|
|
|
4. Bridges: A set of interconnected tools
|
|
. |
|
|
|
 |
|
|
| |
4.1. A prototyping development process
4.1.1. Introducing Bridges main components
4.2. The Bridges Communication System and System Manager (CS)
4.3. Specialised "bridges" to transport models (GTF) and to
GIS and Database managers (GTF_GIS)
4.4. Bridges Core Utilities (NIS): Bridging database structures
attached to transport graphs
4.5. Bridges Expert System: Towards a Decision Support System (ES/DSS)
4.6. Bridges Data Sources Directory: The Bridges Digital Data Sources
Guide (DDG)
|
|
|
|
4.1. A prototyping development process
Bridges research followed a highly interconnected and de-centralised
development process, based on continuous feedback. As a result, all Bridges
elements share fundamental concepts making them fully compatible with each
other.
According to the "Value Stream" approach adopted, each Bridges
Work package was led by one research partner who had to produce a specific
deliverable defined to fulfil a precise user need, as illustrated in Figure
1. Bridges partners were urged by the coordinator to produce "throw
away rapid prototypes" to provide a view of each partner's ideas and
thoughts at an early stage of the research. Based on the analysis of these
prototypes, and the spontaneous cross-fertilisation provided by the
simultaneous viewing of all final deliverables, successive
"evolutionary prototypes" were developed. This process allowed all
partners to "begin with the end in mind", an essential requirement
for highly innovative (so very risky) research projects.
Figure 1 Bridges Technology for building transport policy support systems
In more conventional "Stovepipe" management systems, work packages
respond to the internal logic of the developer rather than that of the final
user (consumer or client), and each one can be divided into different
activities with many developers working with complex interrelations and
input-output dependencies. While this management structure may be well
suited to mature areas and mature partnerships (with little risk of failure
at intermediate input-output steps), this is not the case for immature
technologies (such as software development) and immature partnerships (based
on heterogeneous multinational organisations) as in Bridges. For these cases
the "value stream" approach adopted is much more suitable. This is
illustrated in Table 1.
A Bridges brochure has been produced with a precise and synthetic
description of the final research outcomes. This Final Report constitutes
another step forward from that brochure in that it provides a clearer
picture of what Bridges technology is because it includes the experience of
partners developing real support systems using Bridges. Only the personal
interaction with policy makers and experts, the practical observation of how
much Bridges technology helps to solve their requirements in a productive
manner, has helped the Bridges partners to communicate what Bridges
technology is all about more efficiently.
|
|
|
4.1.1. Introducing Bridges main components
As most Bridges tools remain at the expert level, they have to be used by
system's developers to produce OPEN MULTI-SOFTWARE SUPPORT systems, fully
customised to the needs of end users.
Some Bridges components can be used directly by end users because they
have friendly interfaces (e.g. the DDG, NISystem routines, GTF-GIS
Translator), some tools enjoy friendly interfaces even if they will be used
typically by developers (e.g. the ES/DSS) and others have no friendly
interface as such (e.g. the Communication System has to be configured by
writing ASCII files according to an script language, self-explanatory for
developers with experience but cryptic for most end users). A number of
systems intended for very different end users has already been developed
using Bridges tools for a variety of European and Local Public
Administrations . For each one of these systems (see Section 11.2) a
specific combination of tools has been applied.
In order to understand the technology, the Bridges tools can be
differentiated by the different levels of expertise needed for their use:
- Specialised transport routines and applications, not available or
poorly implemented in the market place, which can be used by non-expert
users (e.g. DDG, some NIS routines),
- Ready-to-use applications and formats to be used by expert users to
link advanced transport models to pre-existing systems (e.g. ES/DSS, GTF
Format)
- Other tools to enable multiple software modules to work together in an
integrated system (e.g. Communication System). They also allow the
organisation and management of a number of personal user workspaces
within the overall system and distributed via a Local Area Network.
