brochure software technology projects 2007
TRANSCRIPT
-
8/14/2019 Brochure Software Technology Projects 2007
1/18
-
8/14/2019 Brochure Software Technology Projects 2007
2/18
3TU.School or Technological Design,
Stan Ackermans Institute oers eleven
two-year postgraduate technological designer
programmes. This institute is a joint initiative
o the three technological universities o the
Netherlands: Delt University o Technology,
Eindhoven University o Technology andUniversity o Twente. For more inormation
please visit: www.3tu.nl/sai.
5 Piotr opatk
7 Bart Golsteij
9 Athanasios C
11 Iulia Dobai;
13 Yuchen Zhan
15 Iulia-Madali
17 Charalampos
19 Jiang Zhang;
21 Panja Sae-Ui
23 Leila Rahma
25 Chetan Nair;
27 Joost Meijles
29 Zan Li; Video
31 Xing Li; Gene
Sotware Technolog
-
8/14/2019 Brochure Software Technology Projects 2007
3/18
It is a great pleasur
out by the PDEng ca
rom 1 December 20
programme particip
organisations, there
This booklet contain
overview o the bro
PDEng candidates a
candidates have con
technology to impro
and quality. In doin
MADE IN HOLLAND.
Made in Holla
Martin Rem2
-
8/14/2019 Brochure Software Technology Projects 2007
4/18
ChallengesProjects at Philips Research study wireless sensor networks.
These networks consist o tiny devices to be used in
healthcare and ambient intelligence applications. To speed
up the development process, tools are necessary to provide
better insight into the behaviour o applications on the
sensor nodes. Considering the nature o the sensor networks,
handheld devices and radio communication should be used.
ResultsA system was designed to give developers monitoring and
wireless control o sensor applications using a tablet PC.
Analysis showed that it was necessary to separate networkapplications logically. A service was added to the sensor node
that enables easy development o multiple network applica-
tions.
BenetsApplication developers can use the implemented system to
check the operational properties o the operating system
running on the node. These properties include clock
synchronisation and task inormation. It is possible to add
new properties to the system.
Wireless sensor netw
networking capabili
The potential bene
developed at Philip
Basic inrastructur
Basic inrastructure
system, synchronise
on the sensor nodes
embedded systems,
or network monito
power, so they cann
remote monitoring
Network service to
Attempting to proto
a good strategy or
project. In our case
includes a network
enable retrieval and
multiplexer and dem
The monitoring and
rom the sensor nod
Handheld cont
or wireless seHis analysis showed thata drastic reorganisation othe sotware architecturewas neededPiotrs work was compli-cated by the constantevolution o the hardwareplatorm on which the sotware was developed.
P.D.V. van der Stok,Philips Research
Piotr opatka4
-
8/14/2019 Brochure Software Technology Projects 2007
5/18
The AutoDJ Improve
o Philips Applied T
algorithms, and so
like a human DJ, no
The mixes produced
moment, the beats
one song to the nex
the gure above.
Producing DJ-like
To be able to produ
about the tempi, th
produce DJ-like mix
ware available that
A rst AutoDJ imple
this implementatio
on the results o thi
implementation.
AutoDJ Improv
the beat goes
Bart Golsteijn
Barts work has yieldedan improved AutoDJsotware implementationthat will allow PhilipsApplied Technologies tomuch better serve ourcustomers.
P. Dillen,Philips AppliedTechnologies
ChallengesA rst version o AutoDJ was available at the start o this
project. As a result o some (unctionally driven) design
choices, the playback part o that implementation was
dicult to understand and maintain, and dicult to change
or extend. In addition, the system as a whole was rather
infexible and dicult to use. The goal o the AutoDJ
Improvement project was to solve these problems by
redesigning and reimplementing part o the system.
ResultsThe problems with the playback part o the system have
been solved by developing a completely new generic systemor mixing songs. The usability and fexibility problems
have been solved by redesigning and reimplementing the
top-level architecture and the public API o the system.
