Web Technology for Forestry and Environmental Information and Decision Support Systems.


Nov 5, 2013 (4 years and 7 months ago)



Web Technology for Forestry and Environmental
Information and Decision Support Systems.


Keith Rennolls


School of Computing and Mathematical Sciences,

University of Greenwich,

London, UK


Comments, corrections or suggestions on this draft document should be sent to t
he author at
K.Rennolls@gre.ac.uk .





1. Introduction



2.2 EFIS

2.3 GFIS

2.4 FEIDSS: Ultimate aim: A European FIS
based decision support system.

2.4.1 The main potential user

2.4.2 Requirements of a pan
European planning tool/DSS. Con
tributed data and information resources. Varying National Standards:



Implication of SCTK on


NEFIS Data Warehouse

The Needs of pan
European Forest and
Environmental Researchers Constructed data structures

A FEIDSS Tools Grid

3. An overview of the main Corporate Web Software Players

4. Web Enabling Technologies

4.1 Metadata

4.1.1 Data and metadata hierarchy

4.1.2 Dublin Core


.3 Web Services : SOAP, WSDL &

4.3.1 SOAP

4.3.2 WSDL

4.3.3 UDDI (Universal Description, Discovery, and Integration)

Engineering of Web Services

Comparison of middleware technologies in relation to a FEIDSS

4.3.6 Federated W
eb Services and Grids Federated Web Services Grids through Web Services

4.3.7 Language Support for Web Services

4.4 The
Semantic Web
: Ontological Languages

4.4.1 RDF

4.4.2 DAML+OIL


4.4.4 The Protégé DAML+OIL
plug in.

4.4.5 DAML
S and OIL
S spec and review:

4.4.6 OWL: Web Ontology Language

4.5 Other (secondary) web technologies

4.5.1 Topic Maps



4.5.3 Browsers

4.5.4 ebXML

4.6 Grids

4.7 Portals

5. Conclusions in relation to WS, the

Semantic Web


5.1 WS and the Catalogue function in FEIDSS

5.2 WS and

functionality of FEIDSS

5.3 Good Software Engineering Management Practice in relation to FEIDSS/GFIS





he W3C Persons Example, using Protégé with the DAML+OIL Plugin.


1. Introduction

This document is meant to be a repository of technical information related to the title of this
document. Some of the original technical material was brought togethe
r for a Technology
review within the EU F5 NEFIS Project. That review has been largely stripped of material
which was NEFIS/EFIS specific so that there is not a particular relationship of this document
with the NEFIS project or the EFIS product. In refer
ring to a generic Forest and
Environmental Information and Decision Support System, the acronym FEIDSS will be used.

There are a number of reasons for constructing a Web
Technology/FEIDSS review as an
information resource.

First, a Technology Review is j
ust a snapshot of the picture at one time. With rapidly
changing web technology, the review will rapidly become dated. However, as one finds new
strands of information and technique becoming available, there is often no place to put
references to such new

material. The last core technology review is as good a context to but
such references to new material as any. In this way the technology review becomes a rolling
technology review.

Second, there are many groups in the world, all doing essentially the s
ame thing, for the same
reason, in developing information systems relating to Forestry and the Environment. This is
clearly a sub
optimal approach on the global scale, though locally, it may be all that has been
possible. In the longer term it is to be h
oped that such parallel groups could come to know of
each other’s activities, and to share some of the best
practice technical solutions to the
operational problems that arise. It is even conceivable that such parallel systems might at
some stage merge in
to a large and more powerful information system. However, for this to
happen technology has to be common, or there has to be a suitable enveloping technology.

These considerations have partly been responsible for the organisation of a DEXA’04
Workshop on

Forestry and Environmental Information and Decision Support Systems
(FEIDSS). Those involved in FEIDSS projects need to look to each other in cooperation and
collaboration. They need to look to the future, and not just focus on the “trivialities” of the
present, funding issues included. This document is prepared as a background document for
the DEXA’04 FEIDSS workshop and will possibly eventually encompass some of the finding
and conclusions of that workshop.


is an ongoing EU FP5 Acco
mpanying Measure Project, which is primarily concerned
with the extension of the utility and design envelop of the EFIS demonstrator developed in the
previous EFIS project, which was initiated by the EFICS Regulation of 1989.



EFICS, the Europ
ean Forest Information and Communication System, (Council
Regulation(EEC) No. 1615/89) had as its objective to “


data concerning the forestry sector and its development …to facilitate the
implementation of

taken at national and regional level concerning the forestry
sector …taking into account
existing information systems

The EFI EFICS Study, 1997, was commissioned “
to analyse in detail the statistical sources of
forest resources in the EU…and to draw

up proposals for obtaining data which is
compatible and comparable
, so as to be able to establish a
reliable and consistent database

European level

The final report documents extensive considerations, research results,
and recommendation

We have underlined various words above to highlight the major concepts forming part of
EFICS thinking. Note in particular the clear realisation of the incompatibility of national
forest inventory data resources, as they stand, without either harmonis
ation or standardisation.

In particular, the Study:

(i) did an extensive user requirements analysis, based on a large questionnaire survey,

(ii) did detailed comparative analysis of the differing measurement standards of forest
inventory throughout Eur
ope, demonstrating, for example, that estimates of forest area, and
timber volume could vary by in excess of 20% purely because of the differing standards

(iii) analysed the potential for harmonisation, or alternatively, the use of conversion form
to achieve standardisation; the harmonisation option was costed as requiring between 40 and
278 person
years of work (i.e. up to 9 M Euro.).

(iv) evaluated various methods of updating different national forest inventories to a common
base date, for c
omparative purposes.

(v) considered various combinations of database location (distributed
not, access options, and the kinds of data
analysis and decision
services provided to users. In this respect, we note that the c
ombination: (unharmonised
national inventory database) + (standardised centralised database on selected important
attributes) + (interactive data
processing by the user) was NOT a combination considered.

All proposals required considerable expenditure to
achieve the EFICS objectives.

2.2 EFIS : Project 17186

In the Executive Summary of the report on EFIS, published by the EU JRC, (2002) the
following are quoted as the sources that motivated the EFIS project:

erence on Environment and Development in 1992 in Agenda 21,
Chapter 40.

EU regulation (EEC) No. 1615/89 of the year 1989 (EFICS).

The Intergovernmental Panel on Forests in 1997 (CSD, 1997)


EC implemented study in 1997 on European Forestry Inventory and Sur
vey Systems, (EC,
1997): as described in section 2.1.

In the last of these, the recommendation was that the system to be developed should: A1. Rely
on existing Forestry Inventory data; B1. All data to reside in national databases only; C3.
Interactive ana
lysis by users through the Internet.

Operationally, there are two main functional aims to the EFIS demonstrator, (and indeed to
any FEIDSS):

1. A searchable
metadata catalogue
, built at a “central” location: the input data required is
the metadata for

the national forest database sources.

Dublin Core (extended to include a ‘quality’ element) was chosen as the metadata format for
information sources, to be compatible with the GFIS (IUFRO) and the FAO (WAICENT

2. A
distributed database access
analysis functionality

(briefly the


Visual ToolKit

at the heart of the EFIS demonstrator is a
data processing and
visualization system
. This system was meant to operate as a central server
side facility on the
ted) national data sources accessed through the internet by the user. A client
architecture was chosen with the analytical engine focusing primarily on geographic and
graphical visualization of data. A decision support facility,
is also a
vailable which
makes use of the Visual Toolkit.

The overall structure of A FEIDSS is shown in Figure 1. (extracted from the A FEIDSS report
(Fig. 2)).

Figure 1.

The A FEIDSS structure.

In terms of software/database/web technology, after
an extensive review, the following
choices made were:


Software environment

Java 2, JDK 1.3

Development tool

Forte for Java

Server side web system

Java Server Pages

Servlet processor

Apache Tomcat


Dublin Core + A FEIDSS qualificatio

Database interoperability

Java FISE, Z39.50

DBMS connectivity


Data transport

MS Access / mySQL


XML (ForestML)

The Visual Toolkit (including CommonGIS and Descartes) is in fact a generic GIS/Graphics
facility, which

was developed under EU research budgets other than A FEIDSS. It is possible
the visualisation facilities of the Visual toolkit could be enhanced through the use of open
source visualisation facilities. MayaVi (http://mayavi.sourceforge.net/) is an open
ource VR
data visualiser with VRML input/output facilities and can be scripted/modified via
Python. The VTK visualisation library (http://public.kitware.com/VTK/) is used in MayaVi,
but is also of independent interest.

