Loc Line
Cost Estimating Software
Abstract
Software cost estimation is the process of prediction the effort required to develop a software system. This document provides an overview of methods for estimating costs of software including recent advances in the field. As some of these models are based on an estimate of software size as input, first give an overview of metrics common size. Then highlight the cost estimation models have been proposed and used successfully. Models can be classified into 2 categories main algorithmic and non-algorithmic. Each has its own strengths and weaknesses. A key factor in selecting a cost estimation model is the accuracy of its estimates. Unfortunately, despite the vast amount of experience with models to estimate the accuracy of these models is not satisfactory. The document includes comments on the results of model estimation and description of the various new
approaches to cost estimation.
Words Keywords: project estimation, effort estimation, cost models.
Introduction 1.
In recent years, software has become the most expensive component of projects computer systems. Most of the costs of software development is due to human effort and most cost estimation methods focus on this look and give an estimate in terms of person-months.
Exact estimates of software costs are essential to both the developers and customers. They can be used to generate requests for proposals, contract negotiations, scheduling, monitoring and control. Underestimating the costs may be in the discharge of the proposed systems, which then exceed their budgets, with functions of underdeveloped and of poor quality, and failure to complete on time. It can lead to an overestimation in too many resources committed to the project, or, during the tender at the result of not getting the contract, which can lead the loss of jobs
.
Exact cost estimate is important because:
- You can help sort and prioritize development projects with respect to a global business plan.
- It can be used to determine what resources to commit to the project and how well these resources will be used.
- It can be used to assess the impact of changes and re-support.
- Projects can be more easy to manage and control resources are better matched to actual needs.
- Customers expect the actual development costs to be consistent with estimated costs.
- Estimated costs of the program involves the determination of one or more of the following estimates:
- effort (usually in person-months)
- project duration (in calendar time)
- cost (in dollars)
Most models cost estimation attempt to generate an estimate of the effort, which can be
become the project duration and cost. If While the effort and cost are closely related, are not necessarily related by a simple conversion function. Effort is often measured in person-months programmers, analysts and project managers. This effort estimate can be converted into a dollar cost figure of calculating an average wage per unit of time of personnel involved, and then multiplying this by the effort that is deemed necessary.
The professionals have struggled with three main issues:
- What model of cost estimating software to use?
- What size to use software metering – Lines of code (LOC), function points (FP), or the point of function?
- What is a good estimate?
The widely practiced method of cost estimation is the opinion of experts. For many years, the draft
Managers have been based on experience and current industry standards as a basis for developing cost estimates. However, estimates based on expert opinion is problematic:
- This approach is not repeatable and the means to obtain an estimate are not explicit.
- It is difficult to find estimates great experience for each new project.
- The relationship between cost and system size is not linear. Cost tends to increase
exponentially with size. The method of expert opinion is appropriate only when the size of the current project and previous similar projects.
- Manipulations management budget designed to prevent overflow and make the experience of previous projects of questionable data.
In recent three decades, many quantitative models of software cost estimation have been developed. Ranging from empirical models such as Boehm's COCOMO model analysis models. An empirical model uses data from previous projects to assess the current project stems from the basic formulas of analysis special database available. An analytical model, by contrast, uses formulas based on global assumptions, such as the speed at which the developer to solve problems and the number of available problems.
Most cost models are based on the size measure, such as LOC and FP, obtained from size
estimate. The accuracy of the estimated size of a direct impact on the accuracy of the estimate of costs. Although common-size measurements have their own drawbacks, an organization can make good use of anybody as long as a consistent method of counting used.
A good estimate of the costs of software should have the following attributes:
- It is designed and supported by the project manager and development team.
- It is accepted by all stakeholders as ongoing.
- It is based on a clear set of software model cost with a credible basis.
- It is based on a database of relevant project experience (similar processes, similar
technologies, similar environments, similar people and similar requirements).
- Defined in sufficient detail to understand their key areas of risk and probability Success is evaluated objectively.
Estimating software costs has historically been a major difficulty in developing software.
