Newsletters

Sign up to receive the free ISBSG eNews direct to your Inbox by clicking here.

eNews Webinar: IT Silver Bullets.  Do they hit the mark? December 2009

Presenter: Peter R Hill (CEO, ISBSG)

Whether your Maintenance & Support (M&S) services are in-house, outsourced or you are an M&S outsource provider, having access to an international, independent source of data on M&S applications will
help you better manage the resources spent on M&S services.

The ISBSG inivites you to submit application data to its M&S repository so that it can offer those involved in M&S, and its management, access to industry data and analysis that will help answer these questions:
• How many support people will you need?
• How does your time-to-repair compare to other similar support environments?
• What proportion of time is spent on maintenance and what on support?
• What percentage of time is being spent on the different categories of maintenance: Corrective & Preventative; Perfective & Adaptive?
• What level of incident reporting is being experienced?
• How do you compare to other organisations in the same industry?

Complete anonymity and security of submissions is guaranteed.

If you submit 5 or more M&S applications before 31 January 2010 you will receive the following:

1.      A benchmark report for each of your applications.
2.      A special M&S analysis report (by April 2010)
3.      A 20% discount off the next release of M&S data.

To enter your M&S application data online or using an MS Word Questionnaire, click here

eNews High Speed, Low Cost Functional Sizing October 2009

Capers Jones’ paper entitled "A New Business Model for Function Point Metrics" looks at the latest developments in the cost effective sizing of software early in its life cycle. Jones contends that current
functional sizing is too expensive to be used by more than ten percent of the industry and that to achieve optimum returns, it is necessary to modify function point practices in a number of important ways:

1.  The starting point for function point analysis needs to be earlier by 6 months
2.  The cost of function point analysis needs to drop below 1¢ per function point
3.  The counting speed needs to be higher than 10,000 function points per day
4.  There should be no lower limit on application size, down to zero function points
5.  COTS packages should be able to be counted
6.  Cancelled projects, if they occur, should be counted
7.  Changing requirements should be counted in real time as they occur
8.  Deleted features should be counted in real time as they occur
9.  Legacy applications should be counted, if they were not previously counted
10. Updates to legacy applications should be counted before they are started
11. Function point sizing should be possible as a service without on-site visits
12. Function point sizing should include requirements changes
13. Function point sizing should include total requirements growth for every release

Jones looks at the current costs and timing of functional sizing and the technology that is starting to become available to the industry to achieve his proposed changes.

Capers Jones has made his paper available free from the ISBSG web site.

Click here to download your free copy.

eNews Real Time vs Business and Component Projects September 2009

NEW SPECIAL REPORT: The Performance of Real-Time, Business Application and Component Software Projects

The performance of, and differences between, Real-Time, Business Application and Software Component projects are examined in this 45 page report.  The different types of projects within these three types
of software are analysed: New Developments, Enhancements and Re-developments.  Important results are reported for:

Effort vs Size
Duration vs Size
Economy of scale
3GL vs 4GL Productivity
Effort distribution – what % of effort should be spent on various project activities?

If you are involved in project estimation or benchmarking, this report will provide you with valuable information to help your accuracy.  It underlines the need to understand the differences involved in developing or enhancing software for different application types.  It emphasises how important it is not to mix their performance results.

Click here to order online.

This Special Report is provided free to ISBSG Special Reports & Web Subscribers.  Click here to subscribe.

eNews ISBSG Web Subscription Special Offer September 2009

We are pleased to offer you a free copy of your choice of either The Benchmark 8 - Software Defect & Quality edition or The Benchmark 10 - Project Planning edition (valued at €90 EURO), when you become an ISBSG Web Subscriber.

Benefits of an ISBSG Special Reports & Web Subscription include:

    * Free copies of all ISBSG Special Analysis Reports (usually €25-€60 EURO each)
    * Exclusive use of the on-line estimation Reality Check tool so you can quickly check the accuracy of your project estimates against industry performance.
    * A free, single user license for the Early Estimate Checker tool. This tool quickly generates estimates of the work effort and elapsed time (duration), required to carry out and complete a software development project, plus project delivery rate and speed of delivery.
 Instant estimates are generated using software size and other available project characteristics.
    * 10% discount off all ISBSG products and services

An annual subscription costs €155 (approximately US$210).