Currently, EFIS had no cat
alogue functionality, no
functionality and makes no direct use
of remotely sensed imagery, and there are no tools for dealing with such imagery.

2.3 GFIS

GFIS is the IUFRO forest information resource cataloguing facility. A brief summary of the
technology choices made by the GFIS developers is made on the basis of Binh & Ibrahim

GFIS makes use of Java, UML for application design, DC metadata, PHP, mySQL, the
Z39.50 protocol, and XML (with a DTD). MetaCube
X is mentioned in relation to

the GFIS
advanced search facility. The DAML and OIL ontology specification languages, XML and
RDF extensions, are mentioned in Binh and Ibrahim (2003), in relation to future semantic
web developments, but the current GFIS prototype does not make use of th
ese. Topic Maps,
also an XML extension, have also been mentioned elsewhere in relation to GFIS, but it is not
clear if they have been used. Binh and Ibrahim (2003) state that the XML DTD for GFIS
initiates a new dialect of XML, ForestML. A FEIDSS also h
as its own XML specification.

GFIS, being a searchable web
based catalogue of forest information objects, with DC being
the basis of the “indexing”, provides the usefulness of a catalogue with hyperlinks, but does
not allow the user to access raw data re
sources. However, it is noted that Bingh (2003), in

the IUFRO Inter
operability Conference proceedings, available from:


he does cla
im that providing direct raw data access is one of the future aims of GIFS.

2.4 NEFIS: Ultimate aims: A European FIS
based decision support system.


The NEFIS targets, stated in a recent EFIS publication linked to a demonstration copy of

1. To

provide efficient information resource discovery, based on catalogued metadata, with A
FEIDSS being aligned with both GFIS and INSPIRE.

2. To

substantial, (influential) and useful statistical and geo
referenced forestry

3. To make use o
f distributed database technology for distributed data resources.

4. To provide information management and processing tools (statistical and geo
which address the requirements of the EFIS user groups.

5. To do 1
4 in as cost
effective and
as user friendly a way as is possible.

The catalogue software system should be designed in such a way to allow GFIS nodes, such
as EFIS, to develop extra facilities and functionality for their nodes, as might consider
necessary. A useful concept here mig
ht be that of “wrapping” and use of the “Adapter
Design Pattern”. This would allow nodes with resources that don’t conform to the standard
to be wrapped by a thin layer of software so that they present a standard face to the outside

It should be r
ecognised, in relation to 3., that NEFIS is

amongst international Forest
Information Systems, since it is the only project world
wide which promises to make detailed
European national forestry data available to users. All other internationally
based FIS
projects are essentially catalogues/indexes to forest information resources, but do not make
such resources directly available to the user.

Target 4 is crucial, since it is the satisfaction of user requirements that is the ultimate
and all the other aspects are really means to that end. It is therefore crucial to
consider the different types of user group, and their varying requirements. In doing so, it has
to be accepted that “service costs”, and full satisfaction of all possible
user group
requirements would likely be very expensive to implement.


The main potential user

The user
group structure defined in the Executive Summary of the EFIS report restricted
itself only to those who were primarily interested in ti
mber information, those who had
general interest in forests, and those who had a secondary interest in forests.

The GFIS list of user groups is rather wider and longer than that stated by EFIS, (but seems
currently not to be available on the GFIS web site


The EFICS Study (1997) reported the results of and analysis of an extensive user

requirements questionnaire survey. This survey was however, limited to governmental and
national level forestry organisations.

The following list indicates mo
st of the EU
groups that can be


1. EU pan
European land
use planners.

2. National and Regional (sub
national) land
use planners and forest managers.

3. Trans
national environmentalists and conservationists.

4. Forest and Environ
ment researchers: data mining and model building.

5. Local forest managers (sub

6. Local Authority land
use and landscape planners (sub

7. Educational users: Schools, Colleges and Universities.

8. Individual personal users
; business or recreation.

The user
groups have been put in a priority order. The EU funding of EFIS was under the
EFICS initiative, and a strong and central concern of the EU Commission has been, and is, to
have available consistent and reliable pan
opean information and tools for centralised
planning activities. The scope of these planning activities cover many the specific interests
of those user groups lower down the list.

The second user
group is the National and Regional planners of states in t
he EU. Such
planners already operate in each of their own states, and use the information and tools that
they have developed nationally. It was hoped that EFIS might enhance such information
management and processing at the national level as well as at t
he pan
European level.

User groups 3, 4 and 5 probably represent an equal level of ranking. With the importance of
conserving biodiversity, and managing forests sustainably in the context of climate change,
groups 3 and 4 are of considerable importance i
n this effort. Group 5 is meant to include the
timber producing industries, which operate largely within national boundaries. Of course,
timber production has long been the primary product of forestry, but current conditions mean
that timber production ha
s to be pursued in the context of the concerns of Groups 3 and 4.
Also, the timber industries,
in general, are part of a commercial enterprise and it might be
argued that they will have already made the investment in information and tool development
they consider they need in their commercial activities. It is possible that information
collected at a pan
European or global scale will not be of much commercial/operational use
to such timber industries, at least in the early stages of the development o
f NEFIS. A request
to NEFIS for a local timber volume inventory, and a map or the timber volume distribution to
aid in extraction, without the company having to do a field survey, must be regarded as at
least “a considerable challenge” to NEFIS, at least
in the early developmental stages. The
potential of NEFIS eventually being able to satisfy such requests, and if this should be an
aim, needs to be considered and discussed.

Similarly, for Group 6., the resolution of the information available in a develo
ping NEFIS is
likely to be less than is already available locally: the detail of the likely interesting questions
from such users will be beyond the information resolution of a developing NEFIS.

User groups 7 and 8 are not involved in the management and f
unding processes of forestry,
and though satisfaction of their interests and concerns in forestry are important from a social
and political perspectives, it is suggested (for debate) that they are should be placed at the
lowest priority user groups. Sati
sfying the requirements of those higher on the priority list


will certainly satisfy any of the requirements of the lower priority groups. Also, the social
and political influences wil1 find themselves entering directly as factors for consideration by

groups 1, 2 and 3.

2.4.2 Requirements of a pan
European planning tool/DSS. Contributed data and information resources.

Tim Richards, in an internal EFIS/NEFIS report on the XML structure of the EFIS, described
three levels of metadata; Leve
ls 1, 2 and 3.

Level 1 metadata is metadata that describes the features of raw data as stored in national or
European databases and data archives. Level 1 metadata essentially captures the data
of the raw
data database. We refer to such a databa
se as a Level 1 database.

functionality of EFIS uses such Level 1 databases, and all the NEFIS data contributors are
essentially Level 1 database contributors.

Level 2 metadata is metadata that describes some of the important general features and
racteristics of an information
object. DC is an example of such Level 2 metadata, and
when such data is stored in a searchable database, (which we call a Level 2 database), then
we have a catalogue of information objects, as in GFIS, and the catalogue fun
component of EFIS. Varying National Standards

Clearly, User Group 1 has a requirement from the Forest Information System for information
that will help in their tasks of pan
European planning, which includes support for regional
tives based on inter
regional comparisons. Clearly, the first requirement is the
availability of the national data in a suitable and

form. Making this a reality is one
of the main aims of EFICS and any pan
European FEIDSS. Simple catalogue sum
information will

be adequate for such purposes. For such comparisons multi
statistical tables may be sufficient tools in some instances, but back
up by graphical displays
of various sorts to simplify the perception, interpretation, and present
ation of more complex
comparisons and trends will be needed.

However, before any comparisons can be made or trends determined, EU Commission
planners need to be assured that they are “comparing like with like”. This is the underpinning
reason for the s
trong Commission push for European harmonisation in terminology,
nomenclature and measurement systems. However, until such a standardisation of
terminology, definitions and measurement standards are achieved, it must be accepted that
unadjusted comparison
s of national statistics are meaningless if they are based on differing
standards. In Europe, as clearly explained in the EFICS Study (1997), definitions of tree
mean diameter, tree volume, forest stand area, forest productivity, site index, forest types,

management regimes, and inventory and mensuration techniques and the frequency of


implementation of them, all vary between countries of the EU. Hence any FEIDSS based on
statistical comparisons of incommensurate national forest data are completely invali
erroneous and meaningless.

