=Paper=
{{Paper
|id=Vol-1322/paper_11
|storemode=property
|title=Enabling Access to Environmental Models, Data, and Services on the Web - Technical Results Summary from the ENVISION Project
|pdfUrl=https://ceur-ws.org/Vol-1322/paper_11.pdf
|volume=Vol-1322
}}
==Enabling Access to Environmental Models, Data, and Services on the Web - Technical Results Summary from the ENVISION Project==
https://ceur-ws.org/Vol-1322/paper_11.pdf
Enabling Access to
Environmental Models, Data, and Services on the Web
– Technical Results Summary from the ENVISION Project –
1
Dumitru Roman, 2Tertre Francois, 3Alejandro Llaves, 4Miha Grcar, 4Maja Skrjanc,
5
Ioan Toma, 6Michael Pantazoglou, 7Silviu Trasca, 1Nils Rune Bodsberg,
8
Morten Borrebæk
1
SINTEF, Oslo, Norway
Dumitru.Roman@sintef.no
2
BRGM, France. 3 University of Muenster, Germany, 4 JSI, Slovenia,
5
University of Innsbuck, Austria, 6 University of Athens, Greece, 7 CS, Romania,
8
Norwegian Mapping Authority, Norway
Abstract. The Environmental Services Infrastructure with Ontologies
(ENVISION) project (2010-2013) provided an IT infrastructure for non
ICT-skilled users for semantic discovery and adaptive chaining and
composition of environmental services. This paper summarizes the core
results of the project with a focus on individual components, relevant
stakeholders, and overall advancements made by the project.
1 Introduction
The ENVISION project (http://www.envision-project.eu/) ran from 2010 to 2012 and had as
its main goal the development of an IT infrastructure supporting users with limited ICT skills in
decision making processes involving environmental services. The project addressed emerging
topics related to environmental services, ranging from semantic discovery, chaining and
execution of environmental services, to migration of environmental models to be provided as
models-as-a-service (MaaS) [1], to use of data streaming information for harvesting
information for dynamic building of ontologies.
The ENVISION project combined and extended tools and components with functionality
for easier use by non ICT-skilled users and with increased semantic technology support in an
incremental development approach. Figure 1 shows the main focus areas and contributions in
ENVISION. Each area is accessible to the one above.
The ENVISION Execution infrastructure provides the basis for resource discovery and
composition. Semantic interoperability is facilitated by the ENVISION Ontology infrastructure,
which contains ontologies, resource descriptions, and supporting tools. Both mentioned
components provide input to the ENVISION Portal and Development tools. This area is also
responsible for providing client components for interfacing with the ontology and execution
infrastructure. All areas in conjunction are applied to the ENVISION Communities. These
communities require application specific decision support. Customised portals are developed
for three pilot cases (landslide, oil spills at sea, and flooding), which serve as a proof of concept
for the project.
Figure 1. ENVISON focus and contribution areas
In the following we give an overview of specific areas of contributions related to the
aforementioned architecture. For each artefact we provide a brief description, discuss relevant
stakeholders, discuss advancements in the project, and summarize the core S&T results.
2 ENVISION Pilots – Scenarios Websites
Landslide Scenario Website
A scenario website1 has been set up to present the results of a simulation predicting
the potential location of landslide in the Mamelles area. In this website, users can
initiate the execution of a model calculating risks for each area a landslide occurs.
Stakeholders: The Landslide website is dedicated to citizens and also to public
authorities who may use it as a help to know if they need to take specific measure
after some meteorological event.
Advances in the project: The components of the landslide model come from a desktop
application, ENVISION had given the opportunity to make them available on the
Web using OGC standards.2
Core S&T results: The project delivered a new way for non ICT-skilled domain
experts to easily create websites for large diffusion of results of their studies. It also
1 The Landslide Scenario Website can be viewed on the BRGM ENVISION portal at: http://envision-
portal.brgm-rec.fr/en/web/rd23-landslides/mamelles-road.
