Project Metrics Research Articles

BMR: Benchmarking Metrics Recommender for Personnel issues in Software Development Projects

This paper presents an architecture which applies document similarity measures to the documentation produced during the phases of software development in order to generate recommendations of process and people metrics for similar projects. The application makes a judgment of similarity of the Service Provision Offer (SPO) document of a new proposed project to a collection of Project History Documents (PHD), stored in a repository of unstructured texts. The process is carried out in three stages: firstly, clustering of the Offer document with the set of PHDs which are most similar to it; this provides the initial indication of whether similar previous projects exist, and signifies similarity. Secondly, determination of which PHD in the set is most comparable with the Offer document, based on various parameters: project effort, project duration (time), project resources (members/size of team), costs, and sector(s) involved, indicating comparability of projects. The comparable parameters are extracted using ...

Thompson isometries of the space of invertible positive operators

We determine the structure of bijective isometries of the set of all invertible positive operators on a Hilbert space equipped with the Thompson metric or the Hilbert projective metric.

Projective metrics and contraction principles for complex cones

In this article, we consider linearly convex complex cones in complex Banach spaces and we define a new projective metric on these cones. Compared to the hyperbolic gauge of Rugh, it has the advantages of being explicit and easier to estimate. We prove that this metric also satisfies a contraction principle similar to Birkhoff's theorem for the Hilbert metric. We are thus able to improve existing results on spectral gaps for complex matrices. Finally, we compare the contraction principles for the hyperbolic gauge and our metric on particular cones, including complexification of Birkhoff cones. It appears that the contraction principles for our metric and the hyperbolic gauge occur simultaneously on these cones. However, we get better contraction rates with our metric.

Cyclic projectors and separation theorems in idempotent convex geometry

Semimodules over idempotent semirings like the max-plus or tropical semiring have much in common with convex cones. This analogy is particularly apparent in the case of subsemimodules of the n-fold Cartesian product of the max-plus semiring: It is known that one can separate a vector from a closed subsemimodule that does not contain it. Here we establish a more general separation theorem, which applies to any finite collection of closed subsemimodules with a trivial intersection. The proof of this theorem involves specific nonlinear operators, called here cyclic projectors on idempotent semimodules. These are analogues of the cyclic nearest-point projections known in convex analysis. We obtain a theorem that characterizes the spectrum of cyclic projectors on idempotent semimodules in terms of a suitable extension of Hilbert's projective metric. We also deduce as a corollary of our main results the idempotent analogue of Helly's theorem.

Employing use cases to early estimate effort with simpler metrics

In order to calculate size and to estimate effort in applications, the standard method most usually used is function points, which has been used with good results in the development of industrial software for some time. However, some aspects should be improved, namely: the time at which the estimation of effort is performed and the margin of error in the effort estimation. Consequently, another size metric which could be used to obtain more accurate estimations should be found. This article presents two other size metrics for projects based on use cases: transactions and entity objects. Effort is estimated by using the technique mean productivity value. There is also a description of two case studies, one which involved four academic projects and the other one which involved four industrial projects. They were developed in order to compare the estimations obtained with each method. The result shows that the current way of estimating effort can be improved by using the number of transactions as a size metric and the technique mean productivity value.

Lightweight Software Metrics: The P10 Framework

Software metrics programs must provide value. Metrics need attachment to what's important about a software project. A thorough and comprehensive way to help ensure that connectedness is the Goal Question Metric. GQM proceeds in an orderly fashion to consider goals and specific questions so that when the metrics are defined they relate directly to project goals. The P10 software metrics framework better reflects the contemporary software development environment. P10 is a simple way to help ensure that the software project metrics you use will provide answers to the important questions you face in creating on-time, high-quality software

Using a role scheme to derive software project metrics

Roles' playing is common in our lives. We play different roles with our family, at work as well as in other environments. Role allocation in software development projects is also accepted though it may be implemented differently by different software development methods. In a previous work [Y. Dubinsky, O. Hazzan, Roles in agile software development teams, in: 5th International Conference on Extreme Programming and Agile Processes in Software Engineering, 2004, pp. 157-165] we have found that personal roles may raise teammates' personal accountability while maintaining the essence of the software development method. In this paper we present our role scheme, elaborate on its implementation and explain how it can be used to derive metrics. We illustrate our ideas by data gathered in student projects in the university.