Harmonisation of standards was the original objective in the EFICS 1089 Regulation.
However, harmonisation is a major and expensive task, likely to take a long time to achieve.
One of the reasons for this is that harmonisat
ion of definitions and standards of forest and tree
measurement would destroy the long term within nation comparability of inventory and
research data. Different national definitions and standards have been often adopted in the
form they exist because the
y reflect local needs and circumstances. Hence, harmonisation is a
long term possibility that we cannot wait for before attempting to meet the objectives of

The situation is NOT entirely hopeless before harmonisation is achieved. The differing
andards for defining and measuring a particular forest attribute generally lie along a
continuum of possibilities and the point that is chosen in any one country may be a matter of
convenience, due to historical accident, or just arbitrary. The question c
an be asked (as it was
in the EFICS Study, 1997): what would the measurement in a particular country have been,
had they used the definitions and standards of another country? This is essentially not a
difficult question to answer, though there is clearly

no “theoretical” or “perfect” answer.
Empirically based statistical studies of the effects of the chosen national definitions and
measurement techniques should be able to come up with a set of empirically determined
conversion models. In some situations,
simple adjustment factors might be adequate, but in
general slightly more complicated (though still simple) conversion models will be needed.
Such standardisation methodology was considered in detail in the EFICS study, with a call for
more empirical data

upon which to base such standardisation work.

Such conversion models might be used to build a
ToolKit (SCTK)

which would be able to convert from the standards of all European nations to an arbitrarily
chosen common standard, or conve
rt all the data to a particular normative national
standard. Implication of SCTK on


In any analysis of data from forest inventory databases distributed across many countries, any
table or view extracted from a database
in one country will have to be subjected to conversion
operations so as to allow any aggregation over countries, or meaningful comparative analysis
between countries. Hence

operability functionality is disrupted. Attempting to
insinuate such co
nversion operations into the ddb operations would likely be messy, ad
and unsatisfactory.

In such circumstances, it makes sense to consider the construction of a centrally located
database of pan
European national inventory information (for important

selected variables,
which have already been standardised). It might be the case that the standardised attributes


stored should also be up
dated to a regular sequence of dates, possibly yearly, or 5
yearly, to
aid subsequent analysis processes. Such a data
base is known as a
Data Warehouse. The Needs of pan
European Forest and Environmental Researchers

We do not consider all the user group requirements analysis in terms of tool development in
detail here.

However, this section concludes with a

consideration of the requirements from FEIDSS of
User group 4: Forest and Environment researchers: data mining and model building,
the activities of this group are central importance in current activities relating to sustainable
forest management an
d biodiversity conservation in a changing climate, and possibly of high
importance to European planning for environmental and ecological stability.

For this group, national boundaries really are arbitrary lines drawn on the map of Europe.
The primary aux
iliary data/information sources are the geo
referenced maps of various soil,
geological and climatic surfaces, etc…. Forest growth data, in a spatially distributed form, is
also important as a dependent variable, as are social and economic indicators rela
ted to local
forest and timber industries throughout Europe. It is clear that national summaries of
inventories for whole countries, or even sub
regions, are not the most suitable data for the
analysis of the interrelationships between the various spatio
temporal ecol
econ factors.
It is likely that remote sensed imagery will be necessary for the apportioning of national
figures in suitable “densities” across countries in Europe. Spatio
temporal models of
changing patterns and relationships need to

be modelled, validated, and used in projection
mode, probably in association with a scenario analysis framework.

A library of forest growth models (they already exist within each nation) should be part of a
FEIDSS to assist such analysis and projection.

Alternatively these models would be stored
on a Forest Model Archive (FMA) and made available as a Web Service.

Of course, they
would need to be subjected to the SCTK for the comparisons between countries and regions to
be valid.

Consideration of the m
ost suitable facility for remote sensed image analysis would best be
made within the context of the current review of Geographic Information Sciences in NEFIS
[Note: GRASS is an open sources system]. Web portability of results is important, and XML
and RD
F, and SVG (Scalable Vector Graphics) (all W3C recommendations) are likely to be
suitable technologies. Use of Data warehousing, data
fusion and data
mining methodologies
are also indicated on the large and distributed data resources.

Similar consideratio
ns apply to
spatio temporal analysis of biodiversity and climatic factors
All of these issues are clearly of central importance to European Forest Science research, and
the planning activities of the EU Commission. Constructed data structure


Raw forest data comes in many “shapes and sizes”; from individual tree measurements, to
forest stand assessments, to national inventories; from remote sensed information from Lidar
devices, to very large scale remotely sensed imagery. Such data is me
asured at “space and
time points” that do not match from one data source to another, and the types and magnitudes
of the errors associated with each type of data differ.
Fusion techniques

attempt to
consider all this data simultaneously and so obtain

the maximal information from all of the
available data sources. Through the use of such techniques, enhanced spatial resolution and
enhanced accuracy in monitoring forest change is possible. Such techniques range from
statistical modelling methods, to ma
chine learning methods like neural networks, but also
include current research activity into the development of hybrid models which appropriately
combine varying
scale models which are suitable at the differing scales of measurement and
data collection. W
hile all these activities are essentially research activities at the present
time, their feasibility can be greatly enhanced by a fully effective FEIDSS.

ways of making hybrid data selections available to hybrid data
fusion algorithms
and other fo
rms of modelling should be considered as a possible useful extra service to be
provided by a FEIDSS.

The need within a FEIDSS for several modelling and analysis toolkits has been discussed
above. These include:
standards conversion (SCTK)
extended dec
ision support
a library of European Forest Growth Models
advanced statistical and
visualisation add
; the availability and use of
mining and data
fusion modelling
; and
a standardised centrally located data

of selecte
d forest inventory

Development of all of these analytical support facilities within a monolithic a FEIDSS is
conceivable, and essentially feasible. However, such a centralised strategy is potentially non
robust and would certainly cause a work over
load for, and bottle at, the central development
team. Alternatively, the modern open
source paradigm would call for de
centralised tool and
facility development, with the tools and analytical facilities being made available over a
federated network of t
ools access/usage. This is just what a grid is
. We therefore propose
the setting up of a demonstrator ‘Global FEIDSS Tools Grid’.

3.0 An overview of the main Corporate Web Software Players

Each of the three main companies, SUN, Microsoft and IBM, hav
e their own strong tradition
and contribution to building the software world of today, with characteristic and competing
product lines as indicated below. In considering the software platforms to adopt for an
extension of A FEIDSS, the commercial strategi
es of these companies needs to be taken into
account. It is necessary to ensure that cross
platform portability of a FEIDDS is not
endangered by changing strategies of software companies and computer manufacturers.


Unix, Java, CORBA : the open arc
hitecture paradigm.


The open architecture and source policy adopted by SUN is well known and draws unto it
a large and dedicated enthusiasts and developers. The Linux operating system is
becoming increasingly popular, and the Java language is the standard

language for web
related programming. Persistence and quality and security are assured. However, it has
to be accepted that the PC with a Windows operating system will continue to be
dominant in the next few years, and maybe even more so in the future.

Java based GUI’s,
though functional, will probably remain second choice to those users already on the MS

SUN has full support for web services in its Sun ONE Web Services Platform Developer

. Sun provide a complete support

framework, including free downloads, tutorials, and application examples. However, the
size of the tutorial primer in Web Services means that this environment is on
e for a
professional web software engineering team.


This is the environment chosen by the A FEIDSS develo
pment team, and would seem the
natural one for continuation of A FEIDSS development.


Bill Gates, Windows, .NET (VB.net or C#.net) and its use of Soap & XML,
Internet Explorer, SQL Server 2000, OLE DM for DM.

The bulk of the potential users of
A FEIDSS are probably locked into the PC with
Windows environment for the foreseeable future. Though accusations of low software
engineering quality are continually associated with MS products, their product range is
highly integrated, and provides a high
degree of ease of use and application to most of its
vast population of users. Its .NET software range brings a high degree of functionality
and programming efficiency, allowing easy access to component based programming,
and use of XML, and SOAP and WS.

The commercial manoeuvring of MS for monopoly is legendary, with the most recent
incident being its discontinuation of Internet Explorer as a product that can be used
without the full MS suite. Its shunning of Java in its .NET environment makes MS an
ward and selfish main player in the world’s software milieu.

However, MS also provides extensive support for web services developers,

). The MS Web Services B
asics web offering
provides a very accessible introduction to the whole (basic) field of web services and
seems rather more readable than the corresponding SUN provision.



mainframe computing, grid computing, autonomic computing, DB2.

IBM is making an increasing impact on the software industry, though its database (DB2),
grid and autonomic computing initiatives. IBM also provides a Web Services
environment, in WSDL Explorer, but their web sites do not seem as clear and transparent
as those of MS.