BenetsA solid oundation has been laid to develop AutoDJ urther
into a commercial-quality product. A wider range o mixes
is now supported, and the mixes can be more complex. A
number o limitations o the old system have been removed
and the generic nature o the newly developed mix model
provides ample opportunities or urther experiments and
improvements. In addition, redesign and reimplementation
o the top-level architecture and the public API have made
AutoDJ more fexible, easier to change and extend, and yet
easier to use and maintain.
6
-
8/14/2019 Brochure Software Technology Projects 2007
6/18
ChallengesProvide insight into the requirements and easibility aspects
o video hardware prototyping within a sotware-oriented
group. Design a re-usable prototyping ramework on FPGA
that will enable an earlier start o video sotware
development, and thereore earlier time to market.
Achieve maximum re-usability o existing components
and architectures.
ResultsAn analysis o the requirements, risks and points o attention
that a sotware group must be aware o to carry out a hard-
ware prototyping project. To gain hands-on experience and tosupport our claims, we went through all the steps o designing
on FPGAs coming up with a proo-o-concept prototype, a
hardware emulation inrastructure implemented on an FPGA
platorm. The ramework enables ast integration o new video
processing hardware IPs, providing all the adaptation layers
required or abstracting rom the NXP-specic protocols to the
FPGA interaces and IOs.
BenetsThe developed prototype, although in an immature state,
is the rst attempt o developing a re-usable and scalable
hardware emulation ramework within and or the needs o a
sotware group. The group can use it as a base or a powerul
FPGA prototyping ramework to expand the capabilities o
the hardware simulation environment or in similar projects.
The documentation produced during this project can also help
in decision-making and risk management in uture projects.
Finally this project enabled the group to gain valuable know-
ledge o FPGA prototyping, as well as o tools and methods or
IC design and integration.
The Nexperia Home
the TV application d
is large and consistcomponents that im
Dependency betwe
The video processin
and hardware devel
early stages o deve
to-market, there is
actual hardware bec
simulation ramewo
limited debugging c
ramework which en
hardware platorm.
simulator with the
its capabilities.
The Video Hardware
within a sotware-o
answered crucial qu
lack o hardware co
o-the-shel protot
valuable insight int
revealing potential
Flexible prototypin
The resulting produ
compatible video pr
as well as by on-chi
re-usable FPGA prot
development projec
Video HW protprogrammableHis enthusiasm and driveincreased with every new
problem encountered.Many problems wereencountered, tackled andresolved, taking largedevelopment eorts.It paid o, ater ninemonths o hard work wenow know better what ittakes to realise an FPGAprototype using o-the- shel hardware.
K. Brink,NXP Semiconductors
Athanasios Chliopa8
-
8/14/2019 Brochure Software Technology Projects 2007
7/18
ChallengesThe project aimed at providing the stakeholders with a
solution that would include a new way to develop experience
demonstrators (based on a high-level denition language) in
a more intuitive way, with a methodology that would allow
greater productivity and easier incorporation o change.
ResultsBesides developing an implementation or the proposed
solution and two experience demonstrators to prove its
easibility, the project provided insight into the process o
speciying user interactions, raised many questions relative
to possible other applications o the approach, and answereda ew questions relative to its advantages and disadvantages.
Ater considering the advantages and disadvantages, we
conclude that the Scenario Engine solution provides good
support or rapid application prototyping and better
maintainability. From a usability point o view, i the initial
learning and understanding phases are overcome, the
developers could reach higher productivity and eciency
levels.
BenetsPrototyping and development o experience demonstrators
play a undamental role in the product development liecycle.
The Scenario Engine solution supports the process by
allowing aster demonstrators or evaluation o the inter-
action concepts, as well as by delivering a set o compact
system requirements.
The ALCOR project a
concepts within me
demonstrators show
images, 2D sets o i
prototyping context
Developing a rapid
The ALCOR Scenario
environment or exp
The approach invest
reasoning mechanis
the Scenario Engine
language). During r
actions should be in
This approach was u
simple rom an inte
solution in the exist
demonstrator. A car
one developed with
o the disadvantage
Managing UI cexperience de
Iulia Dobai
The result o Iulias workis a well documented andworking piece o sotwarethat defnitely will beused in the near uture othe Alcor project.