2 http://www.opengeospatial.org/standards/is
delivered a tool to easily combine atomic process to create new workflow for the
computation of environmental models. The project also delivered to the landslide
community a new technology to share easily the models they create. With this
technology, the public authorities can also access in a faster way the results of
simulation without the need of a landslide expert.
Oil Spill Scenario Website
For the Oil Spill pilot case, a scenario website was established.3 At this website, users
may define an oil spill scenario in the North Sea region, run a simulation that predicts
the fate and effects of the spill, and then visualize the results as animated layers on a
map.
Stakeholders: The Oil Spill Scenario Website is oriented towards researchers (in the
oil spill domain, or in related domains), private companies and public agencies
involved in oil spill contingency planning, and also the general public.
Advances in the project: Traditionally, oil drift models are embedded in applications
and can only be accessed via that application's user interface. In ENVISION,
SINTEF's oil drift model OSCAR4 has been made accessible via a standard web
interface.
Core S&T results: The project delivered a new and user-friendly approach for
creating a domain-oriented website for the oil spill community. The web site enables
invoking and running composite model services, and also visualizing and analyzing
simulation results. The web site may easily be adapted or extended to cover other
geographical regions.
Flood Monitoring Scenario Website
For Flood Monitoring pilot case a website scenario was setup on an instance of the
ENVISION platform5. The scenario website allows the users to observe current (real-
time) and historical water levels and water flows values on Danube River and its main
tributaries in the area controlled by the Iron Gates hydro-power plant
(Romania/Serbia). In case of high-waters events, a model was implemented for water
evacuation through dam gates in order to reduce or avoid the flood risk in this area.
Stakeholders: The scenario website provides valuable information for hydro-power
plant operators for both Romanian and Serbian sides, for the decision makers in the
public institutions (i.e. hydrologic services, waters administration, emergency
situations) and also for the general public.
Advances in the project: Currently, such floods monitoring applications are internally
handled by each interested organisation, using custom desktop or Web applications
which hinder data sharing and modelling results exchange.
Core S&T results: The pilot case allows hydrologists and hydro-power plant operators
to easily create websites for monitoring different interest areas were sensors are
3 The website can be accessed via envision.envip.eu.
4 http://www.sintef.no/Materialer-og-kjemi/Marin-
miljoteknologi/Miljomodellering/Modellverktoy/OSCAR-Oil-Spill-Contingency-And-Response/Model-
Description/.
5 http://envision.c-s.ro.
available. Existing data and processing services can be combined to provide new add-
value services.
Further details about the pilot cases can be found in [2,3].
2 ENVISION Portal
Environmental Decision Support Portal
The Environmental Decision Support Portal plays a major role in the infrastructure,
allowing non ICT-skilled users to manage by themselves the whole workflow of
search, creation, diffusion of environmental services.
Stakeholders: The potential users are domain experts, who are using environmental
models or creating new ones and want to share their results; end users, citizens, public
organizations, who want to consult the results of environmental models on a certain
thematic on a specific area.
Advances in the project: The project has made available some components pluggable
in a web portal such as Liferay.6 These components allow a non ICT-skilled user to
perform the tasks of the ENVISION workflow (semantic discovery, annotation,
publication in the catalogue, composition of environmental services, visualization). In
these pluggable components, some of them are dedicated for the creation of Scenario
Web Site by the domain expert. These components are used by the domain experts to
allow the visualization of the results of the models they have created. These
components allow easily and without any IT knowledge to add in a web site a map for
visualization of data coming from OGC services (raster data from a WMS,7 vector
data from a WFS,8 data series from a SOS9); they also allow to visualize the data in a
chart and to view the evolution of a phenomena in the map during a given time. All
these tools are easily configurable directly in the web site in a visual way and a role
based security can be activated for each web site, page or component to restrict the
access to non-authorized people. The final user can then visualize the results of the
environmental models on a map or on a chart in the context defined by the domain
expert.