IBM leads in certain areas: their Emerging technologies Toolkit (ETTK) covers SOAP,
WSDL, Grid, ReGS, WS
Inspection, and UDDI as well as emerging technologies such as
autonomic computing. Grids and autonomic computing

are areas where SUN and MS do
not compete. Hence if it were envisaged that a FEIDSS were seen to need the distributed
high powered computing of a grid, then the IBM software solution might be appropriate.
This does not seem to be the case in the near fu
ture, even for a FEIDSS development for
Level 1 information services, though this situation might change rapidly once Level 1
database sharing were available in a FEIDSS.


IBM has adopted the Java programming environment (J2EE) so moving in this direction
should not be a problem for the A FEIDSS developer team.

There is howev
er, some doubt engendered from the IBM website about how generic their
software systems and solutions will be, and if they will be bind users to use of IBM
computing equipment.

W3C: (Tim Berners Lee, the semantic web (XML, RDF, …)

W3C is the internatio
nal body responsible for managing the adoption of web software
standards. It is led (intellectually) by MIT, even though there are over 400 major
software institutions/companies involved. W3C is chaired by Tim Berners Lee, the
inventor of the WWW, just o
ver a decade ago. The current list of activities of W3C are
given in the following list:





Device Independence















Quality Assurance


Semantic Web



Timed Text





Web Ontology

Web Services




XML Base

XML Encryption

XML Key Management




… a very wide range of top quality web software engineering pr
oducts which will,
between them have a very major role in the coming decades.

Certainly, the use of a W3C adopted standard will ensure persistence and quality etc…
However, there have been some questions about the vision of the W3C on how the web
will an
d should develop; questions of relevance and usefulness to the real business world.
It has been well publicised that the original idea for the web, by Berners Lee, was
actually closer to the semantic web that he is still moving towards, than to what we ha
today. The main push by Berners Lee, and the W3C towards the semantic web is through
the RDF and OWL initiatives.

However, in reviewing these initiatives, the expressed doubts do seem to have some
substance. These are presented after the DAML+OIL se
ction latter in this paper, so that


the criticisms can be can be seen in context of ontological languages and the semantic

:, SAS,

Oracle, GISystems, Library Science, the Apache Project (Tomcat,
Xerces). There are a wide range of facili
ties and technologies available within web
software and information engineering. Many have been looked at, but we mention only

the SAS Open architecture for SAS applications; some good features, but

DiGER: Distributed Generic Informatio
n Retrieval (DiGIR
) is a client/server
protocol for retrieving information from distributed resources. It uses HTTP as the
transport mechanism and XML for encoding messages sent between client and
server. It is an open source project hosted on Source Forge

and is currently in late
Beta stage of development. DiGIR was originally conceived to be the replacement
for the Z39.50 protocol used in the Species Analyst project, but is intended to
work with any type of information, not just Natural History collection
s. A major
contributor to DiGIR is the MaNIS project. This would seem a suitable upgrade
path for the A FEIDSS/GFIS Level 2 catalogue functionality, since it also makes
use of PHP and extends the Z39.50 protocol that has been used in A
ms to make use of MS SQL Server, rather than mySQL.


4. Web
related and web
enabling technologies

4.1 Metadata

“Data about Data”. Said so often, but needs correct interpretation!

Metadata glossaries


4.1.1 Data and metadata hierarchy

Hence m
etadata is also data. Hence metadata about metadata (metametadata) is data about
data about data. And so on…

The same kind of hierarchy also applies to models. A description of the logical structure of a
dataset or a database, possibly in terms of entiti
es and relations, is called a data
model, (not to
be confused with a descriptive statistical model fitted to data, or a mechanistic mathematical
model developed to explain data). A data model of first level metadata might be called a
model, etc… up t
he hierarchy.


CORBA, the middle
operability architecture for distributed databases,
possibly with high access rates, allows up to FOUR levels of such a hierarchy, before it says

With RDF and the ontological languages allowin
g statements about relations, i.e. reification,
one could in principle talk about meta
relations, (or “rei
lations”); again etc…

The ontological languages actually allow much more logical flexibility, and allow relations
between relations to be specified,

as well as “meta
classes” and recursively defined
classes,…etc…. see the DAML+ OIL section later.

4.1.2 Dublin Core

DC glossary


It is not necess
ary to review DC. It is the accepted standard for all information object
cataloguing activities, and is firmly established in both A FEIDSS and GFIS. DC can be used
in conjunction with simpler forms of metadata specification, e.g. simple HTML metadata
atements, and more complex metadata specifications, in for example RDF etc… Hence there
is no need to question its continued usage, as long as it is not mistakenly taken to be all that is
ever required in a metadata specification.

Note that DiGER, referre
d to above provides a possible upgrade path from DC to XML,
(without loosing DC).

Dublin Core examples, with refs to XML and RDF:


A DC metadata generator, and viewer

(recommended for their simplicity):




XML is the Extensible Mark
up language.

XML history and glossar


In fact, XML is not really a mark
up languag
e. It is a framework in which web mark
languages can be defined, using either a DTD (Document Type Definition) or though XML
schema. Each time a new DTD, or set of schema is defined with a new structure and set of
elements and attributes, a new langua
ge is defined. XML derived languages are therefore
unlimited in number, and the expansion in the number of such languages has been
“exponential” over recent years. Such proliferation of XML derived languages will
eventually lead to overload, and a collap
se, when there are several or many different dialects
to achieve essentially the same function. Very many XML
derived languages have already


fallen by the wayside. Hence, in choosing to use an XML dialect, or to invent its own, a
sortium needs to take some care and consider long
persistence of the dialect.
The idea of a ForestML potentially adds to such confusion,
especially if there turn out to be dialects like EForestML, GForestML, FAOForestML etc.
Coordination on a globa
l scale is required, or it might be better for ForestML to be a subset of
BiomeML, if such a thing were to exist.

The XML derived languages that are not likely to be lost in the confused proliferation are
probably those which achieve W3C recognition, even

though such recognition will not
necessarily guarantee that such languages are really relevant to real world requirements.
XPath, designed to help navigate the web, Xlink, Xquery and Xpointer are all moving to W3C
recognition, and may well become establi
shed dialects of the professional web software

engineer/programmer. However, the complexity of such a range of dialects is not likely to
lead to their common usage, other than as generated code from APIs/wizards.

XML is designed to describe data structu
res, and this is its widest usage in the web world. It
seems fairly clear that it should be used as the core to metadata description within FEIDSS,
(including DC). However, specification of the structure of a database is the function of the

process in classical database theory, whether relational or object
based database structure descriptions should therefore follow the well
principles of database data modelling.

Choice of a software environment or database fra
mework which has XML functionality can
remove the need to write or read XML data
structure specifications; this can done
automatically by the database. This is so for MS .NET and MS SQL Server 2000. It seems
that there is not a native XML functionality t
o mySQL, though there are several proprietary
products that enable this, many in PHP. However, efficiency questions may be asked about
PHP as a mainstream programming environment for a web
based DBMS.

If actual data is to be stored in XML, then a
data bi

framework for access to the data
from the XML (de
marshalling) is to be recommended. Such a framework defines an object
model for the data in the programming environment of choice such there is a direct
correspondence (binding) with the data

in XML. This binding is then used to allow
direct access to XML data from the application programming language without having to deal
with the XML coding.

is such a framework for a Java programming environment,
such as A FEIDSS. However, when de
aling with manipulation of large databases, such Java
based frameworks will be relatively slow.

Note that an XML specification of a database structure has most of the features of a
specification in an ontological language. Hence one might argue either th
at such ontological
language are not required for such database structure definition, or that the ontological
languages are the ideal XML based facility for such database structure specification.

XML is verbose, and while it, or derivatives, are well
ed to describe the structure of
metadata, its verbosity makes it not ideal for embedding and transport of the actual data
instances of a very large data resource transport. It is feasible to include small amounts of


message data in the XML, but for large

database information transport, as might be required
in a Level
1 Forest Information system, (maybe involving database mirroring) the transport of
data needs to be in a comms
efficient form separate but linked to the XML metadata/structure
specification m
essaging. These comms issues does not seem to be fully documented in the
XML literature (e.g. Wrox Professional XML), but there is proprietary software which does
address the problem. However, with such a route NEFIS would be in unstable territory for
e long
term, unless it uses one of the major software players.

4.3 Web Services : SOAP, WSDL &

“A web service is a software application identified by a URL, whose interfaces and bindings
are capable of being defined, described and discovered by X
ML artefacts and supports direct
interactions with other software applications using XML based messages via Internet
protocols.” (Wilson, 2002).

The service provider creates the web service, and publishes the service with the service
registry. A ser
vice requester may use the service registry to find a service, and then uses the
information to bind to the service by invoking one of its methods. The web services approach
exploits the separation of content and business logic in a loose
coupled architect
ure enabling
changes in business logic to immediately propagate to the user interface.