Several reasons for the difficulty have been identified:
- The lack of a historical database of the measurement expenditure
- Software development involving many interrelated factors that affect development activities and productivity, and whose relationships are not well understood
- Lack of training of estimators and estimators with the necessary knowledge
- Penalty Little is often associated with a poor estimate
2. Estimation process
Cost estimation is an important part of the planning process. For example, the top-down approach to planning, cost estimation is used to get the project plan:
- The project manager develops a characterization of the overall functionality, size, process, environment, people, and quality required for the project.
- A macro-level of total effort estimate and schedule is developed using a software cost estimation model.
- Partitions project manager of effort estimation at a higher level work breakdown structure. Also, the partitions in the calendar of dates of important events and determines a profile staff, which together form a project plan.
The current cost estimation process involves seven steps
- Set Cost estimation objectives
- Generate a project plan for the required data and resources
- Specify software requirements
- Work as much detail about the software system as possible
- Using several techniques of independent cost estimate to build on its strengths combined
- To compare different estimates and repeat the estimation process
- After initiating the project, monitor its actual cost and progress and performance feedback to project management
No matter the model estimation is selected, users must pay Please note the following for best results:
- coverage estimate (in some models to generate the effort throughout the life cycle, while others do not include the effort for the requirement of stage)
- calibration and model assumptions
- sensitivity of parameter estimates different model
- deviation of the estimate for the actual cost
3. Software size
The software size is the most important factor affecting the cost of software. In this section
describes five software size metrics used in practice. The line of code and function point metrics are most popular among the five indicators.
Line of code: This is the number of lines of source code of the software delivery, excluding of comments and blank lines and is commonly known as LOC. Although LOC is the programming language dependent, is the most widely used software size metric. Most models relate this measure with the cost of software. However, exact LOC can be obtained only after the project is completed. Estimating the Size code of a program before it is built really is almost as hard as estimating the cost of the program.
A typical method for estimating the size the code is using the opinion of experts, along with a
technique called PERT. This is the view of experts of three possible code formats: Sl, the lowest possible size; Sh highest possible size, and Sm, the size most likely. The estimate of the size code S is calculated as:
S = S1 + SH 4 Sm
6
PERT can also be used for individual components to obtain an estimate of the software system by adding up the estimates of all components.
Science Software: Halstead proposed the code length and volume metrics. length code is used to measure the length of program source code and is defined as:
N = n1 + n2
where N1 is the total number case of the operator, and N2 is the total number of operational cases.
Volume is the amount of storage space required and is defined as:
V = log N (n1 + n2)
where n1 is the number of different operators, and n2 is the number of operands that appear in a different schedule.
Function Points: This is a measure based on the functionality of the program and first introduced by Albrecht. The total number of points depends on the function of the different (in terms of format and processing logic) into the following five categories:
- The user input types: data or control user-input types
- The user output types: types of output data for the user leaves the system
- The types of research: the items that require an interactive response
- Inside types of files: files (logical groups of information) that is used and shared within the system
- The types of external files: files that are passed or shared between the system and another system. Each of these types is individually assigned one of three levels of complexity (1 = simple, 2 = Medium, 3 = complex) and gives a weighting value that varies from 3 (for single entry) to 15 (for the complex internal files).
The unadjusted function point counts (CFU) is given as
5 3
UFC =?? NW
i = 1 j = 1 ij ij
where N ij and W ij are respectively the number and weight of the types of class i, with j complexity.
For example, if the raw function point counts of a single project are 2 inputs (W ij = 3), 2
complex outputs (W ij = 7) and 1 complex internal file (W ij = 15). Then UFC = 2 * 3 + 2 * 7 +1 * 15 = 35.
This initial function point count is directly used to estimate costs or further modified by factors whose values depend on the overall complexity of the project. , Will be taken into account the degree of distributed processing, the amount of reuse, performance requirement, etc. The final function point count is the product of the UFC and these factors, project complexity. The advantage of function point measurement is that it can obtained on the basis of system requirements specification early in software development.