Core S&T results: The project delivered (a) a new approach for the creation of
thematic web sites for the diffusion of environmental data generate by the execution
of chained models; (b) an open source implementation allowing a non ICT-skilled
user to easily create web site containing tools for the visualization of data coming
from OGC services and for results of the execution of environmental models.
Further details about the ENVISION portal can be found in [4].
6 https://www.liferay.com/
7 http://www.opengeospatial.org/standards/wms
8 http://www.opengeospatial.org/standards/wfs
9 http://www.opengeospatial.org/standards/sos
Composition of OGC Services
In order to create the applications of the ENVISION use cases (land slide, oil spill,
cod effects and flood monitoring), there is a need to call several OGC services and
chain them together. This is done with a so-called service composition. The OGC
services used in the land slide scenario: an SOS service to retrieve the amount of
precipitation, a WCS service to get a digital elevation model of the Guadeloupe island
(the area of interest for the land slide scenario), a WPS to calculate a hydrological
model, and finally a WPS to calculate the probabilities of a landslide. The OGC
services used in the oil spill scenario: A WFS to retrieve coastline information, a
WCS to retrieve sea depth information, a WPS to retrieve weather information, and
finally a WPS to predict how the oil spill spreads over time.
The project has developed a new approach for composing OGC services by using a
graphical modeling language to model the control and data flow. It is based on current
standards and tailored for OGC services. Technical details are automatically
registered and hidden from the user to lower the complexity level in using the tool.
Stakeholders: The development of a composition approach is directed to developers
of new services that want a fast and efficient support for putting together
environmental services without the need to code or dig into technical details. The
approach also has the potential to reach non-ICT skilled domain experts with some
future development and maturity of the given platform in the feature.
Advances in the project: The project provided new insights in the composition of
OGC services. The approach is tailored for the specifics of OGC services, where other
approaches only address Web service composition in general. We have identified the
typical challenges when mediating between OGC services so to enable that they can
be chained together. While mediation between Web services in general is a very
complicated matter, it turns out that many typical scenarios of mediating between
OGC services can be handled in a semi-automated and simple manner with the gained
knowledge from the ENVISION project.
When the project started there was no tool available to search, register and integrate
OGC services over the Web. This is now achieved by the ENVISION platform.
Core S&T results: The project delivered a new approach for (a) composition of OGC-
based service, (b) a composition platform integrated with discovery and registration
capabilities (c) a composition platform integrated with a mediation framework
described below, and (d) fully automated generation of WSDL and BPEL for the
deployment and execution of OGC service compositions.
Further details about the composition approach can be found in [5].
3 ENVISION Semantics
Semantic Annotation
Ontologies and ontology management play an important role in the ENVISION
infrastructure because in order to enable efficient browse and search through
resources and efficient composition and execution of Web services, the resources
need to be semantically annotated. Generally, resources such as Web services can be
annotated in different ways. In ENVISION, semantic annotation is defined as a set of
interlinked domain-ontology elements associated with the resource being annotated.
Stakeholders: The potential users are domain experts, which are dealing with
semantic annotations of different data sources, possibly cross-lingual.
Advances in the project: The technology in ENVISION, which enables visual
management (creation and editing) of semantic annotations and ontology querying,
was implemented in a software component named Visual OntoBridge. Visual
OntoBridge integrates several existing scientific methods but also implements several
novel approaches for cross language annotations. Visual OntoBridge is implemented
as a portlet (a pluggable user interface software component) which can be plugged
into a portal (ENVISION employs the Liferay Portal).
Core S&T results: Core results are within the areas of domain ontology querying,
cross-language domain ontology querying, and visual editing of semantic annotations,
as explained below.
Domain ontology querying is implemented by employing text mining techniques, the
PageRank algorithm and general Web search. The user can enter a Google-like query
and receives a list of ontology concepts which are sorted from the most to the least
relevant. To achieve this, Visual OntoBridge implements a number of text mining
methods and a variant of the PageRank algorithm which exploits the ontology
structure enriched with documents obtained from the web.