SPECIAL OFFER:   Subscribe before 31 October 2009 and receive a free copy of your choice of Benchmark book valued at €90.

Click here to read more about the The Benchmark 8 - Software Defect & Quality edition and The Benchmark 10 - Project Planning edition.

To take advantage of this special offer click here to subscribe

eNews COSMIC industry data now available September 2009

Release 1 of the ISBSG COSMIC industry data contains data on more than 350 projects that have been sized using the COSMIC functional sizing method. It includes performance data on 290 projects from the domain of business application software and 62 projects from the domains of real-time applications and of software components for both new developments and enhancements. This data is ideal for benchmarking, for performance comparisons and for project estimating.

The Performance of Real-Time, Infrastructure & Business Applications Software Projects.

In the first half of this year the Common Software Measurement International Consortium (COSMIC) and the ISBSG ran an initiative to attract more history data for software projects that had been sized using the COSMIC method.  On the successful completion of this project Charles Symons, Joint Project Leader COSMIC, performed an analysis of all the COSMIC sized projects in the ISBSG repository and produced a detailed report.

This twenty-seven page report provides an insight into the development and enhancement performance of three different types of applications: Real-Time, Infrastructure and Business.  The analysis is based on a data set of 334 projects that have been sized using the COSMIC method.

The analysis results underline the importance of treating these three types of applications differently when estimating and benchmarking.  Project Delivery rates and duration times are provided, along with
useful charts of the correlations of Effort, Productivity and Duration to software size.  The productivity rates of different Programming Languages are compared, as is the impact of software size on productivity.

This report provides valuable information for benchmarking and estimating COSMIC sized software.  It also provides a revealing analysis of the performance of Real-Time; Infrastructure and Business applications software projects.

Click here to order this Special Report online. 

The report is provided free to Special Reports & Web Subscribers.  

Submitters of 5 or more COSMIC projects recieve a free copy of the report.  To submit data click here.

eNews More on project estimation July 2009

Estimating a software project is not something that you do once and then forget.  There is a need to estimate a number of times during the life of a project. As you gain more knowledge about your project the better your estimate will be.

In the last eNews we looked at the use of a very simple tool (the ISBSG Early Estimate Checker) for “ball park” estimation of effort and duration when a software project is simply an idea or concept.  But as you learn more about the likely attributes of your project you can refine your estimate. The ISBSG Comparative Estimating Tool is ideal for that.

Release 6 of the tool is now available and it allows you to enter a number of attributes for your planned project.  In addition to the predicted size, platform and development type, you can enter values for some or all of the following attributes:

The tool then selects similar projects from industry data within the ISBSG Repository and provides a range of estimates: “Optimistic”, “Likely” and “Conservative” for project effort, delivery rate, duration and speed of delivery.

The ISBSG Comparative Estimating Tool is an MS Excel based tool. The values that it produces are determined from the averages of project delivery rates and speed of delivery from a selection of similar projects in the ISBSG Repository of over 5,000 projects.

This tool filters the ISBSG data to ensure a comparison of “apples with apples”.  It then allows the user to specify project attributes to further refine the choice of projects for the estimates.  Download the free User Guide which includes:

• a description of the estimating process implemented by the tool, and
• details on how to use the tool most effectively.

eNews Early Project Estimation July 2009

It is a challenge to establish an estimate for a proposed software project very early in its lifeoften before it has been confirmed as a funded project.  In a previous eNews we looked at how to estimate the rough size of a piece of software. It is likely that the reason for wanting to know the size is so that you can estimate the likely effort and duration and hence the likely cost.

In the same way that we are able to establish a rough size by using the known attributes of a piece of software, we are also able to establish a rough estimate of effort from the size and other known attributes.

If you have your own project history data you can use productivity figures derived from that history as the base for a new estimate.  Or if you have one of the sophisticated estimating tools, you will use that.

Another option is to use the Early Estimate Checker tool that is included with the industry data on the ISBSG Estimating, Benchmarking & Research Suite R11. The Early Estimate Checker uses the ISBSG data with regression analysis to generate estimates of the work effort and elapsed time, (duration), required to carry out a software development, plus delivery rate and speed of delivery. To obtain a rough estimate you can simply enter software size, or to generate more specific estimates, enter size plus a combination of development platform, primary programming language type and development type.