The UFC is also used to estimate the code – the average size of the following linear formula:
LOC = a + b * UFC
The parameters A, B can be obtained by linear regression and data from previously completed project. The final function point counting practices manual remains by IFPUG (International Function Point Users Group).
Extensions of function point: the point feature expands the role of the points to include algorithms as a new class. An algorithm is defined as the set of rules to be fully expressed to solve a significant computational problem. For example, a square root routine can be considered as an algorithm. Each algorithm used is given a weight from 1 (elementary) to 10 (sophisticated algorithms) and the characteristic point is the weighted sum function algorithms points. Applications This is especially useful for systems with few input / output and high algorithmic complexity, such as mathematical software, discrete simulations, and military.
Another extension of function points is the full function point (FFP) to measure in real time
applications, taking into account the control aspect of such applications. FFP control data introduces two new types operating and four new types of transaction control function.
Points of view: while point features and FFP enlarge point function, the action item under the size of a different dimension. This measurement is based on the number and complexity of the following objects: screens, reports and 3GL components. Each of these objects account and gives a weight of 1 (single screen) to 10 (3GL component) and the point object is the weighted sum of all these objects. This is a relatively new measure and has not been popular. But because it is easy to use in the initial phase of the development cycle and also the measures reasonably sized software, this measure has been used in models to estimate more important, like COCOMO II cost estimate.
4. Cost Estimate
There are two main types of cost estimation methods: algorithmic and non-algorithmic.
Algorithmic models vary widely in mathematical sophistication. Some are based simple arithmetic formulas using summary statistics as means and standard deviations. Others are based on regression models and differential equations. To improve the accuracy of algorithmic models, there is a need to adjust or calibrate the model to local circumstances. These models can not be used outside the shelf. Even with the calibration accuracy can be quite different.
First, give an overview of algorithmic methods.
4.1 Methods not algorithmic
Analogy Cost: This method requires one or more completed projects that are similar to the new project and the estimate obtained through reasoning by analogy of the actual costs of previous projects. Estimating by analogy can be made either in the total project level or at subsystem. The total level the project has the advantage that all components of the system cost will be considered while the subsystem level, has the advantage of providing an assessment more detail the similarities and differences between the new project and completed projects. The strength of this method is that the estimate is based on a real project experience. However, it is unclear to what extent the previous project is actually representative of the constraints, the environment and the functions to be performed by the new system.
Expert opinion: This method involves consulting one or more experts. The experts provided estimates using their own methods and experience. Expert consensus mechanisms such as the Delphi technique or PERT is used to resolve the inconsistencies in the estimates. The technique Delphi is as follows:
1) The Coordinator presents each expert with a specification and a registration form estimates.
2) Each expert fills individually (without discussing with others) and is allowed to ask questions of the coordinator.
3) The coordinator prepares a summary of all expert estimates (including average or median) in an application form for another iteration of the expert estimates and the ratio of the estimates.
4) Repeat steps 2) -3) as many rounds as appropriate.
A modification of the Delphi technique proposed by Boehm and seems Fahquhar be more effective: Before the estimation, a meeting of the coordinating group of the expertise and is willing to discuss the problems of estimation. In the Step 3), experts do not need to give any reason for the estimates. Instead, after each round of the estimate, the coordinator called a meeting to points to discuss with experts in their estimates varied.
Parkinson: Parkinson Using principle "work expands to fill the available volume, "The cost is determined (not estimated) by available resources rather than relying on an objective assessment. If the software has to be delivered in 12 months and 5 people are available, the effort is estimated at 60 person-months. Although sometimes gives good estimation, this method is not recommended as it may provide unrealistic estimates. Furthermore, this method does not promote the good practice of software engineering.
Price-to-win: The cost of software is estimated to be the best price to win the project. The estimate is based on the client's budget instead of functionality software. For example, if a reasonable estimate for a project cost of 100 person-months, but the client can only afford 60 months-person, it is common that the estimator asked to modify the estimation of effort to accommodate person with 60 months to win the project. This again is a good practice because it is very likely to cause a late delivery or poor force the development team to work overtime.