4.3.1 SOAP

SOAP (Simple Object Access Protocol) is a newly developed remote call protocol. It is an
based wire
level protocol that uses HTTP as its transport mec
hanism. This in turn
means it is platform and language independent and can function in any architecture that
accesses the Web, and can use any existing Internet protocol to facilitate its communication,


has managed to es
tablish itself commercially very quickly in an industry accustomed to
vendor specific technologies. The fact that most major vendors have accepted it may come
from the fact that it is a composition of two well
established Internet standards: HTTP and
There will not be an Internet protocol with such omnipresence as HTTP, (even though
an MIT spin
off company has tried to replace HTTP in the last two years with CURL, and
failed) and XML is becoming as ubiquitous as HTTP.

By using HTTP as its transport me
chanism, SOAP has bypassed the problem of access. It can
‘go’ anywhere HTTP can and so one of SOAP’s most significant advantages is the navigation
of the firewall. The standard port for HTTP is port 80, and most firewall administrators leave
this open to g
ain access to the Internet from within the company Intranet. The other side of
this coin is that there have been security doubts about the use of SOAP and WS. However,
extra security protocols are now in place, implemented by the last release of .NET, a
nd are
being implemented for the “Java/SUN” environment


Applications making use of SOAP can be written in many languages, (including Perl,
Java,…) and hosted on many platforms (including Windows, Linux,…).

Another reason for the rapid adoption of WS is t
he triviality of implementing it, for
experienced web software
engineers. Quotes from such professionals include:

“… once I got it working, I thought, is that all there is to it?”; K. McMannus, Greenwich.

“WS can be implemented in .NET in just a few c
licks; all the structure, all the classes are
just set up for you, and you are finished.” Alan Butler, Greenwich.

“With web services, the programmer does not have to know about SOAP or XML, or to
programme either; both are used automatically.” Alan Butl
er, Greenwich.

4.3.2 WSDL

WSDL (Web Service Description Language) is an XML
based language for the description of
the web service, and contains information including the location, behaviour, data types and
bindings of the web service. WSDL can be compa
red to the Interface Definition Language
used in traditional middleware solutions.


relates to the description stage of a web service, namely during the ‘publish’ and
‘find’ stages of a web service lifecycle. For a web service to be consumable, the we
b service
must be expressed in some way to describe what its function is, where it can be located and
how to invoke it. Specifically this would mean expressing in well
defined semantics how the
service should be accessed and what operations it supports.

Distributed systems have a language for interface description: IDL for CORBA or MIDL in
Microsoft’s COM and DCOM. Because the WSDL file is represented as a XML document,
the requesting application does not need to know anything about platform, language or
By using WSDL, the client code is reduced, and a dynamic discovery is initiated, allowing for
easier web service modification.

4.3.3 UDDI (Universal Description, Discovery, and Integration)


is a project to create a network of combined regi
stries using SOAP interfaces as a
central repository where web services can be published. The UDDI contains the web service’s
metadata, a pointer to the web service’s WSDL, and a set of port definitions. UDDI enables
web services to be discovered in an aut
omated manner. It can be thought of as a name lookup
service. Although not yet a formal web service standard, UDDI is responsible for the creation
of a registry of published web services.

It seems as if most of the functionality of UDDI will be provided

within FEIDSS (and GFIS)
by the DC metadata searchable catalogue. In fact GFIS might become a UDDI registry in the
long term.


4.3.4 Software
Engineering of Web Services

den Haan (2003) has prophesised the evolution of web services through three phase
Integration Middleware (Internet Middleware), Cross
Enterprise B2B Integration, and The
Ubiquitous Platform (2003). This topic will be considered further in Part 2.

4.3.5 Comparison of middleware technologies in relation to a FEIDDS.

The tradition
al middleware solution provided by CORBA has significant strengths in some
situations. Situations that demand millions of transactions across large distributed systems
may not be the ideal environment to deploy web services; the traditional solution is bes
Enterprise infrastructures often are tightly
coupled, and again traditional middleware
solutions offer greater control and reliable security mechanisms and are therefore more
appropriate. In a tightly coupled architecture, each component has prior know
ledge of where
to find another component, and what that component does. ‘Tightly
coupled’ systems are
desirable for some systems because of this knowledge between components, security and
privacy are enhanced, and because the endpoints are known, a level o
f service is guaranteed
and maintainability improved. These factors explain why the ‘tightly
coupled’ approach is
used in enterprise computing and B2B applications.

However, traditional middleware is complex to develop and maintain. Also, web access is n
straightforward for traditional middleware technologies. Traditional middleware is often
wired’ to certain ports that may or may not be available on potential partner’s

If flexibility, ease of development, scalability and web ac
cess are primary considerations, and
the traffic is not high, and communication is required between loosely coupled systems, then
Web Services have a lot to recommend them over the traditional solution.

There are some arguments against Web
Services, incl
uding: the possible saturation of port 80,
or the fact that firewall administrators can prohibit SOAP/XML
RPC traffic; XML data is
quite verbose, consuming memory and network bandwidth, and although this is less of a
problem with the falling costs of bandw
idth, it should still be considered in the choice of

4.3.6 Federated Web Services and Grids Federated Web Services

The recent IUFRO Interoperability Conference site at:



includes a presentation by Brachman et.al. (2003) on the Canadian National Forest
Information Service (NFIS) which makes use of Federated Web Services. The Canadian
NFIS Demo site can be found


However, registration seems to be restricted to those valid zip codes and US/Can. phone
numbers. Grids through

Web Services

Grids are networks which make distributed computing and analytical tools available to
authenticated users, making use of WDSL and various networking toolkits including the
Globus 3 Toolkit.

A UK Grid Support Centre exists at

which references:

User's Guide Core Framework on the Globus Toolkit 3 (Alpha) which may be found at:

globus.org/ogsa/docs/alpha/users_guide.html Other FEDS/GRIDS/PORTAL references

1. “Navada Federation” etc. : grids and Federation.


2. IBM Grids/federations/Portals


3. A Portal view of OGSA.


4.3.7 Language Support for Web Services


Microsoft has been involved in the development of SOAP since its initial sta
ges before
standardisation by the W3C. It is not surprising then, that the .NET framework supports the
development and consumption of web services through a series of classes and tools. The
concept behind .NET is to provide a framework whereby applications

can be developed and
deployed through multiple web services. .NET enables the development of applications in any
of the .NET family of languages, ASP.NET, C# or VB.NET, without prior knowledge of

Perl (Practical Extractions and R
eporting Language)

Perl has one of the most comprehensive SOAP implementations and a reputation of being the
language for obtaining and manipulating data on the Web. It is therefore a particularly strong


language for web services development, and its SOAP

implementation SOAP::Lite by Pavel
Kulchenko is in keeping with this. SOAP::Lite is a collection of modules which support a
diverse set of protocols additional to HTTP, including FTP, SMTO, POP3, Jabber and IBM
MQSeries. SOAP::Lite is not restricted to th
ese protocols; if requirements dictate, SOAP::Lite
provides the mechanism to implement other protocols not mentioned above. SOAP::Lite has
partial support for WSDL schema.


PHP (PHP: Hypertext Preprocessor) is a general
purpose web scripting language

that can be
embedded into HTML. PHP is a procedural language that can act in an object
oriented way,
making it suited to the modular architecture of web services development. The case for using
PHP to program web services is particularly strong for develo
pers who lack administrative
access to their hosting servers. The installation of SOAP or XML
RPC within PHP is one of
its significant advantages, as it requires only that the xmlrpc.inc or nusoap.php file are visible
by all the PHP scripts, and so using t
he include statement can be imported from anywhere.


JSP (Java Server Pages) are an integrated component of SUN’s WS framework. These are the
natural environment for the extension of A FEIDSS central server into Web Services, in view
of the software

environment in which the A FEIDSS demonstrator has been built.

4.4 The
Semantic Web
: Ontological Languages

[Semantic web glossary:

Ontology languages are mark
up languages for use with web information resources, as
HTML. Full
blown ontology languages are refinements of RDF (Resources Description
Framework), the basic ontology language. All are written in XML, but they allow complex
relationships in the data to be represented, such as relations between relations. The idea is
that either a browser, or server
side tools, which could parse an ontology language markup
would be able to use an embedded
logic engine

to enhance the precision

and usefulness of the
search, and the information resource usage.