These are tools for expert users and system managers.
|
|
|
|
|
|
|
Table 1 Comparison between the efficiency of two different project
management paradigms: |
| |
Value-stream approach (prototyping) |
Stovepipe management approach |
|
Overall administration (e.g. reporting on time) |
Medium |
Good |
|
Quality of intermediate products |
Bad |
Medium |
|
Quality of final outputs |
Good |
Medium |
|
Partnership motivation |
Good |
Bad |
|
Risk of administrative failure |
High |
Low |
|
Risk of scientific failure |
Low |
High |
|
Recommended for: |
Research of innovative products by immature partnerships composed
by knowledgeable and committed people under efficient scientific
leadership. |
Development of products in well established areas by mature
partnerships composed by not necessarily knowledgably nor committed
groups, under disciplined administrative co-ordination. |
|
Recommended for: |
In the Information and Communication Technologies field (ICT),
where prototypes are cheap to produce and disseminate and meaningful
feedback is feasible and the scientific environment rapidly changing. |
The manufacturing field, where prototypes are expensive, difficult
to duplicate and improve and may even be misleading. |
|
|
|
It is a goal of Bridges developers to make the use of Bridges tools as easy
as possible for all end users, expert users and system developers. As
Bridges developers gain more experience in developing systems with Bridges
and the software industry continues to provide more advanced software and
hardware systems, it is expected that user friendliness will improve.
Bridges components are described below in more detail.
|
|
|
4.2. The Bridges Communication System and System Manager (CS)
The Bridges Communication System harnesses OLE/COM technology to
integrate stand alone applications. The Bridges Communication System is
based on managing flows of command messages between stand alone
applications. Bridges Communication System was designed to work in an
Intranet but its decentralised architecture allows the addition of a new
bridge to the Internet.
The Bridges Intranet Communication System facilitates simultaneous
on-line dialogue with commercial applications . Transport models, even those
supported by the few available OLE/COM compatible modelling packages, pose
additional problems for "bridging" such as when they are based on
independent non-interactive modules or they use incompatible data models.
This requires more specific tools for example common data formats, such as
Bridges/GTF.
In addition, "bridges" which allow different applications to
share the same database were developed. However, the simple import and
export of databases is not the main problem to be solved: The main problem
is to define a common data model and format protocol for the exchange of
complex transport oriented databases, which can grow into a standard for
transport modellers.
The Bridges System Manager helps to personalise an open systems
architecture to the specific needs of each user connected to it. Because of
the modular and scalable character of systems developed by Bridges
technology, each user has their own "personal Work space", which
may contain different applications and different databases.
"Friendliness" is therefore achieved by customising the
architecture of the system to the needs of each specific user and problem.
The Bridges System Manager allows an unlimited number of users to access the
system, each with multiple customised workspaces. The Bridges Manager takes
care of user passwords, work space maintenance and confidentiality issues.
Finally, the Bridges Common Interface allows the development of fully
personalised menus and options, even beyond the design constraints of
Microsoft Windows.
The Bridges Communication System, Management and User interface utilities
have been programmed in Borland C++. Looking ahead, major improvements could
be achieved by achieving Bridges communication through the Internet.
Needless to say, this achievement would be a major step, placing Bridges
technology in the forefront of software innovation.
|
|
|
4.3. Specialised "bridges" to transport models (GTF) and to
GIS and Database managers (GTF_GIS)
A GTF (Generalised Transport Format) has been defined as a format to
store and exchange any transport oriented database. Data exchange between
transport models requires that models share a common data model. The GTF
data model was defined to be sufficiently open to cover all strategic models
at European scale. Bridges' aim is to propose GTF as standard format to be
adopted by EC and international institutions.
The GTF data model is able to deal with the complex graph structures and
topologies used by transport models (including demand, supply and specific
modelling aspects). Following the IDEFIX approach to data modelling, the
following entities were defined: node, link, interchange, route service,
zone, main mode, persons/goods, carrier, modal stage, and flow/movement. The
GTF data model includes the categorisation of relationships. In addition,
the adopted GTF format description is public, compatible with international
standards (UN/GESMES) and complementary to other database and GIS standard
formats. It includes both topological information related to transport
entities and the statistical data attached to them.