From bottom to top: In this approach, each component software system is calculated separately and the results aggregated to produce an estimate for the global system. The requirement of this approach is that the initial design should be in place that indicates how the system is decomposed into different components.
From top to bottom: This is the opposite approach ascending method. An overall cost estimate for the system comes from the global properties, using no algorithms or algorithmic methods. The total cost can then be divided among the various components. This approach is more appropriate for estimating costs in the initial phase.
4.2 Algorithmic Methods
The methods of algorithms are based on mathematical models to produce estimates of the expenses based on a number of variables, which are considered the major cost factors. Any algorithmic model has the form:
Effort = F (x 1, x 2, …, xn)
, Where (x 1, x 2, …, xn) denote the cost factors. Current methods of algorithms differ in two aspects: the selection of cost factors, and the shape of the function f. First, we will discuss the cost factors used in these models, then characterize the models according to the shape of the functions and if the models of are analytical or empirical.
4.2.1 Cost factors
Besides the size of the software, there are many other cost factors. The broader set of cost factors are proposed and used by Boehm et al in the COCOMO model II. These cost factors can be divided into four types:
Product factors: the required reliability, the complexity of product, the size of the database used; required
reuse, documentation match to lifecycle needs;
Factors Equipment: limited run time, main memory constraint, the delivery of computer limitations, the volatility of the platform;
Personal factors: the ability of analyst, application experience, programming ability;
experience of the platform, language and tool experience, staff continuity;
Factors project: development of multiple sites, use of software tool, program development required.
The above factors are not necessarily independent, and most of them are hard to quantify. In many models, some of the factors appear in combination and some are simply ignored. In addition, some factors that take discrete values, resulting an estimation function with a rational piece.
4.2.2 Linear Models
Models have linear form:
n
Effort = a0 + ax
i = 1 ii
where the coefficients a 1, …, A n are chosen to best fit the data of the completed project. Nelson's work belongs to this type of models. We agree with the comment Boehm that there are too many nonlinear interactions in software development of a linear model to work well. "
4.2.3 multiplicative models
Multiplicative models have the form:
n XI
Effort = A0? R
i = 1 i
Again the coefficients a 1, …, n are chosen to best fit the data of the completed project. Walston-Felix used this type of model with each xi having only three possible values: -1, 0, 1. Doty model belongs to this class with each xi having only two possible values: 0, 1. These two models seem to be too restrictive in the values factor cost.
4.2.4 Power function models
The models of the power function has the general form:
Effort = a 'S b
where S is the code size, and a, b are (usually simple) the functions of other cost factors. This class contains two algorithmic models more popular in use, as follows:
COCOMO (Constructive Cost Model) model
This family of models was proposed by Boehm. The models have been widely accepted in practice. In COCOMO, the code size S is given in thousands of LOC (Kloc) and effort is in person-month.
A) Basic COCOMO. This model uses three sets of (a, b) based on software complexity only:
(1) to applications simple and well understood, 2.4, b = 1.05;
(2) to more complex systems, a 3.0, b = 1.15;
(3) for embedded systems, 3.6, b = 1.20.
The basic COCOMO model is simple and easy to use. As many cost factors are not
believes, can only be used as a rough estimate.
B) Detailed COCOMO and Intermediate COCOMO. The intermediate COCOMO, an estimate nominal effort is given by the power function with three sets of (a, b) with a coefficient slightly different from that of the basic COCOMO:
(1) to applications simple and well understood, a = 3.2, b = 1.05
(2) for complex systems, a 3.0, b = 1.15
(3) for embedded systems, from 2.8, b = 1.20
So cost factors fifteen years with values ranging from 0.7 to 1.66 (see Table 1) are determined [5]. The factor overall impact M obtained as the product of all individual factors, and estimating M is obtained by multiplying the nominal estimate.