G.-J. Bloem,Philips AppliedTechnologies
10
-
8/14/2019 Brochure Software Technology Projects 2007
8/18
ChallengesThe initial challenge o this project was to enable
perormance exploitation and optimisation o an existing
image/video processing architecture by analysing and
characterising the perormance characteristics o the
sotware ramework and the underlying platorm.
ResultsThe model o the data fow in the ramework shows a
problematical design choice in the sotware architecture,
which has been proved by experiments to be the main actor
that imposes the bottleneck on the perormance. To remove
this bottleneck, an improved architecture has been proposed.The perormance analysis and exploitation on the platorm
through a set o macro-experiments provides an inventory
o potentials and pitalls in such a platorm that typically
determine/disturb the perormance o the image/video
applications.
BenetsThe perormance investigation and characterisation on
the ramework gives a concrete example or analysing
perormance on the image/video processing systems. Such
experience can be useul or perormance analysis o other
applications, typically or applications with a similar
architecture. The knowledge and ndings rom the platorm
exploitation can help the image/video algorithm developers
to estimate the perormance o their applications, and
should also be taken into account in making perormance
relevant design decisions. At the organisational level, the
resulting knowledge and ndings contribute to the database
o platorm knowledge, which is crucial or projects to
develop perormance-aware real-time image/video
applications.
In the Video Signal
been developed on
provides fexibility
o image/video data
produces a smooth
two original rames
reconstruction algo
ultrasound. Howeve
ramework can guar
applications with re
Methodology and t
The perormance inv
analysing perorman
methodology and te
latency/throughput
viewpoint. The mod
architecture, which
An improved archite
The perormance an
provided a list o po
perormance o GPU
image/video algorit
taken into account
Perormance a
heterogeneouHe was very resourceulin his approach, and hecame up with resh ideasand directions we wouldnot have thought o.
J. van de Ven,Philips AppliedTechnologies
Yuchen Zhang12
-
8/14/2019 Brochure Software Technology Projects 2007
9/18
ChallengesOne o the challenges was how to create a solution that
accurately helps clinicians in easily and automatically
diagnosing AAA diseases. To achieve this, parts o existing
sotware had to be re-used. This raised another challenge:
how to build the new sotware on the existing one.
ResultsThe result o the project consists o a prototype which
measures and monitors the AAA geometry based on 3D CTA
and MRI data. The tests perormed on the prototype show
that the computation o the geometrical parameters
(which describe the AAA geometry) is accurate. In addition,recommendations are made or urther improvement o
the prototype to allow a useul solution to be oered to
clinicians.
BenetsBeore introducing a product to the market, a prototype
should rst be clinically evaluated. The developed proto-
type gives Philips Medical Systems the possibility to show
clinicians an application which measures and monitors
the AAA geometry over time, based on MRI and CTA data.
A clinical evaluation will then show to what extent the
prototype meets the clinicians needs.
MeMo Measureme
Abdominal aortic an
on peoples lives. Th
The rupture o an an
AAA rupture can lea
Improving diagnos
One o the current g
improve the diagno
monitored using tw
is limited. During th
hic Angiography (CT
can allow more accu
To achieve the goal
monitoring tool was
perorms landmark d
computation o geo
visualisation and re
Measuring and mo
The prototype as de
AAA geometry. It th
application which m
A clinical evaluatio
Ater successul clin
Monitoring ab
aneurysm geo
Iulia-Madalina Iota
The prototype sotwareapplication that Madahas developed perectlymatches our expectations!The prototype orms anexcellent basis or utureclinical evaluation and orpotential uture productdevelopment.
M. Breeuwer,Philips Medical Systems
14
-
8/14/2019 Brochure Software Technology Projects 2007
10/18
ChallengesThe major challenge we had to deal with stemmed rom
the considerable fuidity and vast diversity pervading the
establishment o interoperability in healthcare. Ocial
specications or connection to the healthcare IT inra-
structure are either incomplete or still undergoing a long
process o maturing. The two countries we examined
Germany and the Netherlands although neighbouring,
have applied dierent standards and technologies in
their national healthcare inrastructures. Furthermore,
standardisation committees issue their guidelines, while
groups o experts are keen to promote their own solutions.