GTF has been tested with MKmetric VIA and MEAP MEPLAN modelling tools,
and it has been extended to cover GIS in cooperation with the GTF_GIS work.
A GTF Translator to and from VIA Model formats to GTF has been programmed,
with a separate version in Java to allow Internet accessibility.
The GTF_GIS data model is based on adding transport topology to
conventional GIS structures, according to the GTF definition (e.g.
routes-route segments, stops. The format description is very intuitive with
simple ASCII files.
GTF_GIS entities are compatible to GTF but can enrich public transport
entities: links, nodes, turns, routes-route segments and segments, stops,
changes, terminals, zones, fictive links, matrices-flows, vessels,
units-persons/goods and milestones. Topological information related to
entity relationships (e.g. allowed vessels and units on links, routes and
flow matrices) is also included. Names of data tables containing statistical
data and metadata information are listed within GTF_GIS but the actual data
is not included.
Arc/Info, one of the most advanced GIS, is unable to handle the complex
network topologies required for transport modelling (e.g. defining
centroid's connectors, public transport routes and services etc.). Because
of the high GIS productivity of Arc/Info tool, and its widespread use within
European institutions, a specific "bridge" from Arc/Info's
encrypted data format to GTF has been developed. The translator converts the
Arc/Info GIS format into GTF_GIS by adding transport topology. A GTF_GIS-Arc/Info
translator has been programmed based on MapObjects libraries, the only way
to overcome the Arc/Info encrypted data format. The unavoidable use of such
libraries makes the translator not totally royalties free.
Translators to standard CAD (DGN, DXF), GIS and Desktop mapping (MIF, SHP,
GEO) formats have also been programmed. Given the limited GIS capabilities
of these tools, they are unable to support advanced transport database
structures. The interest of bridging these formats lies in the import of
graphs developed on CAD, simple databases on GIS and the export of results
to be displayed graphically. These translators have been programmed in
Borland C++. Translators to DBS applications have been programmed based on
ODBC drivers and Data Access Objects (DAO).
|
|
|
4.4. Bridges Core Utilities (NIS): Bridging database structures
attached to transport graphs
Bridges Core Utilities (Network Information System, NISystem utilities)
have been developed as external stand alone applications to be linked to any
system by the Bridges Communication System, just like any other stand alone
application. This guarantees that systems developed by Bridges are fully
scalable and independent even from Bridges own Core Utilities. Any Core
utility can be removed and/or substituted by others when needed.
The paramount goal of NISystem Core Utilities is to complement transport
modelling, GIS and DBS applications with missing utilities, particularly in
the context of the ETIS reference requirements. NISystem provide:
§ Specialised routines for harmonising heterogeneous transport oriented
databases and graphs. Examples are "Graph matching" utilities to
"bridge" databases attached to graphs with different link
segmentations covering the same zone, "Graph checking" utilities
to check topological properties on CAD files, "Graph editing"
utilities to modify graphs and add topology to CAD files etc.).
- An original GIS binary format (MGS), based on GTF and GTF_GIS data
models has been created to increase the productivity of CAD-GIS core
utilities and also to overcome possible confidentiality concerns for
data dissemination.
- Graphic management routines for CAD, Desktop mapping and GIS have been
programmed and included in the system as "core utilities",
mostly to support "bridges" which required more than simple
data format exchange (e.g. the above-mentioned Graph matching).
Moreover, core utilities facilitate maximum flexibility in customising
user interfaces and also make Bridges technology not completely
dependent on commercial applications but complementary to them.
- A complete database management application has been programmed in
Visual Basic 5.0, allowing an optimum degree of user interaction with
large and complex databases as well as enjoying several management
utilities. Bridges database manager uses Microsoft MDB as default data
format but is open to other formats through standard ODBC drivers and
DAO.
- Operational research and statistical algorithms for analysing
transport databases have been plugged into the system to provide Bridges
users with specialised easy-to-access tools for checking, exploring and
exploiting databases.
Core Utilities have been programmed in Borland C++ (CAD, GIS and
Operational research algorithms) and Visual Basic 5.0 (Database management,
EXCEL macro-language).