Cross-language domain ontology querying: The core idea of the machine-aided
annotation in ENVISION is based on term matching through groundings obtained by
a Web search engine. The user can enter a query in any of the supported languages
and receive a list of relevant results (which may be slightly different from language to
language and in different order).
Visual editing of semantic annotations: Establishing semantic annotations of resource
(such as Web services) using big and complex domain ontologies is not an easy task,
especially for users which are not familiar with underlying technologies. For that
reason, Visual OntoBridge provides technologies which make the creation of
semantic annotations easy and visually appealing. Visual OntoBridge implements an
application independent annotation editor which employs graph representations of
ontologies and resources to simplify the annotation. Thus, the act of creating a
semantic annotation is represented by establishing connections between graph nodes
representing ontology concept instances and the resource, also shown as a graph.
Visual OntoBridge enables visual annotation editing which does not require specific
skills or knowledge. The user is only required to be familiar with the topic of the
domain ontology and its relation to the resource (web service). The process of
creating an annotation is represented through graph editing actions (adding/removing
edges by connecting/disconnecting nodes representing the resource and concept
instances).
Further details about the annotation approach can be found in [6,7].
Ontology Management for the Semantic Annotation of Environmental Models
ENVISION provided an online platform to support the migration of environmental
models to be provided as models as a service (MaaS). The development of ontologies
for semantic annotation of Web services required a proper ontology engineering
methodology and tools which allow ontology maintenance. These annotations are
relevant for the semantic discovery of resources, the mediation between services, and
the execution of the environmental workflows.
Stakeholders: The ontology management strategy presented in ENVISION addresses
scientists who are not ontology experts. They are interested in sharing their
environmental models as service compositions with the scientific community in order
to get feedback and to explore the possibilities of reusability of the model in other
scenarios.
Advances in the project: ENVISION extended and improved previous work on
semantic annotations for OGC services with a methodology for adding semantic
annotations to the OGC service specifications (SOS, WFS, WPS, WCS, and WMS).
The methodology also covers the annotation of WSDL services. Additionally,
ENVISION provided tools for the online management of resources as Web services
(both standard W3C services as well OGC-compliant Web services), WSML
ontologies, and BPEL composition drafts.
Core S&T results: Some of the main components that the project delivered for the
management of semantic annotations of environmental models include: Resource
management (the Resource Portlet manages the access to all resources required to
perform the individual activities of the ENVISION platform); Service Model
Translator (SMT) (The SMT translates a capabilities document provided by a Web
service into RDF-based and WSDL service descriptions. The supported specifications
are: SOS, WFS, WPS, WCS, and WMS. SMT creates both description representations
for each provided feature type or observed property of the corresponding Web
service); and Data Models, Service Models and Domain Ontologies.
Further details about the annotation approach can be found in [7].
4 ENVISION Execution
Semantic Discovery
Semantic discovery is a key component of ENVISION infrastructure enabling users to
find relevant OGC resources and services. It provides an intelligent and precise
discovery mechanism as part of what we call the Semantic Catalogue. The Semantic
Catalogue provides a semantic extension to standard OGC discovery of services,
which uses semantic annotations and reasoning over service descriptions formalized
by means of logics.
Stakeholders: The Semantic Discovery functionality can be used by a wide range of
users with different technical skills. By using the Semantic Discovery Portlet, non
ICT-skilled users can easily search for environmental services registered in the
catalogue. Queries can be easily specified either as keywords or using the Visual
OntoBridge. More technical users i.e. developers can use the results of querying the
Semantic Catalogue to create new services by using the composition component.
Advances in the project: In ENVISION we advanced the state of the art with respect
to discovery of environmental services in several aspects. To implement our solution
we used an open source catalogue (i.e. GeoNetwork) and extended it to support
semantic queries. The Frozen Facts approach for query containment, was used for
semantic discovery. Developed as part of the SWING project, the Frozen Facts
approach for query containment, part of the IRIS reasoner has been extended and
optimized. We have added support for negation, built in predicates, etc.