In the light o these developments, the underlying challenge
proved to be the derivation o an architecture that wouldenable connectivity with the national healthcare inrastruc-
tures, while maintaining a loose coupling with their intrinsic
eatures.
ResultsThe results o the project include an extensive documented
analysis o interoperability in the healthcare domain, high-
lighting the various alternatives, along with a proposed
architecture and design.
Furthermore, we developed a prototype to show the creation
and transmission o a reerral note to the Catharina Hospital,
Eindhoven, and the reception o the corresponding discharge
note, in accordance with the requirements o the Dutch
healthcare inrastructure.
BenetsPhilips Applied Technologies Digital Systems & Technologies
has enriched its inormation inventory with knowledge about
the status, trends, and alternatives in the domain o inter-
operability in healthcare. It also has a tangible demonstrator
in cooperation with a prestigious healthcare institute, the
Catharina Hospital, which it can use or promotional purposes
to potential customers.
Integrating Philips
In recent studies, h
or chronic heart aito monitor patients
acilitating two-way
Engaging patients
Philips has develope
daily, personalised
term behaviour cha
Industry by storm, w
Meanwhile, demand
healthcare provider
lands, specic gove
establishing interop
that could standard
support the intercon
Integration into na
Philips Applied Tech
envisaged integrati
rst steps was there
such as the reerral
The developed demo
and the Health Leve
and concepts to ena
demonstrator, we la
exchange o inorm
Integration inhealthcare inThe work o Charalamposresulted in ..., and better
understanding o thedomain, not only or himbut also or me andprobably others as well.
D. La Hei,Philips AppliedTechnologies
Charalampos Xanth16
-
8/14/2019 Brochure Software Technology Projects 2007
11/18
ChallengesThe project was motivated by the need or highly dynamic,
rapidly recongurable, domain-specic modelling o
mechatronics applications. Having developed successul
systems in the past, Philips Applied Technologies wanted
to nd an ecient and eective way o developing and
modelling new mechatronics systems by re-using components
o legacy products, even in binary orm. In addition, Philips
Applied Technologies wanted to have a domain-specic
language and environment or the modelling o component-
based mechatronics applications.
ResultsThe project resulted in a domain metamodel, which consistso premodelled mechatronics components that contain domain
knowledge. The domain metamodel allows the architect to
design applications by re-using and integrating components
identied rom legacy applications. The project also involved
the design and development o a new graphical modelling
language and environment, which includes new notation and
semantics specically or the modelling o component-based
applications in the mechatronics domain.
BenetsMDE raises the level o abstraction in all aspects o sotware
development throughout its liecycle by making models
primary artiacts. Our component-based MDE approach
especially brings benets such as the division o labour,
because components can be predened by one or a ew
domain experts. In addition, the design models made by
our tools can generate codes or deploying and integrating
binary-orm legacy components. That signicantly improves
developer productivity and code quality. The new domain-
specic modelling language supports higher-level abstrac-
tions than General-purpose MDE, and requires less eort
and ewer low-level details to design an application in the
domain in question.
The project involved
Model-Driven Engin
research project, Ph
ware development o
real-time sotware p
Applied Technologi
o this metamodel.
Designing motion c
The MDE metamode
premodelled mecha
metamodel and des
graphical modelling
In particular, the pr
modelling o motion
was developed withmechatronics sotw
This project succeed
in investigating the
component-based s
Mechatronics
Model-Driven
Jiang Zhang
It is clear that not manydevelopers would becapable o bringing sucha project to a successulend. We are thereore verygrateul to Jiang Zhangor the work that he hasdone
M. Beekveld,Philips AppliedTechnologies
18
-
8/14/2019 Brochure Software Technology Projects 2007
12/18
ChallengesThe challenge was to integrate ConTest into a real-time
embedded system. The characteristics o embedded systems
are very dierent rom those o enterprise sotware systems
like IBM WebSphere. The integration o ConTest into the TV
sotware thereore included a number o modications to
port ConTest to the resource-constrained and platorm-
specic TV environment. As the program timing o a real-time
system is designed accurately or its limited resources,
it can also be problematical to add an additional program to
run in such a system without impairing its real-time
properties. Moreover, the act that our case study was not a
small real-time system but a large industrial system made
the project even more challenging.
