Ontsluiting ISBSG gegevens met QSM
-
Upload
nesma -
Category
Technology
-
view
195 -
download
4
description
Transcript of Ontsluiting ISBSG gegevens met QSM
Ontsluiting ISBSG gegevens mbv QSM
Werkgroep BenchmarkingBaarn, 12 april 2007
NESMA&
ISBSG
12 april 2007 2
Agenda
Nesma en ISBSG-
ISBSG benchmark-
QSM-
Ontsluiten ISBSG via QSM
Theo Kersten
12 april 2007 3
Bestuur
ISBSGIFPUG
BenchmarkingTelrichtlijnen …
GeïnteresseerdenLeden
werkgroepen
12 april 2007 4
� het uitwisselen van ervaringen met benchmarking van applicatie ontwikkeling en applicatie beheer om zo het vakgebied te professionaliseren � het periodiek organiseren van een workshop
Prestaties meten en begroten waarin een belangrijke rol is weggelegd voor benchmarking in het algemeen en ISBSG in het bijzonder.
Nesma Werkgroep Opdracht
12 april 2007 5
� Dhr. A.J.E. Dekkers, Shell Information Technology International BV � Dhr. H.S. van Heeringen, Sogeti Nederland B.V. � Dhr. F. Gerwig, Ordina � Dhr. ir. T.A.G.M. Kersten, ATOS Origin (voorzitter) � Dhr. P. de Ruijter, ABN AMRO Bank N.V. � Dhr. drs. M.P.E. Pereboom, Mediaan/abs bv � Dhr. W. Visser, Capgemini B.V.
Nesma Werkgroep Samenstelling
12 april 2007 6
ISBSG
THE INTERNATIONAL THE INTERNATIONAL SOFTWARE BENCHMARKING SOFTWARE BENCHMARKING
STANDARDS GROUPSTANDARDS GROUP
......or ‘Icebags’......or ‘Icebags’
12 april 2007 7
Doelstelling ISBSG
“… to improve the global understanding of software
practices so that they may be better managed.”
12 april 2007 8
ISBSG leden
� ASMA / SQA (NSW) – Australian Software Metrics Association / Software Quality Association Definieren benchmark standards (+tool)
� CSPIU (China Software Process Improvement Union)� FiSMA (Finnish Software Measurement Association)� DASMA (Deutschsprachige Anwendergruppe fur Software Metrik and Aufwandschatzung)� NASSCOM (National Association of Software & Service Companies)� GUFPI - ISMA (Gruppo Utenti Function Point Italia - Italian Software Metrics Association) � JFPUG (Japan Function Point User Group)� KOSMA (Korea Software Measurement Association)� NESMA (Nederlandse Software Metrieken Gebruikers Associatie)� AEMES (Asociacion Espanola de Metricas de Software)� SWISMA (Swiss Software and Service Metrics Association)� UKSMA (UK Software Metrics Association)� IFPUG (International Function Point Users Group)
12 april 2007 9
Software project & Software beheer
Productie platformProductie platform
Gebruik& Beheer
OntwikkelOntwikkel--procesproces
resourcesresources productproduct
Ontwikkel platformOntwikkel platform
Team Prog.taal Type-project Levertijd Bedrijfstak HardwareUren Hardware Omvang (FP/LOC) Kosten Aantal gebruikers Oper.syst.Ervaring Oper.syst. Methode Fouten Aantal locatie DBMSTijd DBMS Technieken & Tools Herstelkosten Aantal BU’s Prog.TaalBudget Ervaring Type Tevredenheid
Planning, Planning, Rapportage, Evaluatie & Besturing
12 april 2007 10
plan
test
specify
designbuild
SoftwareProject
Maintenance &
Support
Database Maintenance
Application Maintenance
implement
Welk project: Project afbakening
Support
12 april 2007 11
Vragen in Product Questionnaire
InspanningInspanning DD OmvangOmvang FF
Organisatie A
Product E
Project B/C
OpleveringOplevering GG
V 1V 1--66
V 7V 7--5252 V 53V 53--5959
V 60 V 60 -- 8282 V 89 V 89 -- 118118 V 119 V 119 -- 134134
V 83V 83--8888
12 april 2007 12
Vragen in Questionnaire