4.4.1 RDF

RDF: W3C has led in the development of the Resource Development Framework (RDF) to
remedy limitations of XML (to qualify relations).


the Resource Description Fram

is a foundation for processing metadata; it
provides interoperability between applications that exchange machine
information on the Web.” (Ora Lassila and Ralph Swick (1999))


In 1999 the W3C released its RDF Recommendation, and toda
y, after growing interest in the
topic, RDF is at the very forefront of Semantic Web activity.
Although RDF is proving
successful as a future language of the Semantic Web, some limitations have been identified
that could be resolved through the use of a mo
re heavyweight XML schema
based language.

RDF was created by largely as a result of initiatives of the W3C (led Tim Berners Lee) as its
first step towards the semantic web that TBL had always envisaged when he invented the
internet. Some, such as Hjelm
(2001) argue that RDF is all that will ever be needed for the
creation of the semantic web. W3C would seem to disagree with Hjelm’s view, and have
been the main force behind, first the linking up of DAML and OIL, and are now pushing
OWL (Web Ontology Lan
guage) as the way to the semantic web.

Basically, RDF inherits the ‘object
oriented’ element/attribute hierarchical structure from
XML but extends itself to include set theoretic logic and graph theoretic descriptions of the
metadata structure of a data

In general, an “object” is a set of possible instances, usually defined by some shared property
of those possible instances. Such objects, as sets, (or even more generally as bags), with
properties, can have relationships between them. There are
simple set theoretic relations,
such as inclusion, (e.g. all men are human), but other relations might be defined, e.g. ‘men
like blondes’. In graph theoretic terms, each object is a node, and each directed link is a
relation. Hence RDF is generally spec
ified in terms of
, in the form :

ArB (A has relation “r” to B),


Subject verb Object (e.g. <#pat> <#knows> <#jo> .)

RDF then allows us to talk about relations, and give them their own properties. This
“adverbial facility” is sometimes call

The article by Tim Berners Lee (2003) is a simple tutorial on the N3 notation (i.e. ArB).

While the simple graph theoretic characterisation of RDF makes it seem simple, some
authors, such as
Hjelm (2001) obscure this clarity in a plethora of detailed examples of the
application of RDF (from labelled images, DC, though to ontology specification for
realization of the semantic web), with no clear explanation of the logic.

4.4.2 DAML+OIL



Family encompasses a group of mark
up languages from the DARPA Mark
Language Project. Artificial Intelligence inspired and knowledge
based DAML technologies
are serious contenders in the development of the
Semantic Web. DAML+OIL and the newer
L and DAML
S (DAML for services) go beyond XML and RDF because they have
“well defined model
theoretic semantics as well as an axiomatic specification that determines


the languages’ intended interpretations. DAML+O
IL is unambiguously computer
interpretable”. (McIlraith, Son and Zeng, 2001)

Standardised web services descriptions may currently be achieved using WSDL. However,
web services can be described using DAML and RDF, and DAML
S has been developed

for semantically enabled web services.

The ontology “languages” of DAML, OIL, DAML+OIL, OWL extend the ability of RDF to
describe logical relations, allowing relations between relations to be described.

DAML+OIL, until recently, has been regarded as th
e state
art in ontology languages. It
is a result of the union of efforts of the
DARPA Agent Markup Language (DAML) Project

funded by the US Defence department, and the
Ontology Inference Layer

(OIL). The OIL
initiative is funded by the European Union IST programme for Information Society
Technologies under the On
Knowledge project (IST
1013) and IBROW (IST

DAML+OIL was developed by a
joint US/EU Agent Markup Language committee
, with
participation from the DAML project and from the OIL project, as well as the W3C. This
document is a submission to the
World Wide

Web Consortium

from Lucent Technologies.
Note that W3C is an MIT led initiative, chaired by Tim Berners
Lee (of MIT). DAML+OIL
may therefore be seen as the front
edge of W3C efforts to fully implement the concept of the
semantic web
, as originally envisaged by Berners


for the full reference documentation, and
DAML+OIL walkthrough

for an introductory accoun

There are no browsers which parse DAML+OIL at the currently, though there are some server
side tools. This whole topic is possibly one for the future use, rather than now!

However, in view of the prestige of the institutions involved with DAML+OIL (
and in
particular the EU IST programme), it would seem appropriate to consider the potential
use of DAML+OIL in a future FEIDSS project. See the Appendix for an example of

[An Alternative view:
A criticism of W3C vision on the semantic web. Se


It would seem that in terms of current FEIDDS development, all that should be aimed for in

near to middle future is the capturing of Level 1 metadata for contributed forest data
sources, in XML or an XML derived language, possibly RDF, reflecting the semantic
structures that are already in existence in the DBMS’s for those data sources. It see
ms that
automating such a procedure should be the primary aim of a FEIDDS, rather than hoping that
future intelligent web agents will assist more efficient resource search and discovery.




is a free open source ontology editing facility, which through the storage of instances
can be used to build knowledge bases.

2000 was developed by
Stanford Medical Informatics

the Stanford University
School of Medicine

with support from the following agencies:

Defense Advanced Research Projects Agency

National Cancer Institute

National Institute of Standards and Technology

National Library of Medicine

ational Science Foundation

The developers are clearly working in a major application area of web technologies and have
developed a product that is matched to the requirements of its users.
Since the area of
Medical Informatics is so much more advanced
than that of Forest and Environmental
Informatics, the lead given by medical Informatics is a good one to consider following.

4.4.4 The Protégé DAML+OIL plug in.


Such a plug in is available at the above ULR, but it seems slightly limited at the moment. See

for the restrictions on the use of DAML+OIL through t
his plug in to Protégé.

However, Protégé, with this plugin can be used for authoring DAMl+OIL ontologies. The
W3C ‘Person’ example source, when loaded, with a .daml extension, seems to loose the
implicit hierarchical structure of the ontology specificati
on, though this can easily be
corrected using the Protégé GUI. Some of the other features and constraints of the original
DAML+OIL source are also not rendered. The plugin does not support user
Metaclasses, and seemingly not DAML+OIL feature of an
onymous class definition by
restriction. Even saving an edited ontology seems problematic at the current time. This
demonstrates the limited current functionality of DAML+OIL, in the Protégé context at least,
but future potential should not be underestimat
ed for this reason.


4.4.5 DAML
S and OIL
S spec and review:


4.4.6 OWL: Web Ontology Language


seems to be a serous competitor to DAML+OIL for recognition by W3C as the markup
language for sites with useable information resources.

It is not clear which of these competitors is superior, or if there are any major differences
etween the functionalities.

Recent state of W3C OWL documentation:


The following

are our Working Drafts in Last Call:

* OWL Web Ontology Language Overview


* OWL Web Ontology Language Reference


* OWL Web Ontology Language Guide


eb Ontology Language Semantics and Abstract Syntax


* Web Ontology Language (OWL) Use Cases and Requirements


4.5 Other web technologies

4.5.1 Topic Maps


Another XML
derived language speci
fication. See Park (2002) the standard text. Topic maps
arise by analogy with mind/concept
maps, essentially diagrammatic views of concepts and
their links. The main advantage is the immediacy of the diagrammatic from of illustration
used to represent the

topic map; essentially a directed graph. It is claimed that Topic Maps can
represent RDF (not surprising since RDF is a directed graph consisting of
“NodeRelationNode” links), and implement the semantic web. Possibly used in GFIS, and
not un
naturally, s
ince Topic Maps do seem like a logical extension of the catalogue
framework, based on the “More General term”, “More restricted term”, and “Related term”.
However, Park (2002) is not impressive, and seems rather hyped and confused. XTM not
being consider
ed by W3C.



Sharing mathematical models over the web.

Mathematical models can be regarded as a meta
representation (probably better to avoid the
term ‘meta model’ which is reserved for logic
based models of relations between metadat
a) of
data resources that they encompass or encapsulate. Statistical models can be regarded
similarly, but with measures and models of uncertainty included. Alternatively, when the
models arise from analysis of the data resource, then the models may be r
egarded as
discovered knowledge.

Such models are and extremely efficient representation of the data resources from which they
derive, and they have the benefit that they can be used directly and flexibility in Decision
Support Systems to aid in the optimi
sation of management choices.

The sharing of such mathematical and statistical models over the web is therefore important.
There seem to be two current markup language routes by which this can be achieved.


“MathML is intended to facilitate the u
se and re
use of mathematical and scientific content
on the Web, and for other applications such as computer algebra systems, print typesetting,
and voice synthesis. MathML can be used to encode both the presentation of mathematical
notation for high
ty visual display, and mathematical content, and for applications
where the semantics plays more of a key role such as scientific software or voice synthesis.