ResultsC-ConTest was successully integrated into the TV sotware
stack. Based on the current evaluation results, C-ConTest still
has many points o improvement in the real-time embedded
system domain in terms o its eectiveness in revealing
and eliminating concurrency bugs. Moreover, although the
implementation o C-ConTest adheres to the POSIX standard,
this does not mean much in the real-time embedded system
domain where each company is likely to have its own
predened basic unctionalities. C-ConTest cannot assume
what resources would be available in the target system.
BenetsThis project made the technical issues in integrating a test
tooling into the TV sotware stack clear to Philips Consumer
Electronics (Philips CE). The same integration approach can be
re used when Philips CE wants to integrate other tooling into
its sotware. The project also allowed IBM to gain more insight
into technical issues in the real-time embedded system
domain, and to improve ConTest or this domain. Moreover,
the project established the cooperation between Philips CE
and IBM. In the near uture, Philips CE can expect improve-
ments o ConTest in many aspects, and it could turn out to
be a potential concurrency-test tooling or Philips CE in the
uture.
Testing a multi-thre
with each other and
same test may have
is discovered it is v
Philips CE oresees d
to investigate bette
Forcing concurrenc
ConTest, a concurre
means to orce conc
systematically and t
conditions, deadloc
blems early in the t
ConTest has been us
has never been used
eectiveness o Con
Successul integra
C-ConTest, the C ver
with the pthread lib
project, one o the
C-ConTest was split
latter was integrate
as possible to avoid
TV sotware stack. A
The evaluation o th
embedded system d
Feasibility ana
ConTest or emHis drive or results andsolid way o working keptthe project on track... thetechnical work was donein a proessional way.No cutting corners to getit working, but creatingproduction-qualityimplementations.
C. Aarts,Philips ConsumerElectronics
Panja Sae-Ui20
-
8/14/2019 Brochure Software Technology Projects 2007
13/18
ChallengesExtending the architecture o devices that can collaborate
to ull specic user requests by building an ad hoc network
o unctionalities. The extension includes eatures such as
learning capabilities between the devices, and context-
awareness. The architecture should be validated by a working
prototype.
Results Architecture o collaborative devices that can learn
rom each other and are context-aware.
Knowledge base design and generic sotware design that
ull the architecture. A prototype proo o concept that implements this
architecture using UPnP technology and development
boards based on PNX0106.
BenetsThe result o this project was a proo o a new concept
or semiconductor devices, in which these devices can auto-
nomously collaborate to ull user requests. The prototype
showed that these devices can auto-congure themselves
as new devices come along, enabling the user to evolve
the network by buying small devices which add new
unctionalities to the network. The prototype also showed
the easibility o context-awareness in the concept.
The Advanced Syste
ahead with the aim
tests them with pro
meet these concept
As a step towards t
called Collaborative
unctionalities and
provides the user w
were a single device
Developing the Col
In this project, we w
For example, when t
this new device and
Another aspect is th
as relaxing, sleepinrequest according t
We have extended t
technology. By emb
themselves as each
to dierent context
dierent devices.
Ontology-base
context-aware
Leila Rahman
Leila has contributedsignifcantly towardsthe goal o realising anetwork o autonomousdevices. We are veryhappy with the results,and urther work willbe undertaken based onLeilas results.
A. Syed,NXP Semiconductors
22
-
8/14/2019 Brochure Software Technology Projects 2007
14/18
ChallengesA model-based ramework or awareness in an embedded
system may only have minimal impact on the existing
(possibly real-time) behaviour o the system. This imposes
many design constraints, or example limited processing
power, memory and bandwidth. Another challenge is the
accuracy o error detection, which is infuenced by actors
such as correctness o implementation o the model,
proper timing o measurement and comparison o behaviour,
and variability in the acceptable behaviour o the system
(determining what is a deviation and what is not).
