3 Bridges Vision
3.2. The software components of multi-software support systems
3.3. Specific Bridges research areas
The aim of Bridges is to solve as far as possible the lack of compatibility and openness of advanced modelling tools and also between models and other software tools (e.g. those supporting database structures and facilitating interactive viewing of results). Instead of a new transport modelling tool including GIS and database management, the outcome of the Bridges research is a software technology (a "toolbox") "to bridge" models, GIS and database managers in productive and user friendly multi-software support systems. Therefore, Bridges' major goal was to enhance the productivity of information and modelling tools by "bridging" them, including the building of multi-software platforms.
By closing the "software gap" between transport models and other software tools, Bridges technology can help to reduce the more fundamental gap between advanced modelling tools and end users (e.g. experts, decision makers). Systems developed using Bridges tools can put advanced analytic tools in hands of people who are not software specialists; highly customised interfaces can be programmed to give users friendly interaction with advanced tools.
Bridges research is a Windows NT compatible technology, but its Communication System was designed to allow future Internet compatibility and some promising avenues towards such compatibility have already been explored. In particular, a Bridges Server, able to receive remote messages from Internet browsers and interact with Bridges Intranet Communication System is under development as a follow up to the research. The development of such a Server may require modification of some elements of the whole Bridges technology (e.g. the Communication System and NISystem), but the fundamental Intranet design concepts presented in this Final Report will remain as they are.
The fact that end users are demanding more and more advanced software is moving the software industry from closed and proprietary tools towards open systems and public formats. No single application can optimally meet all user requirements by itself and therefore open multi-software platforms, making computer aided design (CAD), database managers (DBM), geographic information systems (GIS), statistical applications and other modelling tools work together, are needed to carry out more and more specialised work with higher productivity. Current applications with a limited degree of openness will tend to be replaced by specialised intercommunicating applications and software components, stand alone routines and controls open to data sharing with other applications and to being driven externally by fully customised user interfaces. Needless to say, this emerging trend in the software industry is providing better software tools to develop open multi-software support systems. However this if true only to a certain extent, since most research efforts are nowadays concentrated in key business areas such as electronic commerce or personal communication through Internet, with much less effort to fulfil the specific needs of scientific-oriented systems (those requiring the use of advanced computation tools) or policy-support systems (those making the advanced tools accessible to decision makers).
A support system, or more generally, a Decision Support System (DSS), is a software system under the control of one or many decision makers that assists in decision making. It includes information management, modelling and decision support. Bridges "technology" is a set of communication tools, data models and format protocols needed to develop transport oriented DSSs as open multi-software systems.
The fact that transport models and transport databases are relatively complex and non-standard software tools, with compatibility problems in working with more advanced GIS applications makes the integration of advanced transport models into productive and user friendly support systems an extremely difficult task, almost always requiring specific solutions for each model and each database. These technical problems, as well as other institutional and organisational factors, explains why very few transport administrations have developed such policy support systems despite their clear advantages given the growing complexity of the transport decision making process, especially at European level. While there are commercially available technologies to develop support systems for business management (now linked to electronic commerce) and almost all major companies already have or are presently developing such systems both for management and strategic planning, nevertheless in the policy field, especially in the transport planning sector, specialised technologies are not available and there are only ad hoc, often sub-optimal solutions, unsuitable for general application.
A Software SYSTEM for end users is built by implementing links ("bridges") between MULTIPLE independent software applications. Links ("bridges") between these multiple applications supporting databases and model formulation are needed to make all of them to work together in an integrated and as interactive as possible a manner, using common data formats where this is feasible.
This multi-software system approach of Bridges is becoming ever more necessary given the increasing requirement to deal with more complex questions, which no single database, model or software application can satisfy completely. The first concern of a software technology converting multiple stand-alone applications into an integrated system is, then, improving the productivity (e.g. automating data import and export between applications, reducing running times by interconnecting routines etc.). For this reason, many modellers have already developed their individual ad-hoc solutions to integrating their models with other tools.