MathML is cast as an application of XML. As such, with adequate style sheet support, it will
mately be possible for browsers to natively render mathematical expressions. For the
immediate future, several vendors offer applets and plug
ins which can render MathML in
place in a browser. Translators and equation editors which can generate HTML pages
the math expressions are represented directly in MathML will be available soon.”

MathML 2.0
, a W3C Recommendation was released on 21 Feb 2001.
“A product of the W3C
Math working group, MathML is a low
vel specification for describing mathematics as a
basis for machine to machine communication. It provides a much needed foundation for the
inclusion of mathematical expressions in Web pages.”

MathML source can be generated from Scientific Word, and from M
athType (the full
equation editor of which the MS Word equation editor is a sub
set.), or Mathematica, or by
hand. However, the current functionality seems very limited.


PMML (Predictive Modelling Markup Language); another XML derived language. T
seems to be a MS initiative, but with many other prestigious players involved, so it may go
somewhere. The source of the activity is the MS OLE DB for DM which is an MS open


architecture aimed to assist Data
Mining software suppliers provide their sof
tware products in
a form that may be easily linked into data warehousing facilities such as SQL Server 2000.

The MS OLE DB for DM example code consists of about 6000 lines of dense C++ code, for
the very simple “Naïve Bayes Classifier”. This would take

at most 6 lines of code in almost
any statistical language/package such as S
PLUS/R. It seems the PMML code is merely for
rendering the documentation/specification of the model used. There seem to be no facilities
of parsing it into executable (and ther
efore useful) code/algorithms.

Not suitable for A FEIDSS usage at this time.

4.5.3 Browsers

It seems that none of the standard browsers can parse and render RDF, (or make use in any
way of the Ontology languages, DAML+OIL & OWL), or MathML, other than
the W3C
browser Amaya which can render RDF and MathML. Hence we consider it here.


From the July10 2003 release of Amaya by the W3C.


“Amaya is
a very powerful and flexible browser which is available for free download fro the
W3C. It has the advanced facilities for MathML and XML authoring and RDF rendering.

Amaya is a Web client that acts both as a browser and as an authoring tool. It has been
designed by W3C with the primary purpose of demonstrating new Web technologies in a What
You See Is What You Get (WYSIWYG) environment. The current version implements the
Hypertext Markup Language (HTML), Extensible Hypertext Markup Language (XHTML),
matical Markup Language (MathML), Scalable Vector Graphics (SVG), the animation
module of Synchronized Multimedia (SMIL Animation), Cascading Style Sheets (CSS),

Hypertext Transfer Protocol (HTTP).

With Amaya, you can manipulate rich Web pages contain
ing forms, tables, and the most
advanced features from HTML. You can create and edit complex mathematical expressions
and simple SVG graphics within Web pages. You can style your documents using Cascading
Style Sheets. You can publish HTML, XHTML, MathML,
and SVG documents on local or
remote servers through HTTP. A first support of generic Extensible Markup Language (XML)
allows you to display the content of every XML document and to associate with it a CSS

However, Amaya does not verify RDF o
r DAML+OIL documents yet, and has not been
developed to the stage that it can be used as an authoring tools for RDF. Other software, such
as Protégé has to be used for this.


4.5.4 ebXML

June 2002 issue 190. Simon Bisson takes a look at ebXML, a tool

for global business to
business XML application development. Probably not relevant to FEIDSS.

Introducing ebXML



5. Conclusions in relation to WS, the
Semantic Web


5.1 WS and the Catalogue function in FEIDSS

In view of the considerations in this section and the earlier XML section, it would seem that
there are no

essential problems to providing the a FEIDSS/GFIS cataloguing services as web
services. Doing so would allow platform independence for the service users, and as a WS
GFIS would be more easily discovered by user groups who might have indirect interests in

forest information. Also, being a WS would potentially allow the catalogue into federated
SOAP provisions in a wide number of areas linked to forestry.

However, in terms of service provision to those who are already service users, the movement
to the ca
talogue to being a WS would seem to offer few major advantages.

On specific technical issues, the following seem to be self

Use of Dublin Core as the basic cataloguing structure, as adopted by GFIS and A
FEIDSS, should be continued, with any qua
lifications and extensions being shared
though all nodes of GFIS.

Use of XML should be adopted as the general standard for resource metadata and
structure specification. Expression of the DC core in XML will enable easy extension of
DC if and when require

Web services technology, both for enhanced resource discovery and information sharing,
using SOAP, XML and WSDL, are easy to implement in all computing environments,
and are not inconsistent with the continued use of DC. WS also offer the potential to
extend the architecture to grid functionality. Use of WS is therefore recommended for
use in a FEIDSS cataloguing function.

The system development for the cataloguing function needs to be kept flexible enough to
adapt to future moves in web technology tha
t might be associated with the semantic web
concept. At this stage the use of RDF for some semantic
web oriented functions should
be considered for prototype development.


5.2 WS and

functionality of a FEIDSS

The crucial requirements in respect to


functionality of FEIDSS are:


automatic capture of the structure of the raw data, preferably in XML, in a set of
distributed databases containing data on a common theme.


a means of querying the ddb and efficiently transporting the selected d
ata to the user.

(iii) ensuring that the selected data is consistent if drawn from source databases in regions
with differing standards.

It seems that use of a DBMS environment(s) which have automatic native XML functionality
would seem preferable. It mi
ght be that a data
binding framework (probably Java based)
should be used in conjunction.

For a FEIDDS Level 1

functionality, involving the transport of large amounts of data
from large databases with known structure, Web Services may not be adequat
e. However,
WS/UDDI could be used as an add
on feature to enhance web
based accessibility of a
FEIDSS Level 1 data resource.

There seem to be three alternative technical approaches that may be considered and discussed.
All of these are in addition to
a base DC core metadata characterisation.

CORBA is the well
established middle
ware solution for the management of stable distributed
databases. It is high
quality, high
security, and robust. However it required substantial
professional personnel for de
velopment and implementation.

Federated WS (XML + SOAP +WSDL) are easy to implement in all environments. Use of
WS, in association with UDDI, aids in resource discovery. There have been some security
concerns that are being addressed, and the technolog
y is said to be most suitable for low
traffic requirements. The technology is most suited to loosely coupled and dynamically
changing networks of information resources.

The “Semantic Web”, the “brain
child” of Tim Berners Lee is a much
hyped topic, and
basis of substantial efforts by the W3C. RDF seems to offer considerable scope for detailed
resource metadata characterisation, and could possibly linked with a means of data access,
and transfer, possibly using WS. Use of RDF would put the developme
nt in line with the
W3C vision for the semantic web, and additional XML
based functionality, such as X
Pointer, Xquery, should be easy to include as they become accepted standards. The down
side of this option is that it is “experimental” and there is a l
ack of current client
side tools
that can make use of the newest of semantic
technologies. However, sever
side tools do


5.3 Good Software Engineering Management Practice in relation to a “Global FEIDSS”

Many organisations are currently int
erested and involved in Forest Information Services. The
European Commission and IUFRO are independently interested in this area, but the
commonality of interest has led to the aim that EFIS will be a node of GFIS.

Clearly, the nomenclatures of any new
FEIDSS and GFIS must be consistent, and shared.
Ideally the core cataloguing systems developed in the various nodes of GFIS should be of a
common form, to allow complete functional transparency between nodes within GFIS. If
possible, this identity of for
m should be extended to include all computing, software and web
features, with a common and shared design and software environment being used. The
catalogue software system should ideally be designed in such a way to allow GFIS nodes,
such as EFIS, to dev
elop extra facilities and functionality for their nodes, as might be
considered necessary. Such a policy essentially amounts to an “open” GIS framework.

To support the software development within a “Global FEIDSS” both in the


lities, it is suggested that the following two options (not exclusive) be

That responsibility for the development of the core structures and software should rest
with a “Global FEIDSS Software Design and Implementation Group”, involving all
ose with an interest and capacity to contribute.

That the core FEIDSS system should be open architecture, and open source, so as to
allow wider contributions to the system development, and also to allow consistent
nodal extension facilities to be developed

by nodal development groups.


Modeling XML Applications with UML Practical e
Business Applications”,
David Carlson

Addison Wesley, New Jersey, 2001

Professional XML Web Services”
, Patrick Cauldwell, et al, Wrox, 2001

Engineering Distri
buted Objects
”, Emmerich, Wolfgang, Wiley, West Sussex, 2000

Software Engineering A Practitioners Approach European Adaptation”,

Fifth Edition,

Pressman, Roger S., adapted by Darren Ince, McGraw
Hill, Berkshire, 2000

Programming Web Services with SOAP
Snell, James, Tidwell, Doug and Kulchenko, Pavel
O’Reilly, Sebastopol, CA, 2002

Programming Web Services with XML

St. Laurent, Simon, Johnston, Joe and
Dumbill, Ed, Foreword by Winer, Dave, O’Reilly, Sebastopol, CA, 2001

RPC Specification
”, Wi
ner, Dave, 1999



Apicella, Mario (2002) “
Beyond Integration
”, InfoWorld, September 2002



Apperly, H. (2003) “

and Component
based Development Usin
g the Select
Perspective and UML
”, , Addison Wesley, London.