Table 1: Factors of cost and weight in COCOMO II
Cost factors
Description
Rating
Very Low
low
nominal
High
Very high
Product
TRUST
required software reliability
0.75
0.88
1.00
1.15
1.40
DATA
size of the database
—
0.94
1.00
1.08
1.16
CPLX
product complexity
0.75
0.85
1.00
1.15
1.30
Computer
WEATHER
execution time constraint
—
—
1.00
1.11
1.30
Stor
Main storage constraint
—
—
1.00
1.06
1.21
VIRT
virtual machine volatility
—
0.87
1.00
1.15
1.30
TURN
computer response time
—
0.87
1.00
1.07
1.15
Staff
ACAP
analyst ability
1.46
1.19
1.00
0.86
0.71
AEXP
application experience
1.29
1.13
1.00
0.91
0.82
PCAP
Capacity
1.42
1.17
1.00
0.86
0.70
VEXP
Virtual machine experience
1.21
1.10
1.00
0.90
—
Lexpa
language experience
1.14
1.07
1.00
0.95
—
Project
MODP
modern programming practice
1.24
1.10
1.00
0.91
0.82
TOOL
software tools
1.24
1.10
1.00
0.91
0.83
SCED
development program
1.23
1.08
1.00
1.04
1.10
Meanwhile COCOMO basic and intermediate estimate the cost of the system software
level, detailed COCOMO works on each sub-system separately and has a clear
advantage for large systems containing heterogeneous subsystems.
C) COCOMO II. Perhaps the most significant difference of the models of COCOMO is that changes early exponent B according to the following cost factors:, precedentedness flexibility in the development, architecture or resolution risk, the team cohesion and process maturity. Other differences include the cost factors and added new models to consolidate and reduce software architecture risk.
Putnam and SLIM Model
Putnam is derived from their model based in Norden / Rayleigh distribution and its workforce found in the analysis of many completed projects. The central part of Putnam's model name equation software as follows:
S = E '(Effort) 1 / 3 t d 4 / 3
d where t is the time of delivery of software, E is the media factor environment that reflects the ability of development that can be derived from historical data using software
equation. S is the size the state of consciousness and the effort is in person-years. Another important relationfound Putnam is
Effort = D 0 'td 3
where D 0 is a parameter called manpower build-up ranging from 8 (completely new software with many interfaces) to 27 (rebuilt software). The combination of the above equation with the software equation, we obtain the power function form:
Effort = (D04 / 7 x E-9 / 7) X S9 / 7 and
td = (D0-1 / 7 x E-3 / 7) x S3 / 7
Putnam model is also widely used in practice and SLIM is a software tool based on this model for cost estimation and planning of manpower.
4.2.5 Calibration of the linear regression model
A direct application of the above models no local circumstances into account.
- However, one can adjust the cost factors using local data and the method linear regression. We illustrate this calibration of models using the general model of power function: Effort = a 'b S.
Take logarithm of both sides and Y = log (effort), A = log (a) and X = log (S). The formula is transformed into a linear equation:
XY = A + B '
Applying the standard method of least squares to a set of data from previous projects (Y i, x i: i = 1, …, k), we obtain the necessary parameters b and A (and therefore a) for the function power.
4.2.6 Discrete models
Discrete models have a way table, which usually relates the effort, duration, difficulty and cost of other factors. This class of models contains the model Aron, Wolverton model, and Boeing model. These models gained some popularity in the early days of the cost estimates as they were easy to use.
4.2.7 Other models
Many other existing models and the following have been used quite successfully in practice.
Price-S is the proprietary software cost estimation model developed and maintained by RCA, New Jersey. From an estimate of the size of project, the type and difficulty, the model calculates the project cost and schedule.
SoftCost relates the size, effort and duration to address the risk using a form of the Rayleigh probability distribution. It contains heuristics to guide the estimators in coping with new technologies and the complex relationships between the parameters involved.
Algorithmic models can be grouped as shown in Table 2.