� Meer vragen beter voor analyse� Weinig vragen goed voor gebruiker
� Evenwicht zoeken tussen aantal vragen
� Geen minimale set gemaakt� Wel verplichte vragen: vetgedrukt
� Echt minimum: omvang en uren
12 april 2007 13
Project (B)
� Type projectNieuw
Aanpassing
Herontwerp
� Project fasen– Plan, Specify, Build, Test, Implement
– Per fase opgeleverde producten
12 april 2007 14
Project: Technologie C
� Operating systeem
� Programmeertaal
� Database
– Primair en niet primair
12 april 2007 15
Mensen en inspanning D
� Ontwikkelteam en omvang� Samenstelling ontwikkelteam� Gebruikersomgeving
- Bank, Regering, Medisch etc� Validatievragen
Maar ook kennis en ervaring wordt gevraagd:� Teamleden met materiekennis� Teamleden met ontwikkelomgeving� Ervaring projectleiding
12 april 2007 16
Oplevering G
� Productiedatum� Doorlooptijd� Aantal fouten
– Minor– Major– Extreme
� Klanttevredenheid� Kosten per fase
– Daarbij aangeven welke mensen meegeteld zijn Software ontwikkelteam, Gebruikersteam, Operators
� Validatievragen
12 april 2007 17
Vergelijken
� Productiviteit: Uren per fp per– Ontwikkeltype– Ontwikkelomgeving– Programmeertaal– Type organisatie– Type applicatie
� Opgeleverde kwaliteit: Aantal defects per 1000 fp
– Ontwikkeltype– Ontwikkelomgeving– Omvang– Type organisatie– Methode
12 april 2007 18
Productiviteit per categoriePresentatievorm (uren / fp)
N Min P10 P25 Median P75 P90 Max Mean Std Dev
categorie aantal laagste hoogste gem.
projecten waarde waarde waarde
mainframe
cobol 70 0.9 3.0 5.9 11.8 18.0 21.6 48.1 13.3 9.9
7.37.3
12 april 2007 19
KwaliteitPresentatievorm (fouten /1000 fp)
1e maand productie
Je project Min Mean Max
Extreme
defects 0 0 0 6
Major
defects 4 1 5 120
Minor
defects 19 1 16 300
12 april 2007 20
NESMA & ISBSG, producten
� Producten & diensten– Project profile report gratis
– Application Benchmark report gratis
– Benchmark report R8 € 70,- / € 100,-
– Repository CD R10 € 800,- / € 1200,-
– Project Estimation Toolkit € 90,- / €. 100,-
– Compendium € 90,- / € 110,-
– ISBSG Special Analysis Reports € 0,- / € 25,-
– NESMA workshop € 200,- / € 350,-
Bestellen [email protected] web www.nesma.nl
12 april 2007 21
QSM
Standaard
Parameters Productiviteit
� FP� Uren Uren/fp
QSM
Parameters Productiviteit
� Elocs� ManMaanden PI� Doorlooptijd
12 april 2007 22
QSM Uren per FP neemt toe bij tijdsdruk
0,00
10,00
20,00
30,00
40,00
50,00
60,00
0 2 4 6 8 10 12 14 16 18 20
Doorlooptijd (maanden)
Uren
/Fun
ktiepu
nt
250
500
750
1000
1250
1500
1750
2000
2500
3000
12 april 2007 23
QSM
3 basis parameters samenhang
� Elocs Elocs� ManMaanden PI� Doorlooptijd (maanden) MBI
� PI Productivty Index (productviteitsmaat)� MBI: Manpower Buildup Index (tijdsdruk)
Daarnaast is invloed op kwaliteit belangrijk� Number of defects
12 april 2007 24
QSM effort, size and elapsed time: Rayleigh Curve
Avg Staff (people)<Single Goal - MBI 3,6041>
1 2 3 4 5 6 7 8 9 10jan'07
feb mrt apr mei jun jul aug sep okt nov dec
0
2
4
6
8
Avg S
taff (people)
109876543
TO+RENzorg
Project: MMSP
12 april 2007 25
QSM products based on the QSM method
� SLIM Datamanager (de database)� SLIM Metrics (voor analyses)� SLIM Estimate (voor het maken van alternatieve
plannen)� SLIM Control (voor projectbewaking)
12 april 2007 26
SLIM Datamanager
12 april 2007 27
SLIM Estimate S ta f f in g & P r o b a b il ity An a lys is
A v g S ta f f ( p e o p le )<O v e r a ll>