Measuring performance: Moving NASA Earth science products into the mainstream user community

Abstract Demonstrating performance of the applications of Earth observation satellite-based science data products and services is increasingly a requirement of government research agencies. We present efforts from the NASA-funded Earth Observing System Data and Information System's Synergy Project to measure performance in the development of applications from NASA research and development projects. We summarize challenges in monitoring performance and share our experience in evolving metrics over a 5-year project life. We demonstrate how to adapt project management processes and metrics from the information technology (IT) industry to Earth observation applications research and development. A roadmap for adapting IT processes and developing metrics and examples of quantitative and qualitative metrics are provided. Our findings suggest that designing and implementing these IT metrics will enhance project success, as defined by the degree of penetration of NASA products into the user community and level of non-NASA funding secured.

Projective spaces with Cayley-Klein metrics

Projective metrics were first introduced by A. Cayley and F. Klein who constructed analytical models over the field of complex numbers. The aim of this paper is to give for the first time a purely synthetic definition of all projective spaces with Cayley-Klein metrics and to develop the synthetic foundation of projective-metric geometry to a level of generality including metrics over arbitrary fields of characteristic ≠2.

Auto-induced semantic classes

Advanced computer dialogue agents contain a natural language understanding component that requires knowledge of semantic classes and concepts. These are frequently generated manually for new tasks. We avoid this time-consuming procedure by using a two-step unsupervised clustering process. First, a semantic generalizer automatically induces semantic classes using training data from well-studied applications (domains) for which large transcribed corpora of human–human dialogues are available. Candidate word pairs are grouped into similar semantic groups according to the similarity of their lexical bigram contexts. We show that the proposed algorithms for automatically inducing semantic classes perform very well for typical spoken dialogue applications. We exceed 90% precision for the first 100 cluster assignments for narrowly defined tasks such as a movie information task. For a heterogeneous task such as the text-based WSJ, a precision of only 24% is increased to 73% by including context thresholding and part-of-speech tagging. Second, we determine the degree of domain independence for each class by using concept comparison and projection metrics to rank order semantic classes by degree of domain independence.

Value creation and capture: a model of the software development process

Landmark Graphics supplies software and services to the upstream oil and gas industry. Our software portfolio, which ranges from exploration and drilling to data management and decision analysis, includes more than 60 products consisting of over 50 million lines of source code. For many years, Landmark has been collecting project metrics we wished to harvest to gain insight into key business questions in three areas: optimal release cycle duration (scope/time trade-off), optimal project staffing levels, effects of uncertainty. We set out to develop a relatively simple project dynamics model to use in conjunction with market sensitivity and economic analysis to help optimize profitability. Some of our ideas and results are similar to those of Preston Smith and Donald Reinertsen, who examined the impact of time-to-market sensitivity. However, our approach is a more detailed model tuned to software development issues.

The Perron-Frobenius theorem for homogeneous, monotone functions

If A A is a nonnegative matrix whose associated directed graph is strongly connected, the Perron-Frobenius theorem asserts that A A has an eigenvector in the positive cone, ( R + ) n (\mathbb R^{+})^n . We associate a directed graph to any homogeneous, monotone function, f : ( R + ) n → ( R + ) n f: (\mathbb R^{+})^n \rightarrow (\mathbb R^{+})^n , and show that if the graph is strongly connected, then f f has a (nonlinear) eigenvector in ( R + ) n (\mathbb R^{+})^n . Several results in the literature emerge as corollaries. Our methods show that the Perron-Frobenius theorem is “really” about the boundedness of invariant subsets in the Hilbert projective metric. They lead to further existence results and open problems.