Table 2. Classification of algorithmic models
Algorithmic models
Linear
Multiplicative
Function Power
Discrete
Other
Empirical
Nelson
Walston-Felix et al Herd
COCOMO
Aron Boeing Wolverton
Price-S
Analytical
Putnam
Soft Cost
Table 3 compares the strengths and weaknesses of different methods. From the comparison, we conclude that the
- No method is best for all projects.
- Parkinson and Price-to-win methods are not suitable for organizations that aim to win more business.
- Using a combination of techniques can provide the best estimate. For example, the combination of top-down estimating with expert advice and methods of analogy, can provide superior results.
Table 3. Summary of strengths and weaknesses of different methods
Methods
Strength
Weaknesses
No algorithmic
Case Experts
Experts with relevant experience can provide good estimate;
Depending on the "experts";
It may be prejudice;
They suffer from incomplete recall
Analogy
Based on actual project data and past experience
Similar projects can not exist;
Historical data may not be accurate
Parkinson's Price-to-win
Often with the contract
Poor practice;
You can have large cost overruns
From top to bottom
System-level approach;
Faster and easier than the bottom-up approach;
Requires a minimum detail of the project
Provide little detail to justify
estimates;
Less accurate than other methods
From the bottom up
Based on detailed analysis;
Project follow-up support better than another method, such as the address of its estimates of low-level tasks
They can ignore the factor level of the system cost;
They require more effort estimation
compared with the top down;
Difficult to estimate early life cycle
Algorithms
Objective, reproducible results;
Acquire a better understanding of the estimation method
Subjective inputs;
Adjusted for previous projects and may not reflect the current environment;
The algorithms may be specific to companies and not be suitable for software
development in general
4.3 Performance Measurement Model
Several researchers have used different error measurements. The most popular is the mean absolute relative error error (MARE):
n
MARE = a (| (estimatesi – updated) / updated |) / n
i = 1
where an estimated effort estimated from the model, real self is the true stress, n is the number of projects.
To establish whether the models are partial, the average relative error (MRE) can be used:
FM = = a (| (estimatesi – updated) / updated |) / n
A large positive MRE suggests that the model generally overestimates the effort, while a large negative value indicates otherwise.
The following criteria are can be used to evaluate the cost estimation models:
1) Definition – Have you clearly defined the cost model is the estimate, and excludes the costs?
2) Loyalty – Are the estimates closer to actual costs spent on the projects?
3) Objectivity – Does the model avoid allocating most of the variation in the cost of software to improperly calibrated subjective factors (such as the complexity)? Is it difficult to adjust the model to get any results that the user wants?
4) constructive – Can a user know why the model provides estimates it does? Does it help the user understand the software work can be done?
5) Detail – Is the model fit easily into the estimation of a software system composed of a number of subsystems and units? Does it give accurate phase and activity breakdowns?
6) Stability – Small differences in inputs produce small differences in estimates of production costs?
7) Scope – Does the cover model of the kind of software projects whose costs the user needs to estimate?
Ease of use – Are the model inputs and options easy to understand and specify?
9) Prospectiveness – Does the model avoid the use of the information will not be known until the project is complete?
10) The parsimony – Does the model avoid the use of highly redundant factors, or factors that make no significant contribution to the results?
5. Performance of estimation models
Many studies have attempted to assess the cost estimation models. Unfortunately, the results are not encouraging, since many of them turned out to be not very accurate. Kemerer out an empirical validation of four algorithmic models (SLIM, COCOMO, Estimacs, and FPA). No recalibration of the models was performed on data from the project, which was different from that used for model development. Most models showed a strong on the estimation bias and estimation errors overall, ranging from a mare from 57% to 800%.
Vicinanza, Mukhopadhyay and experts Prietula to estimate the project effort data from whole, without formal Kemerer algorithmic techniques and found that the results beat patterns in the original study. However, the tide goes from 32 to 1107%.
Ferens and Gurner evaluated three models (SPANS, Checkpoint, and COSTAR) with 22
projects Albrecht database and 14 projects from Kemerer data set. The estimation error is also large, with MARE ranging from 46% to the checkpoint model to 105% on cost model.