A System is OPEN when any additional module or independent external application can be included easily and it is possible for a substitute to be provided for any internal one without changing the overall structure of the system. Openness is a key characteristic in facilitating the evolution and sustainability of the system. Openness requires the existence of standardised formats and communication protocols supporting the links between system modules. This leads to a second reason for developing a Bridges technology, that is the need to increase the capacity for continuous improvement and modernisation of information and modelling systems. There is empiric evidence suggesting that rigid systems and obsolete software solutions are key problems blocking modelling improvement.
An information and modelling system can be considered a SUPPORT system when it is adapted to the solution of end user problems with a "knowledge" interface able to translate information and modelling outputs into meaningful answers for the end user and, at the same time, translate legitimate end user questions into suitable inputs for models. For advanced models, such an intelligent interface has to be developed as an Expert System; for simplified models, access to policy indicators and database management, it is sufficient to understand end users' analytic and decision processes and design user interfaces adapted to these requirements. While there are many software tools to design and develop such user interfaces, two crucial elements are missing: advanced analytic tools working at the end-user level (to be used also by expert users directly as in the design process) and tools specialised in defining and developing Expert Systems for advanced transport models in a systematic and productive manner.
Finally, an Open Multi-Software Support System can be considered USER
FRIENDLY when all modules are driven by fully personalised user interfaces
and the whole system is able to engage in an INTERACTIVE dialogue with the
user, whenever this is feasible. Within a user friendly system, there may be
modules either with their own user friendly interfaces or with no user
interface as such. These latter modules remain internal to the system,
accessible only to expert users and/or the system developers. There are many
multimedia, virtual reality and 3D, desktop mapping and graphic design
commercial applications for building such user interfaces. Bridges adds to
these standard tools advanced transport modules (NIS), some of which can be
run by end users themselves. Systems developed using Bridges can be
simultaneously driven by multiple user interfaces which govern a number of
applications and databases (a user workspace is this group of applications
and databases with a user friendly interface).
The final, and probably most important reason for developing Bridges
technology is to narrow the gap between end users and advanced assessment
tools, converting the latter to actual "learning and communication
tools" for the mutual benefit of end users and scientists (modellers,
system developers).
There are no commercial software tools for all the "bridges" needed to build up efficient transport-oriented open multi-software systems like the ones envisaged here, and even the available Windows-compatible tools to link independent applications are sub optimal for these specific purposes. Therefore, Bridges research has developed an entirely new set of tools to fill the most important gaps.
As discussed in the previous section, advanced transport information and modelling systems ideally require the integration of CAD, GIS, DBM, modelling and Desktop Mapping capabilities, knitted together behind a user friendly, fully personalised, interface:
- CAD applications (such as AutoCad or Microstation) facilitate the
digitisation of physical networks and reference environmental and
geographic features. With additional GIS oriented routines (programmed
in AutoLisp or UCM-DML, or more generally in Visual Basic), simple
graphic databases can be supported by CAD applications as well as basic
graph utilities (such as checking the consistency between nodes, links
and polygons)
- More sophisticated GIS applications (such as ArcInfo, Small World, MGE,
Geomedia or Geographics) have specialised network modules with many
graph oriented routines (albeit incomplete and in many cases
sub-optimal). They provide complex geographical routines for projection
change and adjustment, raster analysis and even remote sensing,
superimposing both vectorial and raster information. Alphanumeric
database management and interactivity with external model algorithms and
specialised database managers (DBM) are among the main factors arguing
for their general use in transport database management. However, because
they are unable to handle advanced transport topologies, they are often
used simply to view the final modelling results.
- Desktop Mapping applications (such as Mapinfo, ArcView) enjoy basic
GIS options and provide easy graphic display of data and results. They
tend to be open to external applications, such as Microsoft Excel or
Access, where data and basic calculations can be more conveniently
handled, or to complementary GIS applications.