Stability and uniform approximation of nonlinear filters using the Hilbert metric and application to particle filters

We study the stability of the optimal filter w.r.t. its initial condition and w.r.t. the model for the hidden state and the observations in a general hidden Markov model, using the Hilbert projective metric. These stability results are then used to prove, under some mixing assumption, the uniform convergence to the optimal filter of several particle filters, such as the interacting particle filter and some other original particle filters.

Projectively flat Finsler metrics of constant flag curvature

Finsler metrics on an open subset in R n with straight geodesics are said to be projective. It is known that the flag curvature of any projective Finsler metric is a scalar function of tangent vectors (the flag curvature must be a constant if it is Riemannian). In this paper, we discuss the classification problem on projective Finsler metrics of constant flag curvature. We express them by a Taylor expansion or an algebraic formula. Many examples constructed in this paper can be used as models in Finsler geometry.

Project Management in the Information Systems and Information Technologies Industries

For many enterprises, sustainable success is closely linked to information systems (IS) and information technologies (IT). Despite significant efforts to improve software project success, many still fail. Current literature indicates that most of the software project problems are related to management, organizational, human, and cultural issues—not technical problems. This paper presents results of a survey of 36 software owners/sponsors, contractors/suppliers, and consultants on 12 projects. The empirical results address answers to questions related to success, performance metrics, and project business drivers. A lack of alignment on these critical issues emerge consistently by phase as well as across the entire project. The results of this study also are compared with others that span seven additional industry sectors. As a result, the authors have developed an approach that links project critical success factors (CSFs) to corporate strategy, and project metrics to the CSFs. An important finding of this study is the critical need to identify and manage realistic expectations of the stakeholders to achieve perceived project success.

Collecting and analyzing Web-based project metrics

Metrics generated from real project data provide a valuable tool for evaluating and guiding performance. The authors describe their experience developing a Web-based project metrics system and encouraging its adoption. They also examine how such a system might encourage the adoption of metrics within a company.

Projective metrics and mixing properties on towers

Projective metrics and mixing properties on towers

Information System for Risks, Costs, and Benefits of Infrastructure Improvement Projects

Several approaches to performance assessment for selecting projects for highway infrastructure improvement are synthesized. In a planning cycle, highway agency divisions face many diverse candidates for maintenance and improvements, originating from public demand, accident data, engineering judgment, and local officials. Resources are limited, and selection factors to consider include crash risk reduction, travel time savings, economic development, environmental protection, aesthetics, and capital and maintenance costs. This article describes the adoption of quantitative metrics for some factors, their estimation with varying precision for about 30 actual projects, graphical project comparison, and a related application of multiple factors to interchange design. Identified needs include to (a) select appropriate metrics for diverse projects; (b) coordinate information systems; (c) avoid preassigned factor weighting, enabling juxtaposition of quantitative and qualitative factors and invitation of newfactors in a deliberation; (d) interpret public agency planning for citizens and government officials; (e) appreciate any tort liability of generating risk-related data; and (f) address the impacts on saturated, urban networks.

Inside Intel - coping with complex projects

In this paper, the authors describe how team-based project management processes and metrics enable design teams at Intel to plan, execute and evaluate performance on complex projects, on a quarterly basis.

Software project control and metrics

With the aim of providing scientific guidance for the application of diamond cutting tools to the machining of stone-plastic composite, this work presents results on the influence of tool geometry and cutting parameters on cutting forces and temperature during helical milling of stone–plastic composite with diamond cutters. Four factors—helical angle, spindle speed, feed rate, and cutting depth—were assessed using a response surface method. Mathematical models were developed and identified by verification testing to accurately predict changes in cutting forces and temperature during composite helical milling. Then, the significant contributions of each factor and of two-factor interactions were determined by analysis of variance, and the trends of cutting forces and temperature were studied using response surface methodology. The optimal conditions in terms of low cutting forces and temperature were determined to be a helical angle of 70°, cutting speed of 51.3 m/s, feed per tooth of 0.24 mm, and cutting depth of 0.5 mm. These parameters are proposed for use in the industrial production of stone–plastic composite material to improve machining efficiency.