calculating the price of the order as the sum of the price of the order items. In
Fig. 1 the object types of the order is shown and their relations. The notation was developed
for information modelling and is briefly explained in [JE19] page 10)
Fig. 1: Object types of an order system, where Orders can have relation to many (circle symbol) order
items of type Amount and an Amount can have a relationship to one (rectangle symbol) product type.
Fig. 2: Panel for defining expressions, in this case order price as the sum the price of order items.
210 Tomas Jonsson, Matts Rimfors
Fig. 2 shows the panel for expression definition, where properties of orders and related
amounts are listed, as well as the sum button which conveniently generate the javascript
pattern of summing.
3 Calculation for Summing Node Values of a Tree Structure
In this example, we define a tree graph of nodes and relations, with an expression calculating
the sum of node values and the consecutive sum of all tree branches. Instead of the recursive
tree traversal approach of imperative or functional languages, in the expressional data-flow
approach nodes sum is just declared as the sum of its children plus it own value. Node
object type and relations is shown in Fig. 3
Fig. 3: Node object type and its relations to subnodes and super node
Expression in principle: Node.treeValue = sum subnodes.treeValue + nodeValue.
Javascript pattern is in this case partly generated by selecting the relation (subnodes) and
the property to summarise (tree value) and then adding the node value itself. Expression in
javascript is shown in Fig. 4 and runtime system in Fig. 5
Fig. 4: Panel for defining expressions, in this case sum of tree nodes
CoreWEB and Semantic Expressions 211
Fig. 5: Partial node tree during execution displaying tree-value by the generated user interface. All
nodes have a node value of 10 and there are 17 nodes in total.
4 Concluding remarks - Models as the New Software
CoreWEB is an experimental environment which lacks several features that would be
required for commercial use, however it demonstrates the concept of Models as Software.
CoreWEB is currently used for teaching, semantic modelling and application prototyping in
research, as well as for a tailor made and deployed system for a role play organisation with
600 members. A commercial environment based on similar principles, used to implement
an ERP system is described in [JE15].
CoreWEB environment registration and some demonstration videos is available at
https://www.ameis.se/cml. Educators and researchers are encouraged to use the free
environment CoreWEB to model, execute and experiment with models. We urge researchers
and practitioners with interest in model driven is to join in a triple helix effort
https://www.ameis.se to develop and disseminate knowledge and practice.
References
[JE15] Jonsson, T.; Enquist, H.: CoreEAF -A Model Driven Approach to Information
Systems. In: CEUR Workshop Proceedings. CAISE2015. Vol. 1367, Stockholm,
pp. 137–144, June 2015.
[JE17] Jonsson, T.; Enquist, H.: Semantic Consistency in Enterprise Models-Through
Seamless Modelling and Execution Support. In: CEUR Workshop Proceedings.
ER2017. Vol. Vol-1979, Valencia, p356–359, Nov. 2017.
212 Tomas Jonsson, Matts Rimfors
[JE19] Jonsson, T.; Enquist, H.: Phenomenological Framework for Model Enabled
Enterprise Information Systems. In (Welzer, T.; Eder, J.; Podgorelec, V.; Wrem-
bel, R.; Ivanović, M.; Gamper, J.; Morzy, M.; Tzouramanis, T.; Darmont, J.;
Kamišalić Latifić, A., eds.): New Trends in Databases and Information Systems.
Vol. 1064, Springer International Publishing, Cham, pp. 176–187, 2019.
[MD08] Mohagheghi, P.; Dehlen, V.: Where Is the Proof? - A Review of Experiences from
Applying MDE in Industry. In (Schieferdecker, I.; Hartman, A., eds.): Model
Driven Architecture – Foundations and Applications. Vol. 5095, Springer Berlin
Heidelberg, Berlin, Heidelberg, pp. 432–443, 2008.
[Pa04] Pawson, R.: Naked Objects, Dublin, Ireland: University of Dublin, Trinity College,
June 2004, 223 pp.
[PG92] Pastor, O.; González, A.: OO-Method: An Object-Oriented Methodology for
Software Production. In (Tjoa, A. M.; Ramos, I., eds.): Database and Expert
Systems Applications. Springer Vienna, Vienna, pp. 121–126, 1992.
[Re77] Reenskaug, T.: PROKON/PLAN-A Modelling Tool For Project Planning And
Control. In: IFIP Proceedings. IFIP. Toronto, Canada, pp. 717–722, 1977.