Core S&T results: The project delivered two main results: (1) a robust service
discovery mechanism with full support for semantic queries based on WSML goals
and query containment algorithms and (2) a well integrated solution of an open source
OGC catalogue (i.e. GeoNetwork10) with the semantic service discovery mechanism
mentioned before. We have also provided a user interface for the discovery inside the
ENVISION platform, wrapped in a portlet and provided the necessary interface to
connect the portlet with both the Resource Module and the WSML goal editor in
order to fully integrate the functionality provided by the ENVISION platform.
Further details about the discovery approach can be found in [8].
Mediation Framework
The Mediation Framework enables the rapid prototyping of data mediation
algorithms. Developers can thus quickly build, deploy and evaluate mediation
services on specific or public data sets. It also permits to communicate data models,
and to evaluate the effectiveness of existing solutions. The resulting mediation
services can be easily shared with the data mediation community – if needed, but can
also be integrated into existing software architectures, as done in the ENVISION
project.
Stakeholders: This framework targets users who face data mediation issues (e.g.
software engineers, etc.), typical in service compositions, document exchanges,
interoperability, etc. It lightens the development of algorithms by providing basic
mediation features as a library, and supports the evaluation of the resulting mediation
algorithms by automating comparisons with alternative solutions on predefined data
sets.
Advances in the project: Data mediation is a complex issue, for which there is
unfortunately no "silver bullet". Recent approaches strive for the development of
generic algorithms able to solve any mediation problem. This framework concretizes
an alternative approach focusing on the rapid prototyping of application-specific
algorithms, where one accepts to lower algorithms reusability in order to provide
higher effectiveness by making the most of application specificities.
Core S&T results: The project delivered two main results: a new approach for rapid
prototyping of data mediation and the supported open-source implementation
available on Github.
Further details about the mediation approach can be found in [5].
10 http://geonetwork-opensource.org/
Stream Mining of Environmental Data
In the environmental domain Stream Mining techniques are still not widely present.
With the developed EnStreaM component in ENVISION we have shown the usability
and applicability of Stream Mining in the environmental domain. One of the
environmental data features is that for complete analysis of environmental
phenomena, one has to combine stream data on one hand and “static” data on the
other hand. In the development of stream mining component (EnStreaM) this fact is
introduced and resolved in a way, that the component is able to “listen” and analyze
the data streams as well as offering the possibility of import the history data about the
observed environmental phenomena in the conventional data format. The developed
component enables stream mining methods and a prototype system for handling
semantic data streams and stream ontologies including the information from the
sensor data streams.
Stakeholders: Potential users of stream mining in environmental data are domain
experts, specialized for observed environmental phenomena. Part of EnStreaM
component is also user friendly graphical interface, which offers complex
visualization of stream and historical data analysis, which is intuitive enough also for
non-ICT skilled users. However, the configuration process demands some basic
understanding regarding semantic annotations and its usage.
Advances in the project: The EnStreaM component is designed in a way that supports
import of various data (sensor data and static-metadata about selected environmental
phenomena) and provides wide range of functionalities: from simple to complex
browsing of sensor data, discovery and validation of expert rules (which could be
used for alarm triggering), anomaly detection in almost real-time (which enables
detection of broken sensor) and prediction of selected phenomena in various time
spans. Visualization of selected sensor on the map and ability to selected sensors from
the map itself enable users to combine data from appropriate geographical areas in
very straightforward way.
Core S&T results: Developed functionalities enable support for: Data stream
summarization in a configurable manner; Discovery of anomalies in the data stream;
Prediction of events; The expert domain user to semi-automatically generate, validate
and export rule for selected events in environment, measured by sensors; Rule export
in DataLog, JSoN and RuleML format.
Further details about the mediation approach can be found in [9].
Notification Infrastructure based on Semantic Event Processing
In domain applications dealing with environmental change, like flood monitoring, it is
normally required to exchange geospatial information about relevant occurrences as
they are detected. Different information communities may use diverse models to
represent changes in our environment, thus causing interoperability problems when
the information they produce is shared. In ENVISION, we proposed a layered
ontology model based on event processing to detect and classify occurrences derived
from sensor observations, and described them using domain knowledge. Such model
is integrated into our Notification Infrastructure which offers tools for event
subscription and notification by email.