1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 2 0 2 1 2 2o k t' 0 4
n o v d e c ja n'0 5
fe b mr t a p r me i ju n ju l a u g s e p o k t n o v d e c ja n' 0 6
fe b mr t a p r me i ju n ju l a u g s e p
0 .0
2 .5
5 .0
7 .5
1 0 .0
1 2 .5
1 5 .0
1 7 .5
Avg S
taff (people)
10964321
R &DC &TP_ Mn t
M i l es tones 1 - S RR 2 - H LDR 3 - L LDR 4 - C UT 6 - S T C 9 - 9 9R 10 - 9 9 .9R
M i l es tones 1 - S RR 2 - H LDR 3 - L LDR 4 - C UT 6 - S T C 9 - 9 9R 10 - 9 9 .9R
SO L UT IO N P A NE L <O ve r a l l>
D ura t i o nE ffo rtC os t
P eak S ta ffM T T D
S tart Da te
C&T10.913
163814.51 .1
6 -3 -2005
L ife Cyc le21 .019
231914 .547 .0
1 -11-2004
M onth sMHR (K )E U R (K )peopl eDay s
P I=15 .0 M B I=3 .1 E ff F P =575
R ISK GA U GE < O v e r a ll>
< N o c o ns t r a in t s >
D u r a tio nEffo r t
Pe a k S ta ffQ u a li ty
% 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0
C O N TR O L P AN E L <O v e r a l l >
P I 1 2 .0 1 8 .0
1 5 .0
P e a k S ta f f 1 1 . 6 1 7 . 4
1 4 . 5
E f f F P 4 6 0 6 9 0
5 7 5
P ro j ec t : B CDB
12 april 2007 28
SLIM Estimate Realisatie v s SWH historie en QSM gemiddelde
PI v s Ef f ectiv e SLOC
0,1 1 10 100 1.000
Ef f ectiv e SLOC (thousands)
-5
0
5
10
15
20
25
30
PI
MBI v s Ef f ectiv e SLOC
0,1 1 10 100 1.000
Ef f ectiv e SLOC (thousands)
-2
2
6
10
14
MB
I
TO+RE Duration (Months) v s Ef f ectiv e SLOC
0,1 1 10 100 1.000
Ef f ectiv e SLOC (thousands)
0,1
1
10
100
TO
+RE
Duration (M
onths)
TO+RE Ef f ort (PM) v s Ef f ectiv e SLOC
0,1 1 10 100 1.000
Ef f ectiv e SLOC (thousands)
0,1
1
10
100
1.000
10.000
TO
+RE
Effort (P
M)
Current Solution Historical Projects QSM 2005 BUSINESS GOVT GRP Av g. Line Sty le1 Sigma Line Sty le Project: MMSP
12 april 2007 29
SLIM controlStop Light Overview
FTE Staff
3 6 9 12 15 18 21 24 27 3019-10'02
09-11 30-11 21-12 11-01'03
01-02 22-02 15-03 05-04 26-04 17-05
0
2
4
6
8
10
12
ppl
765432Cum Cost Life Cycle
3 6 9 12 15 18 21 24 27 3019-10'02
09-11 30-11 21-12 11-01'03
01-02 22-02 15-03 05-04 26-04 17-05
0
100
200
300
400
500
NLG
(thousands)
765432
Cumulative Product Construction
3 6 9 12 15 18 21 24 27 3019-10'02
09-11 30-11 21-12 11-01'03
01-02 22-02 15-03 05-04 26-04 17-05
0
5
10
15
20
25
SLO
C (thousands)
765432Defects Found
3 6 9 12 15 18 21 24 27 3019-10'02
09-11 30-11 21-12 11-01'03
01-02 22-02 15-03 05-04 26-04 17-05
0
10
20
30
40
50
60
Defects
765432
Current Plan Actuals Current Forecast Green Control Bound Yel low Control Bound Project: Number Redesign
12 april 2007 30
Ontsluiting ISBSG via QSM
Parameters voor QSM wijken af van paramaters ISBSG
� QSM richt zich sterk op Main Build fase� ISBSG richt zich op alle projectfasen
– In geleverde ISBSG db staan genormaliseerde uren naar “volledig project”
� ISBSG bevat voornamelijk FP� QSM gaat in principe uit van elocs (effective) lines of code� QSM heeft doorlooptijd Main Build nodig� ISBSG werkt in uren QSM werkt met maanden
Daarom gezamenlijk keuzes gemaakt voor vertaalslag
12 april 2007 31
Ontsluiting ISBSG via QSM
Filter criteria:� Betrouwbare projecten� Size Measure: IFPUG of Nesma� Projecten met doorlooptijd > 3 maand� Omvang < 7.000 fp� Resource level = 1
� En dan de conversie slagen o.a. voor fp naar elocs
� Het resultaat is via een api opgenomen in SLIM DM
12 april 2007 32
Tot slot
� U kunt de waarde van de benchmark vergroten
� U kunt invloed uitoefenen op de benchmark standaards
Door mee te doen!
Vragen?