Software Productivity Measurement Using Multiple Size Measures
Recommend Documents
ICEAA International Training Week. October 17-20, 2016, Bristol (UK) ... and authors. ICEAA Bristol (UK), Oct. 2016. 2 ... X Technology-dependent, no standards.
Oct 20, 2016 - Sizing method options: ⢠Lines of code: X Can't estimate until software designed. X Technology-dependent, no standards. ⢠Functional size.
Keywords: software engineering, software size measures, ISO/IEC stan- dards .... of the so-called Functional Size Measurement (FSM) of software products.
Insourced and Outsourced Projects. Addison Wesley Professional (2007). 18. Standish Group: CHAOS Summary 2009. West Yarmouth, Massachusetts (2009)
A leading German global automotive supplier has applied COSMIC sizing for change requests to. Electronic Control Unit (ECU) software. In this case, the sizing ...
Rehabilitation Research Center of Muscle-Skeletal, School of Rehabilitation Sciences, Ahvaz Jundishapur University of Medical. Sciences, Ahvaz, Iran.
GRP is a product of the reach of the advertisement and the average of its distribution of exposures delivered to a target audience. GRP is a direct measure of ...
Abstract The design, control and optimization of software engineering processes generally require the determination of performance measures such as.
plots were measured with tape measure, and the recent LSMS Integrated Surveys on Agriculture. (ISA) in Ethiopia, Malawi, Nigeria, Niger, and Tanzania, which ...
the distance between the input and prototype grows, the response falls off exponentially towards 0. The shape of the RBF neuron's response is a bell curve, as.
a strategy phase, a mapping phase, and a measurement phase. ...... Montpellier, France, June 18-20 2008b, Z. Bellahsene, C. Woo, E. Hunt, X. Franch AND R.
Aug 17, 2011 - The size of software project teams has been considered to be a driver of ..... benchmarking information to companies submitting data to the ...
pooled to create a single aggregate measurement .... plete their research, make a decision, and then visit a store ....
In the U.S. National Income and Product Accounts (NIPA) produced by the ... hours paid but not worked for reasons such as annual leave, sick leave, public holidays, ..... survey of how people living in the United States spend their time.
Maintenance has a broader scope of configuration management with more operational ..... If it is found that software maintenance does not account for close to.
May 1995. A DoD Software Measurement Pilot: Applying the SEI Core Measures.
James A. Rozum. William A. Florae. Software Process Program. Accesion For.
Keywords: Productivity Measurement, Imperfect Competition, Productivity Dispersion, ... that assume perfect competition and constant returns to scale.
Analysis â A Case Study for a Mobile Application. VolNan Tunalı ... the actual size of the project after development, and results were presented. Index Terms ... be performed, anticipating time, budget, and quality constraints. Estimation and ...
The Productivity Commission is the Australian Government's independent research and advisory body on a range ..... accou
Others, such as Mark ... different in Mark II Function Points. Mark II. Function Points define a different set of .... Garmus and Herron [10] and Lokan [20] both note.
AbstractâAs indicated by earlier researches, software products behave in a manner that is almost similar to the process of human ageing. Just as human beings ...
Sep 16, 2016 - Greater marketing of software. ⢠More complex systems ... Green software metric survey [P.Bozzelli, Q.Gu And P.Lago, 2013] ... [1] C. Seo, S. Malek, and N.Medvidovic. ... The recommended metric(s) becomes the next metric.
process and modern software development in order to ... acquisition and development capability maturity. ... submitted proposals have to be evaluated. It is clear ...
perspective of embedded software system families in the automotive domain. ... the development costs of process oriented SSF in embedded control units of ...
Software Productivity Measurement Using Multiple Size Measures
Introduction |Productivity - Simple ratio of product size to project effort - Broadly used to benchmarking - Difficult to measure No standard model for aggregating these measure Productivity =
Amount of output Unit of input used
Software Engineering Lab, KAIST
3/20
Background |Research trend - Common assumptions Size must be related to effort Reused components should not be included
- Productivity measure method Used single size measures • FP, SLOC, System Meter, Magnitude, Use Case Point • Standard size/effort ratio
Approach on multidimensional size measure
Software Engineering Lab, KAIST
4/20
Related work (1/4) |Reifer’s work [IEEE SOFTWARE, 2000] - Estimating quick-to-market software - Size of web application Using Halstead’s equation V* = Nlog2(n) = (N1* + N2*) log2 (n1* + n2*)
distinct number
Volume Volume total number Software Engineering Lab, KAIST
5/20
Related work (2/4) |Reifer’s work (cont’d) - Web based predictors of size
Software Engineering Lab, KAIST
6/20
Related work (3/4) |Stensrud and Myrtveit’s work [TSE’03] - Identifying high performance ERP projects Productivity is an indicator The smallest project is always the most productive
- Relative measure Variable returns to scale (VRS) P=
y x
x=
1 P
y
B
P: productivity x: effort y: FP or SLOC B > 1 (COCOMO)
Software Engineering Lab, KAIST
Effort = a × Size b 7/20
Related work (4/4) |Stensrud and Myrtveit’s work (cont’d) increasing return to scale (small project) decreasing return to scale (large project) constant return to scale (medium project)
Software Engineering Lab, KAIST
8/20
Motivation |Mendes’s work [METRICS’03] - Early web size measures and effort prediction for web costimation (133 companies) From Tukutuku database Company-specific dataset Multi-company dataset
Case-based reasoning
best?
Stepwise regression
Software Engineering Lab, KAIST
9/20
Measurement model (1/3) |Considerations - Finding elements related to efforts Web Web pages pages
High High effort effort functions functions
New New images images
- Combinations of size measures Multiplicative parameters (non-linear)
- Obtaining the total effort Person-hours • No information about the proportion of total effort spent constructing new elements and the proportion of total effort spent on reused elements
Software Engineering Lab, KAIST
10/20
Measurement model (2/3) |Assumptions - Effort is related to several different size measure - Size-based effort estimation model The expected value of productivity is one A value greater than one is indicative of good productivity (Expected effort)
Software Engineering Lab, KAIST
11/20
Measurement model (3/3) |Assumptions (cont’d) - Economies and diseconomies of scale should be account for