Stakeholders: There are two types of users for the Notification Infrastructure. Domain
experts are needed to register semantically annotated event definitions that are
essential for the semi-automatic classification of occurrences. On the other hand, the
event subscription interface is designed for non ICT-skilled users, but can be used by
anyone.
Advances in the project: The application of Semantic Event Processing to geospatial
information is a relatively new field. In the last years, there has been some research
work mostly on event-based ontologies and detection of occurrences in time series of
observations, but not focusing too much on the interoperability problems that
different perspectives and application purposes can cause. Our solution is designed to
accept representations of events from multiple domains involved in environmental
monitoring. Additionally, we use a semantic-based notification system to avoid that
users have to deal with technical event definitions.
Core S&T results: The main components that are included in the Notification
Infrastructure based on Semantic Event Processing are: Event Processing Service
(EPS); Event-Observation ontology; Semantic Notification Broker (SNB);
Subscription Portlet.
Further details about the mediation approach can be found in [10].
Stream Reasoning on Environmental Data
Environmental data is becoming more and more available as streams. To be able to
derive new knowledge based on such data, new reasoning techniques are needed. We
have developed Streaming IRIS, a Datalog stream based reasoned, that support
reasoning with rules on streams of Environmental data.
Stakeholders: Stream reasoning is mostly intended to be used by technical users with
a background in knowledge representation and reasoning.
Advances in the project: Streaming IRIS fills a gap that emerged due to the increasing
amount of available live environmental data. Various system are available which are
able to process the data in terms pattern matching and filtering as well as light weight
reasoning, e.g. Complex Event Processing (CEP). But these systems lack in extracting
implicit knowledge of these data streams. While providing the full reasoning
capabilities of Datalog, Streaming IRIS extends Integrated Rule Inference System
(IRIS) with the ability to work in a streaming environment. Complex reasoning tasks
can be performed in a continuous way, by registering queries into the system, taking
the dynamic streamed data and eventual static background knowledge into account.
Core S&T results: There are several results delivered as part of the work performed
on stream reasoning: Streaming IRIS, a Datalog stream based reasoned build on top of
the Datalog IRIS reasoner; WSML2Reasoner extensions (support to integrate the
Streaming IRIS); Sparkwave (a system that supports schema-enhanced pattern
matching over RDF data streams).
Execution of Environmental Models as WS-BPEL Processes
In order to facilitate the development, delivery, and reuse of environmental software
models service orientation has been recently pushed forward by several important
initiatives, such as INSPIRE, GMES, and SEIS, and international standardization
bodies, such as OGC. In the light of those efforts, both geospatial data and geo-
processing units are exposed as Web services, which can be used as building blocks
for the composition of environmental models, in the form of WS-BPEL processes.
However, several challenges arise upon this paradigm shift. Efficient execution and
monitoring of long-running environmental processes that consume and produce large
volumes of data, in the presence of multiple concurrent process instances are among
the prominent issues that one should effectively deal with.
Stakeholders: During the course of the project, ENVISION has particularly targeted
companies and organizations that require an efficient and relatively low-cost setting to
execute environmental WS-BPEL processes, which are long-running, produce and/or
consume voluminous data, and are concurrently accessed by multiple users. IT
entrepreneurs who implement and deliver resource-demanding environmental models
as WS-BPEL processes, but cannot afford expensive servers to host a cluster of BPEL
engines, are first-class citizens of the target group of the ENVISION Adaptive
Execution Infrastructure.
Advances in the project: The ENVISION Adaptive Execution Infrastructure departs
from the various existing solutions for WS-BPEL process execution in numerous
ways. More specifically, the work conducted in this context revolved around the
implementation of the following two innovative features: data-driven adaptation and
decentralized execution.