The Early Estimate Checker tool can also be used to validate existing project estimates (eg. bottom-up estimates generated from a project's work breakdown) or assess the reasonableness and likely risk
associated with a quoted estimate (where on the range from most conservative to most optimistic is the estimate positioned?).  The Early Estimate Checker has been designed to complement existing
software estimating tools and techniques.

Click here for more information on The Early Estimate Checker tool which is available free when you license the ISBSG Estimating, Benchmarking & Research Suite or when you become an ISBSG Special Reports & Web Subscriber.

The ISBSG Estimating Benchmarking & Research Suite R11 can be licensed.  Upgrade prices are available for R10 and R9 licensees.  Email admin@isbsg.org if you are interested in upgrading.

eNews There are many options for software sizing June 2009

One of the major problems in trying to get better estimates and benchmarks for software projects is the need to establish how big the software is.  People who will happily estimate the size of a room in their house by pacing along the walls, seem unaware that they can use similar size approximations for their software. Of course, it is much better to measure a room accurately with a tape measure, but if you
haven’t got a tape measure, or don’t need such accuracy, then an estimate of the size is acceptable.

The size of software is best expressed as a functional size. In February we looked at a very rough way to estimate functional size from Use Cases.  If you don’t have Use Cases there are other rules of thumb for size estimation that use other known attributes.  For example:

FPs = 35 x Internal Logical Files + 15 x External Interface Files (Steve McConnell, Software Estimation, pg 203)
Identify and count the entities/classes in a domain model as the ILFs.
External Interface Files are as identified in the interface analysis.

There is a more complex method that analyses the GUI components to calculate approximate FPs. This is documented in Steve McConnell, Software Estimation, pg 204.

Another alternative is the 3-factor method documented in Capers Jones, Estimating Software Costs 2nd Edition, pg 103 to get an order of magnitude FP estimate for a project. This method tends to provide
conservative figures.

In his presentation to ACOSM in 2008 Michael Stringer listed the following as good predictors of project size/cost that can be counted easily: Use cases; data entities; major features; users; installations; computers; work locations; number of existing records; tables, screens and reports in an existing system.

Click here for Rules of Thumb for early lifecycle software sizing.

At least two new tools that will estimate functional size from these types of predictors are close to release. These will be a major breakthrough in the use of functional sizing.

The level of accuracy that you require for software size will depend on what you want to use the size for.  The level of accuracy that you can achieve depends upon the amount of information that you have about
the planned software.

Different purposes will require different degrees of accuracy.  If, for example, you need a rough size simply to establish whether the project you have in mind is close to your budget, then a size approximation, based on the known project attributes, could be satisfactory.  If, on the other hand, you need to produce an accurate quotation to build some software, then you should have the size established via a full functional size count.  That will require a detailed Functional Requirements document.  For benchmarking purposes an accurate size is essential.

Six levels of sizing accuracy and a link to a count sizing paper can be found here

There are many options for software sizing.  You simply have to choose the most appropriate for your needs and the information that you have available at the time.

The ISBSG has established a repository of Maintenance and Support (M&S) Application data that now has 320 M&S Applications. Our revamped web site has a dedicated Maintenance & Support section where you can find a free example of an Application Benchmark Report, a free copy of the ISBSG analysis report on the M&S data, and guidelines for the M&S metrics that might be useful for your operation.

You can now license a copy of the Maintenance & Support Industry Data Release 2, for €200. Visit www.isbsg.org/products

eNews The Benchmark Release 10 Book now available April 2008

The Benchmark Book Release 10 is now available in pdf format on a CD. This book provides analyses of Release 10 of the ISBSG development & enhancement data. It investigates the factors that impact the planning, management and benchmarking of software development and enhancement projects.

The information in this book will help you find answers to the following questions:

• What percentage of project effort should I expect for each phase of my project?
• What percentage of effort will be spent by each of my development resources?
• Are there techniques and tools that can help to make my project successful?
• How will the web project I have to plan/manage differ from a non-web project?
• Will it help if I employ a re-use strategy?

Visit www.isbsg.org/products for more information.  The Benchmark publications are a series of releases each of which covers specific topics based on the analysis of the project data in the ISBSG Development & Enhancement Repository. As the Repository grows we are able to analyse new topics.