- DBS (such as Access, Dbase, Oracle, Informix etc.) are indispensable
for the storage and management of increasingly large data sets coming
from many different sources. ODBC drivers and SQL language provide
interconnectivity options between database managers. Recent Oracle
developments allow the expert user to send "spatial queries",
so that Oracle already supports advanced GIS capabilities. Other expert
solutions for database management could be, for instance, to program
specific ad-hoc routines in database management-oriented languages, such
as Clipper, using DBF as data format; since DBF can be read efficiently
by advanced statistical tools (e.g. SPSS) and programming languages such
as Borland C++, a complete, fully personalised and highly productive
software architecture for modelling can be developed without any
specific database manager application.
- Spreadsheets, statistical and mathematical packages provide useful
routines and programming languages for building mathematical models.
Models developed with these tools often have good links to database
managers and even provide internal graphic applications (including 3D
views, dynamic mapping etc.).
- Transport Modelling tools (e.g. SATURN, TransCad, EMME2, Transplan,
Trips etc.) internalise a minimum number of CAD and GIS facilities and
are becoming more and more open to CAD, GIS and DBS applications.
Because of the inherent rigidity of modelling formulations, they are
less likely to be used in research. At European scale also, the
heterogeneous and incomplete nature of the available databases often
requires ad-hoc formulations).
- For the design and development of multimedia user interfaces there are
an increasingly number of applications, currently providing Intranet and
Internet viewing . On the other hand, relatively advanced programming
languages, such as Microsoft Visual Basic or Borland Delphi, are very
efficient, integrating text editors (Word etc.), spreadsheets (EXCEL
etc.) and database managers (ACCESS, SQL Server etc.) under fully
personalised interfaces.
- Since the late 70s there has been growing activity aimed at integrating all these disparate software elements required to meet the specific decisional needs of particular users to convert software systems into functional "decision support systems". To move from information and modelling systems to "support" systems, it is not enough to develop customised user interfaces; first a deep understanding of the decision process is needed to make the interface intelligent so it can translate users' questions into legitimate inputs to predictive models and return information and modelling results as meaningful answers to users. Artificial Intelligence systems, learning-systems and other kinds of tool have been developed over recent decades, often using programming languages such as Prolog instead of commercial packages.
All in all, ongoing software research and development is moving towards greater openness, scalability and user friendliness. Commercial transport models enjoying GIS options are emerging (such as TransCad by Caliper) and GIS applications well linked to sophisticated DBS applications with relatively good data warehouse capabilities (such as Intergraph Geomedia). Database managers (e.g. through the Internet JDBC extension), are in the process of achieving productive data format exchange and, to some extent, intercommunication through the Internet. Electronic commerce through the Internet is inducing software companies to produce new software technologies capable of developing communication systems which can, to some extent, link remote stand alone applications managing different databases.
The vision of Bridges is to bring software innovation to the transport
planning field. In particular, bearing in mind the ETIS requirements, the
objective is to adapt and/or develop efficient software solutions to build
open information, modelling and support systems capable of becoming
efficient decision support systems in hands of experts and policy makers, as
close a possible to the ideal of "maximum capabilities with minimum
complexity".
- "Data links" (to access transport and GIS database formats)
- "Command links" (to drive stand alone applications),
integrated into a "Communication System" built on Windows
95'NT and managed by a System Manager on Intranets.
- Tools to build "Expert systems" for advanced transport
models, helping to formulate legitimate queries to them and translate
their outputs into meaningful policy questions.
- "Core utilities", currently missing or sub-optimal for
transport planning purposes in the software industry, for transport
applications, in standard commercial software tools. Basically, they are
GIS routines linked to transport data management and modelling. These
components make Bridges a very productive software technology for
building transport oriented multi-software support systems. Core
utilities are encapsulated into stand-alone components that are linked
to any system similarly to commercial applications.
- Customised "user friendly and intelligent interfaces",
helping multiple users to interact with the system according to
personalised sub-systems or user workspaces.
- A Directory of European Transport Data Sources
The succeeding chapters will present each Bridges tool in depth, as well as the research process leading to them.