Apshanka (2002). “
Professional Open Source Web Services
”, Wrox, June 2002

Baiura, R. et.al. (2001). “
Professional ASP.NET Web Services
”, Wrox, 2001

Box, D., Ehenebuske, D., Kakivaya, G., Laymen,
A., Mendelsohn, N., Nielsen, H. F., Thatte,
S. and Winer, D. “
Simple Object Access Protocol”.
W3C Note.


Bellman, (2002) “
Web Services

A Waste of Words
”, IT
Director, October 2002


Lee, T. (1990). “
Information Management: A Proposal
”, CERN.


Lee, T. (1998) “
Semantic Web Road Map
”, September 1998.


Lee, T. (2001) “
Business Plan for the Semantic Web
October 2001.


Lee, T., Connolly, D., and Swick, R. (1999). “
Web Architecture: Describing and
Exchanging Data”
, W3C Note, June 1999.


Lee, T., Hendler, J., and Lassila, O. (2001) “
The Semantic Web
” The Scientific

American, May 2001.


Lee, T., Hendler, J., and Miller, E. (2002)
ting Applications on the Semantic
", Journal of the Institute of Electrical Engineers of Japan, Vol 122(10), October,
2002, pp. 676

Bertolini, C., Preist, C., and Trastor, D. (2002) “
Semantic Web Support for the Business
Business E
commerce Life
, WWW2002, May 2002, Honolulu, Hawaii, USA,

Booch, G. Jacobson, I., Rumbaugh, J. (1998) “
The Unified Modelling Language User Guide

Addison Wesley, 1st edition, September 1998.

Brachman, B., R. Morrison, B. Low. (2003) A Di
stributed Access Control System for
Secure Delivery of Web Services in a Federated Organization.


Brodie, M. L., and Stonebraker, M.(1995) “
Migrating Legacy Systems: Gateways, Interfaces
& the Incremental Approach
”, Morgan Kauffman, 1995.

Brown, Tony M. (2002) “
Web Services

EAI Killer The Final Word
”, eAI Journal, April 2002

Brown, John, J
ohnson, Simon and Kelly, Kevin. (2002) “
Using Service
Oriented Architecture
and Components Based Development to Build Web Services Applications


Software White Paper, TP032, October 2002.

Burbeck, Steve. (2000) “
The Tao of e
business Services
”, P
h. D, October 2000.


Carlson, David. (2001) “
Modeling XML Applications with UML Practical e
Applications”, Addison Wesley, 200.

de Jong, Irmen , et al. (2002)

Web Services/SOAP and CORBA
”, April 2002



den Haan, Andre. (2003) “Web Services: A Sea Change Taking Place”, eAI Journal,
February 2003.


Dumbill, Ed. (2
000) “
The Semantic Web: A Primer
”, November 2000.


Emmerich, Wolfgang (2000) “
Engineering Distributed Objects
”, Wiley.

Gottschalk, K, et al.(2002) “
Introduction to Web Services Architecture
”, IBM S
Journal, Vol 41, NO2, 2002

Gnaho, C. and Larcher, F. (1999) “
A User Centered Methodology for Complex and
Customizable Web Applications Engineering
”, Proc. 1

ICSE Workshop on Web
Engineering, ACM, Los Angeles, May 1999.

Hammersley, Ben. (2002) “
eks go hack to the future
”, The Guardian, 23 May 2002.


Hjelm, J. (2001)

Creating the semantic Web with R
DF. Wiley.

Lassila, O., Swick, R R. (1999) “
Resource Description Framework (RDF) Model and Syntax
”, February 24 1999.


McIlraith, S.A., Son, T.C. an
d Zeng, H. (2000)
“Semantic Web Services
”, IEEE Intelligent
Systems, 1094
7167/01, 2001.

Ogbuji, U. (2002) “
The Languages of the Semantic Web
”, The New Architect, June 2002,


, U. (2000
) “
An introduction to RDF
Exploring the standard for Web
”, December 2000


Park, J. (Editor) (2002) XML Topic Maps: Creating and using Topic Maps for the Web.
Addison Wesley.

Peters, L.J
981) In
Forward to Software Design: Methods and Techniques
. Ed. Belady, L.,
Yourdon Press.

Powel, T.A. (1998). “
Web Site Engineering”
, Prentice
Hall, 1998

Siddiqui, B. (2003) “
Web Services Security, Part 1
”, March 2003


Systinet, (2002a) “
Introduction to Web Services

Systinet 2002b “
Introduction to Web Services Architecture

Szyperski, Clemens (1998) “
Component Software Beyond Object
Oriented Programming
Addison Wesley.

Wilson, M. (2002) “The Future of

the World Wide Web, Web Services

Peer to Peer Data
Processing”, March 2002.


Winer, D. (2001) (2001) Foreword in “
Programming Web Services with XML
St.Laurent, S., Johnson, J., and Du
mbill, (Eds). pg xi, First Edition, O’Reilly &
Associates, June 2001.

Yager, T. (2002) “
The Windows Way to Web Services”
, InfoWorld, January 2002.




WS Glossary :

WS Glossary:

Metadata glossarie


Web Ontology glossary and RDF Glossaries:




DC glossary:

Semantic web glossary:

Glossary Grids:

See also:


Internet Glossary:

XML history and glossary:



IMESH: Technology Reviews:


The W3C Persons Example, using Protégé with the DAML+OIL Plugin.

Here, we give some extracts, in sequence from this “Person” example, just to give a flavour of
the language. It is a set/logic
theoretic and may be regarded as ‘object
oriented’, with
‘classes’ having inheritance. ‘instances’ are the ‘data cases’ which might be used to build up a
knowledge base in Protégé, but are not used in the DAML+ OIL plugin mode.

The name spaces used are indicated below. These are generated auto


$Revision: 1.9 $ of $Date: 2001/05/03 16:38:38 $


xmlns:rdf ="http://www.w3.org/1999/02/22




xmlns:dex ="http://www.daml.org/2001/03/daml+oil

xmlns:exd ="http://www.daml.org/2001/03/daml+oil

xmlns ="http://www.daml.org/2001/03/daml+oil

Some classes are defined, and some relations
hips between them:

<daml:Class rdf:ID="Animal">


[Actually, “label” is not supported]


This class of animals is illustrative of a number of ontological idioms.



ass rdf:ID="Male">

<rdfs:subClassOf rdf:resource="#Animal"/>


<daml:Class rdf:ID="Female">

<rdfs:subClassOf rdf:resource="#Animal"/>

<daml:disjointWith rdf:resource="#Male"/>


<daml:Class rdf:ID="Man">

<rdfs:subClassOf rd
f:resource="#Person"/> [Note that Person has not yet been defined]

<rdfs:subClassOf rdf:resource="#Male"/>


A Property is defined

<daml:ObjectProperty rdf:ID="hasParent">

<rdfs:domain rdf:resource="#Animal"/>

<rdfs:range rdf:resource="


And, a relation between properties:

<daml:ObjectProperty rdf:ID="hasFather">

<rdfs:subPropertyOf rdf:resource="#hasParent"/>

<rdfs:range rdf:resource="#Male"/>


Eventually, the person Class is d
efined in a rather complex way:

<daml:Class rdf:ID="Person">

<rdfs:subClassOf rdf:resource="#Animal"/>




<daml:onProperty rdf:resource="#hasParent"/>

<daml:toClass rdf:resource="#Person"/>






<daml:Restriction daml:cardinality="1">

<daml:onProperty rdf:resource="#hasFather"/>





<daml:onProperty rdf:resource="#shoesize"/>





Note, DAML+OIL is declarative.

As, a final snippit, an ordinal scale for height is defined

as a class:

<daml:Class rdf:ID="Height">

<daml:oneOf rdf:parseType="daml:collection">

<Height rdf:ID="short"/>

<Height rdf:ID="medium"/>

<Height rdf:ID="tall"/>



Clearly, such coding is tedious, but not too diff
icult because the logic and syntax are clear and
simple. However, an ontology editor with a GUI interface, and a DAML output will be
quicker, and probably less error prone.