ResultsA ramework or model-based awareness was designed. Itallows quick experimentation with dierent models, and is
congurable to suit dierent strategies or detecting and
reporting errors. An Awareness Monitor was implemented on
Linux using this ramework, and was shown at the TRADER
Demo Day. TRADER research partners are using th e ramework
in experiments in monitoring MPlayer (a popular open-source
media player application). The Awareness Monitor can be
ported to embedded platorms such as a TV with relatively
low eort.
BenetsThe researchers at ESI now have a ramework available or
conducting experiments on model-based awareness. They have
increased knowledge o topics in model-based awareness such
as modelling behaviour and detecting and reporting errors.
The research partners o TRADER are using the rameworkto aid their research. Trials o model-based awareness on
embedded systems such as TVs can be started using the
ramework. The ramework is easily used or black box
testing or test-time improvements in reliability.
Reliability is a prim
printers. Trends in e
cycles demand bett
TRADER project at t
reliability o consum
Model-based aware
Model-based awaren
awareness makes a
desired behaviour o
as an error. An Awa
Subsequently, the a
recover the erroneo
not technically erro
ESI researchers inve
systems (e.g. in a Ta ramework to expe
Developing an awa
In this project, an a
behavioural models
Linux to monitor a L
quick experimentat
or detecting and re
allow reporting rele
perormed, initial p
on the system is re
Chetans work has alsotriggered new researchon the simulation o theawareness ramework:we expect substantialimpact, both scientifcallyand industrially, in thenear uture.
T. Hendriks,Embedded SystemsInstitute
Chetan Nair
A ramework
model-based e
24
-
8/14/2019 Brochure Software Technology Projects 2007
15/18
ChallengesCreating a ramework that supports the creation o tangible
interaction consoles and enabling the creation o dierent
applications or those consoles was a major challenge. The
ramework combines a simple run-time environment with the
possibility o speciying complex behaviour, while at the same
time oering fexibility and ease o use, all in one system.
In addition, embedded hardware constraints had to be met to
allow the run-time environment to be deployed.
ResultsA ramework or ast, easy creation and execution o
autonomous tangible sensor-based applications was created.The ramework enables the specication o tangible inter-
action consoles, and orms a basis on which dierent types
o tailored user interaces can be built, to allow a variety o
users (rom children to manuacturers) to create applications
or the consoles.
BenetsThe ramework is considered to be an important step towards
moving the creation o autonomous tangible sensor-based
applications rom technologically savvy people to those who
are more oriented towards toy and game design. In addition,
the ramework speeds and simplies the creation o tangible
sensor-based applications or all types o users.
The project was car
studied since early
story-telling enviroThis platorm is now
The rst target o ES
applications such a
o these application
dierent (wireless)
o the platorm is e
Fast, easy creation
The objective behin
tangible application
sensor programming
people who are capa
to build un applic
sensor applications
savvy and more des
Those problems are
The design environm
application to the raccording to the ap
number o sensors a
is known as a tangi
o users, the system
abstraction layer.
To validate the syst
StoryToy is a single-
matching game (as
autonomous tangib
Edutainment sapplication ra
Joost Meijles
Joost did his workdiligently and ollowing
an organised and welldocumented process.It gives us the evidenceand confdence that thefrst version o theplatorm is already agood and useul one.In all, a ormidable jobwell done.
Willem Fontijn,Philips Research
26
-
8/14/2019 Brochure Software Technology Projects 2007
16/18
ChallengesThere were three main challenges:
Exploration o state-o-the-art video processing algorithms
which should not only give pleasing visual results but also
should be commercially easible or use in the product.
Real-time constraints in sotware development.
Analysing and realising the synchronisation mechanism.
ResultsA prototype that contains three sotware video processing
algorithms was designed and implemented. A fexible
architecture which denes clear video algorithms, interaces
and control hardware on the board was successully
implemented.
BenetsThe VI2B delivered the sotware which is able to control the
hardware and integrate three new advanced video processing
algorithms. The buer manager scheme is designed
and implemented. One o the processing algorithms gives
impressive results. This is probably commercially viable
or use in the product i the computation complexity
demands on the h ardware are greatly reduced.