COCOMO Another study also found high error rates, an average of 166%. Jeffery and Low investigated the need for model calibration, both in industry and organizational levels. Without model calibration, the estimation error was large, with Mare, from 43% to 105%.
Jeffery Barnes and later low compared to SPQR/20 FPA model with data from 64 projects in a single organization. The models were adjusted for the local environment to eliminate the bias of estimation. Improved estimation was observed with a mare of 12%, reflecting the benefits model calibration.
There were also studies based on the use of analogy. Using a program called
Angel based on the minimization of the Euclidean distance in n-dimensional space, Shepperd and Schofield found that estimation by analogy, exceeded the estimate based on algorithms derived statistics.
Heemstra surveyed 364 organizations and found that only 51 models used to estimate effort and that users of the model did a better estimate than non-users of the model. Similarly, the use of estimation models was no better than expert opinions.
A study of software development within JPL found that only 7% of estimators of Use
algorithmic models as a primary focus of the estimate.
6. New approaches
Cost estimates remains a complex problem that continues to attract considerable research attention. Researchers have tried different approaches. Recently, models based on artificial intelligence techniques have been developed. For example, Finnie and Wittig applied artificial neural networks (ANN) and case-based reasoning (CBR) for the estimation of effort. Using a set of data from the Software Metrics Association of Australia, RNA is capable of estimating the development effort within 25% of real effort in more than 75% of the projects, and with a mare of less than 25%. However, CBR results were less encouraging. In 73% of cases, the estimates were 50% of actual effort, and 53% of cases, estimates were 25% of staff.
In a separate study, Mukhopadhyay, Prietula Vicinanza and found that an expert system based on the reasoning analog outperformed other methods.
Srinivasan and Fisher used machine learning approaches based on regression trees and neural networks to estimate costs. Learning approaches have been found to be competitive with SLIM, COCOMO and function points, compared with the previous study by Kemerer. One of the main advantages of learning systems is that they are adaptable and nonparametric.
Briand, El Emam, and proposed a hybrid method Bomarius cost models, COBRA: Cost
estimation, benchmarking and risk analysis. This method relies on knowledge of quantitative data experts and the project of a small number of projects. Encouraging results reported a small data set.
About the Author
CURRICULUM VITAE
Objectives: -
Develop the desktop and Distributed Database Applications through the
application tools like :
C/C++, Visual Basic (Database Connectivity With Oracle or SQL Server2000,Ms-Access,Crystal Report)
Java, HTML, DHTML, Visual C++, ASP (Active Server Pages), VbScript, JavaScript.
Personal: – E-Mail: nomaniba@yahoo.com
Name: Muhammad Noman Siddiqui
F/Name: Muhammad Saeed Siddiqui
D.O.B: 13th June 1980
Sex: Male
Marital Status: Married
Nationality: Pakistani
Address: B-6/229 Indus Mehran Society,
Malir Extension Karachi.Pakistan
P.O.Box: 75080
Cell No: 0345-3623475
Experience: -
I have One year Teaching Experience in an Global Computer Institute.
C/C++,Visual Basic,Ms-Office(Ms-Word,Access,Excel)
SQL Server 2000,Crystal Report.
Qualification:-
Academic:-
Graduation:
MS (Network &Telecommunications;)
from Mohammad Ali Jinnah University
BS (Computer Science)
from Orasoft Training Institute
Affiliated With University Of Shah Abdul,Sindh
Intermediate: H.S.C from Govt Dehli Science College
Matric : S.S.C From F.J Grammar School
Strength:-
Handling and Design the Desktop Applications Through Visual Basic With Database Connectivity Like Sql Server-2000,Ms-Access and Oracle.
Handling and Design the Distributed Applications Through ASP and Jsp with the database connectivity like Ms-Access,Sql-Server-2000 and Oracle 8i.
Projects:-
• Library Information System
• Inventory Control System
• Sale/Purchase System (Designed in Visual Basic)
• Fixed Assets Management System(Under Process)
Reference:-
I have listened through my colleague.So consider me as an employee candidate in your concerned department of your organization.
