Project modeling using iterative variable defect forecasts
AssigneeInternational Business Machines Corporation
InventorsJonathan Dunne, Paul H. McCann, Jeffrey B. Sloyer, Ian D. Stewart
Life of this patent
Estimating the term…
Abstract
Project modeling is conducted using variable defect arrival rate or variable defect rate density parameters. These defect rates may be updated on an iteration by iteration basis and may be used to provide remediation and further project modeling, remediation, and prediction.
Claims20
A computer system configured to forecast project status using variable defect rates, the system comprising:
- a module running on a processor, the module configured to repeatedly generate a defect trajectory model for a software project using a set of parameters representative of the software project, the set of parameters including one or more of the following:
- lines of code, available development resources, project duration, and a variable defect rate, wherein the variable defect rate is variable upon one or more iterations when generating the defect trajectory model;
- a module running on a processor, the module configured to perform functional and system testing for code defects, code maturity and code complexity of respective programming code;
- a module running on a processor, the module configured to infer when the software project will reach virtual zero defects using the defect trajectory model;
- and
- a defect repository in communication with one or more modules over a network.
A plurality of computing devices for assessing variable defect arrival rates, one or more of the devices configured to perform one or more steps comprising:
- at a processor, analyzing lines of code of a software product as an input parameter for a predictive forecaster to determine when a defect arrival rate of a software project will reach zero;
- at a processor, using a predictive forecaster to take point and time analytics from defect repositories and test management systems to determine where project testing currently is relative to an arrival rate model;
- at a processor, using a predictive forecaster to forecast how far ahead or behind a software project is in terms of a period of time;
- and
- calculating a time frame by using an arrival rate backlog plus the amount and type of testing left to complete and by estimating how many additional defects are predicted to be added to an open queue.
A computer program product comprising:
- a non-transitory computer readable medium that stores computer processor instructions, which when executed configure one or more computer processors to perform steps comprising:
- in a first iterative cycle, generate a defect trajectory model, the defect trajectory model including a predictive project model and forecasted project model, the forecasted project model generated using measured project parameters or reported project parameters, and a variable defect rates, the variable defect rate indicating a predicted amount of future faults in programming code in an iterative cycle subsequent to the first iterative cycle;
- and
- in a second iterative cycle, generating a revised defect trajectory model, the revised defect trajectory model including a forecasted project model updated form the first iterative cycle and generated using measured project parameters or reported project parameters, and a variable defect rates, the variable defect rate updated from the first iterative cycle and revised to predict an amount of future faults in programming code in an iterative cycle subsequent to the second iterative cycle.
Dependencies are inferred from the claim text.