NISystem utilities have been encapsulated in a number of modules (e.g.
the DATABASE module contains all Visual Basic DAO-based functions to manage
databases, as well as Microsoft Graph; VIEW contains all C++ routines to
provide GIS visualisation capabilities; GRAPH contains all C++ routines
providing edition, quality control and analysis of transport networks as
graphs etc.). Each NIS module is an executable file which enjoys its own
user-interface. It can be activated in the user-interface of the system
similarly to any stand-alone application, simply by placing a button linked
to the execution of the module.
|
|
|
4.5. Bridges Expert System: Towards a Decision Support System (ES/DSS)
ES/DDS has been developed and integrated into Bridges' toolbox to provide
a tool to build intelligent translators between end users questions and
sophisticated transport model outputs. The Bridges Expert System helps users
to establish legitimate queries to models and interpret the model results.
The ES has been tested by building the expert translators to MKmetric VIA,
MEAP MEPLAN, and with the Spata Airport environmental impact model. ES/DSS
has been defined and programmed with a medium and long-term ambition: the
more sophisticated and numerous are the models "bridged" to ETIS,
the more useful the ES/DSS tool will become in providing the expert
interface.
ES/DSS architecture is composed of a user-interface (Object Oriented
Interface, OOI) and the Expert System (ES). The OOI comprises a set of
menus, forms, dialogue boxes etc. which are either pre-set or generated at
run-time to facilitate user dialogues. The design part of OOI guides the
expert user though the standard options of calling up a DSS module (data
dictionary, template parser, expert system etc.) or a utility (Visdata,
DSSview etc.).
The main task of the Expert System (ES) is to analyse user queries,
decompose them into queries to be passed on to other modules, and to combine
results into a meaningful form for the user to understand. In doing so, the
ES may apply default values, assumptions and rules to fill in any missing
data required for running other modules. ES comprises a Template parser, a
Query processor, an ES Shell and a Multicriteria evaluator.
ES/DSS has been programmed in MS Visual Basic 5.0, Visual C++ 5.0, Access
7.0, ESRI's MapObjects 1.2 and Amzi! Prolog 4.0, according to OLE/COM. DSS
is an OLE container application, therefore easily "bridged" to the
rest of the components of a multi-software system by using the Bridges
Communication System.
|
|
|
4.6. Bridges Data Sources Directory: The Bridges Digital Data Sources
Guide (DDG)
DDG has being developed as a stand alone application which contains
European data sources and database descriptions (based on ETIS definitions).
DDG follows a similar approach to the "Directory of Transportation Data
Sources" published each year by the US Department of Transportation
(Bureau of Transport Statistics).
The rational for including DDG in Bridges Toolbox was the assumption that
ETIS (as with any other support system) would be limited to a minimum core
of data useful for most studies, with links to additional data source
providers as needed. On the other hand it was important to develop a public
directory of transport information sources in Europe in the framework of
Bridges to demonstrate the value of such a tool in the frame of policy
support systems.
Currently, DDG is a data sources directory which contains updated
information related to the main European Data providers and Data products,
as well as available transport models and software products. Initially, DDG
was focused on information databases and sources available in electronic
format (CD-ROM) and sources accessible throughout the Internet but it was
subsequently enriched by including sources available in conventional
paper-form. It was also extended to cover available models/modellers and the
major software products/providers.
DDG contains:
- 558 European database products, providing information such as
availability, cost, periodicity, legal reference, data etc.,
- 732 European information providers with contact information and
Website summaries,
- 85 modelling products with information based on previous studies
undertaken by DGVII,
- 76 modelling providers with contact information,
- 71 software products and
- 74 software providers have also been included.
DDG has a user friendly interface to define queries and achieve
customised answers, as well as to establish direct links to data sources
through their Internet Websites. It also allows users to update the
directory themselves; therefore, it includes a precise codification of
fields and checks to validate any new information introduced. DDG has been
programmed in MS Visual Basic 5.0, and information has been obtained from a
variety of sources.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
info@mcrit.com |