Making a vaccum hose adapter
|
|
Adjustable Loc line 1/2 inch with connector $5.50 |
|
|
Loc-Line 1/2 inch Tee Fitting $5.25 |
|
|
Loc-Line 1/2 inch Ball Socket Inline Valve $9.75 |
|
|
Loc-Line 1/2 inch Ball-Socket End Cap $1.60 |
|
|
Loc-Line 1/2 inch Ball Socket 90 Deg Nozzle $2.25 |
|
|
Loc-Line 1/2 inch Ball Socket Circle Flow Assy $35.00 |
|
|
Loc-Line 1/2 inch Flat Slot 125 Nozzle $2.50 |
|
|
Loc-Line Modular Pipe Y Fitting – 3/4″ $4.99 |
|
|
Eductor Accelerator 3/4 Inch Loc line MaxFlow 2x to 4x $15.50 |
|
|
Loc-Line 2 1/2 inch Swivel Nozzle 75 for 1/2 inch Loc- $7.25 |
|
|
Loc-Line 1/2 Inch Flat 5 Hole Nozzle $2.50 |
|
|
Loc-Line 1/2 inch Ball Socket FPT Valve $9.75 |
|
|
Loc-Line 1/2 inch Ball Socket Y $4.25 |
|
|
Loc-Line 1/2 inch Ball-Socket End Cap $1.60 |
|
|
Loc-Line 1/2 inch Ball Socket 90 degree Elbow $4.50 |
|
|
Loc-Line 1/2 inch Ball Socket 90 Deg Nozzle $2.25 |
|
|
Loc-Line 3/4 inch x 3 inch Ball Socket Flare Nozzle $5.99 |
|
|
Loc-Line 1/2 inch X 1/4in Reducing Nozzle $1.75 |
|
|
Loc-Line 3/4 inch Ball Socket 90 degree Elbow $5.29 |
|
|
Loc-Line 3/4 inch Ball Socket MPT Valve $10.79 |
|
|
3M Filtrete 3US-PS01 Under-Sink Advanced Water Filtration System $37.80 Filtrete undersink filtration kit is easy to install; it can be done in less than 30 minutes. No dedicated faucet is needed. All fittings and detailed instructions are included…. |
|
|
Revere Copper Clad Bottom 1-Quart Saucepan $8.99 1033394 Revere high quality stainless steel teakettles feature a copper bottom for quick and even heat distribution. Polished stainless steel exterior is low maintenance while the black phenolic handles are cool to the touch. Features: -Saucepan. -Stainless steel construction. -Copper bottom for quick and even heat distribution. -Capacity: 1 Quart. -Overall dimensions: 11.25” H x 2.75” W x 5.75… |
|
|
Sprite Chrome Universal Hose Filter $22.49 Sprite Chrome Universal Hose Filter. The Sprite hose filter, attaches to any existing shower handle. Swivel-ball attachment allows the filter to adjust to any angle. Replaceable cartridge is rated for three months. Softens skin & hair by removing chlorine & pollutants from water. Reduces water scale build-up on shower walls & glass…. |
|
|
2Cellos $6.82 2CELLOS (Luka Sulic & Stjepan Hauser) became a worldwide YouTube sensation with their passionate, dueling cellos version of Michael Jackson’s Smooth Criminal. Their debut album is a thrilling collection of chart-topping rock and pop songs performed in their signature style from Guns N’ Roses, U2, Coldplay and Kings of Leon, among others. Their follow-up single is an explosive version of the Guns N… |
|
|
Kind of Blue $4.44 This is the one jazz record owned by people who don’t listen to jazz, and with good reason. The band itself is extraordinary (proof of Miles Davis’s masterful casting skills, if not of God’s existence), listing John Coltrane and Julian “Cannonball” Adderley on saxophones, Bill Evans (or, on “Freddie Freeloader,” Wynton Kelly) on piano, and the crack rhythm unit of Paul Chambers on bass and Jimmy C… |