NXP Semiconductor
integrated circuits (
The video group o t
business line digita
Signicant diere
The diversity o vide
their quality (resolu
have increased in si
resolution displays
quality video source
the perceived low q
Nowadays, the key e
This project aims at
resolution displays
existing state-o-thNXP Semiconductors
Flexible architectu
The VI2B applicatio
The proposed soluti
denes clear sotwa
type has been desig
Video Improve
In-A-Box (VI2Zan Li was able to showher skills in (advanced)video processing,and ater one weekendo hard work she showedthe frst results o herimplementation oan edge-directedupscaling algorithm.
K. Brink,NXP Semiconductors
Zan Li28
-
8/14/2019 Brochure Software Technology Projects 2007
17/18
ChallengesOne o the challenges o this project was to dene a repre-
sentation o a programming language which acts as an inter-
mediate media in projects o code analysis in LaQuSo. This
representation should be generic so that it is easible to map
any programming language into it, and the metrics and the
diagrams that LaQuSo needs can be extracted rom it. Another
challenge was to generalise the mapping rom specic
language(s) to the dened representation, as well as metrics
and diagram extraction rom the representation so that deve-
lopers will have guidelines to map another language to the
representation and extract other kinds o metrics or diagrams
rom it. In addition, the extraction system should be re-used
in the projects o code analysis or dierent languages.
ResultsThe Graph eXchange Language (GXL) ormat was used as the
generic representation according to LaQuSo requirements.
Furthermore, a fexible metamodel was dened to accom-
modate a broad range o p rogramming languages and metrics.
Two sotware projects, which were written individually in
Java and MicroFocus COBOL, were successully mapped into
GXL. Additionally, 13 kinds o metrics ranging rom the
number o operators to Depth o Inheritance Tree (DIT) were
extracted rom the GXL representation. The extracted metrics
were veried. Experiments showed that the extraction system
is re-usable. Above all, the mapping and extraction guidelines
were also proposed.
BenetsGiven the act that as long as a programming language can
be mapped to GXL, a uniorm way can be employed to extract
metrics and diagrams, the time spent on code analysis or di-
erent languages can be reduced with respect to the metrics
and the diagram extraction. Moreover, i GXL is used with a
re-usable extraction system to acilitate code analysis, the
only eort that should in principle be required is or mapping
a programming language to a GXL representation. Case studies
showed that the time spent on mapping was between one
and two weeks, depending on the complexity o the language
grammar.
This project was per
analyse the quality
Diagrams reer to cl
Metrics are normally
and diagrams, the q
The sotware that c
LaQuSo does or the
intermediate repres
a dierent language
methods are built to
on extracting metri
Dening a generic
To tackle that probl
representation, rom
intended to be re-ucan thereore be sav
tion together with a
is in principle requi
The most important
any programming la
research groups. On
were used as case s
tation and extractin
processes to obtain
designed during the
Generic interm
representation
Xing Li
LaQuSo is now ableto perorm source codeanalysis, metrics andmodel extraction or new(legacy) languages moreefciently.
M. van den Brand,TU/e LaQuSo
30
-
8/14/2019 Brochure Software Technology Projects 2007
18/18
Piotr opatka; Hand
Bart Golsteijn; AutoAthanasios Chliopa
programmable hardw
o experience demon
analysis or heteroge
Vatau; Monitoring a
Charalampos Xanth
healthcare inrastruc
meta-model or mod
Feasibility analysis oLeila Rahman; Onto
Chetan Nair; A ram
Joost Meijles; Eduta
ramework. Zan Li; V
Xing Li; Generic Int
Edited by:
H.T.G. Weers
Coordinator Sotware Technology
Eindhoven University o Technology
Text:
PDEng candidates o the Sotware Technology programme
Production:
Lian van Hout communicatie, Eindhoven
Translation:
Nigel Morrison, Eindhoven
Photography:
Rien Meulman, Eindhoven
Design:
vanRixtelvanderPut ontwerpers, Eindhoven
Printing:Drukkerij Lecturis, Eindhoven
Credits32