Core S&T results: Through its three-year course, the ENVISION project delivered
numerous technologies related to the execution of WS-BPEL processes. The most
prominent ones include: A decentralized WS-BPEL engine; A Semantic Context
Space (SCS) engine, a Process Optimizer, and a Semantic data mediation engine.
Further details about the execution approach can be found in [11].
5 Summary
ENVISION targeted the implementation of Web-enabled pluggable user interface
components for the creation of domain-specific Web sites for the environmental
modelling community. This included tools and approaches for the discovery, semantic
annotation, and adaptive composition of workflows representing environmental
process models. With these new mechanisms in place, ENVISION has made a
number of advances beyond state of the art approaches as outlined in this paper,
opening new paths for better access to environmental models, data, and service on the
Web.
The ENVISION platform, together with its components has been made available
under open source licenses.11 A set of demonstrators showing the various aspects of
the platform presented in this paper have been made available online.12 Extensive
materials produced by the project, including technical reports, scientific publications,
11 The ENVISION open source project is available at https://kenai.com/projects/envision/pages/Home.
12 Demonstrator can be accessed via http://www.envision-project.eu/resources/screencasts-demonstrators/.
ontologies, annotate services, user stories, installation guides, results of end-user
workshops are available on the project website.13
Acknowledgment. This work was funded by the EU project ENVISION (FP7-249120).
References
1. D. Roman, S. Schade, A. J. Berre, N. R. Bodsberg, J. Langlois. Model as a Service
(MaaS). Position Paper at Grid Technologies for Geospatial Applications Workshop,
2009. Paper available via http://ifgi.uni-muenster.de/archives/agile/submissions/AGILE-
GridWorkshop-Roman.pdf.
2. ENVISION Deliverable D1.6. Report presenting validation of ENVISION results, 2012.
Available at http://www.envision-project.eu/wp-content/uploads/2013/01/D1.6v1.0.pdf.
3. ENVISION Deliverable D1.8. Development of an operational system by applying the
ENVISION results to the pilot cases, 2012. Available at http://www.envision-
project.eu/wp-content/uploads/2012/10/D1-8_v1-0.pdf.
4. ENVISION Deliverable D2.4. Environmental Semantic Web Portal Version 3 – Final
development and user guide, 2012. Available at http://www.envision-project.eu/wp-
content/uploads/2012/12/D2.4-v1.pdf.
5. ENVISION Deliverable D3.4. MaaS Composition Portal Version 3 – Final development
and user guide, 2012. Available at http://www.envision-project.eu/wp-
content/uploads/2012/12/D3.4-Maas-CompositionPortal.pdf.
6. ENVISION Deliverable 4.4. Cross language and cross domain annotation software.
Available at http://www.envision-project.eu/wp-
content/uploads/2012/02/ENVISION_D4.4_FINAL_for_pdf.pdf.
7. ENVISION Deliverable 4.6. Annotation management report and final components.
Available at http://www.envision-project.eu/wp-content/uploads/2010/01/D4.6_v10.pdf.
8. ENVISION Deliverable D5.4. Integration of the open source catalogue with the semantic
discovery engine – Version 2, 2012, Available at http://www.envision-project.eu/wp-
content/uploads/2012/05/D5.4-Integration-of-the-open-source-catalogue-with-the-
semantic-discovery-engine-Version-II-_M24_v2.0.pdf.
9. ENVISION Deliverable D4.5. Semantic data streams and stream ontologies software,
2012. Available at http://www.envision-project.eu/wp-content/uploads/2012/11/D4-5_v1-
01.pdf.
10. ENVISION Deliverable D4.7. Event processing of observation data, 2012. Available at
http://www.envision-project.eu/wp-content/uploads/2013/02/D4.7-1.0.pdf.
11. ENVISION Deliverable D6.4. ENVISION Adaptive Execution Infrastructure Version 3 –
Final prototype, user guide and evaluation, 2013. Available at http://www.envision-
project.eu/wp-content/uploads/2012/12/D6.4_v1.0.pdf.
13 Accessible via http://www.envision-project.eu/resources/.