Measurement is a wonderful thing.  However measurement without standards and definition can be worse than no measurement at all.  This paper which I will be presenting at the 2010 ISMA conference begins the attack, highlighting the need and proposing that additional work commences in standards for estimation and measurement.  Software Estimation and Measurement 2010

Related posts:

1. Practical Software Measurement Conference June 22- Orlando
2. Weight Management = Measurement and Estimation
3. Function Point Simple Introduction

Related Posts Computer Generated

• Optimism: The Enemy Of Successful Acquisitions
• Function Point Simple Introduction
• Forecasting Guidebook from the UK Ministry of Defense
• Creative Cost Efficiency Article From 2008 European SEER Users Conference
• Cost Estimating Throughout the World

Dr. Ricardo Valerdi of MIT sent me a pre-publication copy of his upcoming IEEE article. Here is the abstract:

“Engineering cannot wait until all phenomena are explained. Engineers may work effectively, often for centuries, with heuristics. This paper provides thirty one heuristics that have been inspired by the development and application of a systems engineering cost estimation model. The objective of this paper is to present such heuristics in a simple manner so that they can benefit systems engineering researchers and practitioners that develop, calibrate, and use cost models.”

I enjoyed the article (as I do with pretty much everything Ricardo produces).  Such simple truths. A few heuristics quoted from the paper follow:

“Don’t assume the original statement of the problem is necessarily the best, or even the right one.”

“Let the available data drive the application boundaries of the model.”

“Design the rating scale according to the phenomenon being modeled.”

The full article is available, as provided by Ricardo here: “Heuristics for Systems Engineering Cost Estimation.”

Related posts:

1. Estimating Systems Engineering…. The Upcoming Book.. and SEER
2. Make or Break: Why Accurate Cost Estimation Is Key
3. CECIM: A Capability Improvement Model For Cost Engineering

Related Posts Computer Generated

• Estimating Systems Engineering…. The Upcoming Book.. and SEER
• Considerations for CMMI Level 4 for Systems/Software product development and deployment

This list was adapted from ITBusinessEdge

Lack of governance: Project criteria, roles, processes & outcomes not used or accepted by management.  Not understanding project risk.

Internal politics: Territorial fights.  Its not my job, or “they” messed up.

Communication issues between the business and IT: IT talking with the business stakeholders about bandwidth and blobs rather than end user oriented benefits.

Unclear expectations: Bad estimates and ambiguous expectations.

Lack of fact based analysis: Plans not based on facts but on opinions.  Studies have shown, for example, that projects of any magnitude can’t produce a viable estimate without a model like SEER.

Lack of input from users: IT may know how to do it but users probably know what they need better.

Changes in project without re-planning

Unplanned changes in key personnel

Unrealistic schedules: Projects on death marches.

Unanticipated operations costs: These must be estimated well up-front

Related posts:

1. More Software Project Failure / Challenge Information From CAI
2. Poor Estimation: Major Root Cause of Project Failure
3. Software Project Failure Costs Billions.. Better Estimation & Planning Can Help

Related Posts Computer Generated

• Step One: Establish Estimate Scope and Purpose
• SEER TO Remediate Troubled Corporations
• Eight Causes of Project Failures: UK Government IT Projects
• COSMIC Function Points In SEER

Estimating the slowdown on the freeway.  Interesting article from Wired quantifies and estimates the occurrence of phantom traffic jams.    You all know them… traffic slows to a crawl.  There must be an accident.  But no, it is just a phantom traffic jam.  Living in Los Angeles I find this really interesting.

Related posts:

1. Estimating Web Effort / Schedule
2. Cost Estimating Throughout the World
3. Estimating Your Fitness Level

The University of Southern California has been developing and updating line of code counters for a number of years.  Such code counters can be very handy when using lines of code as a size measure.  I know, many object to using lines of code, but when used correctly they can work well.  We see users who have just as much success with lines of code as they do with function points, use cases, etc.

Even if you are a functional size user, knowing the SLOC for legacy can be useful, rather than counting the function points, etc.  Here is the announcement from USC:

We are pleased to announce that a new version of the Unified CodeCount (UCC) tool is now available to the public at http://sunset.usc.edu/research/CODECOUNT/. This Release 2010.07 supports new programming languages (e.g., Fortran, Python, ColdFusion, Bash and C-Shell script) and the CSV output format among other enhancements and bug fixes. Please refer to the release notes document for further details at http://sunset.usc.edu/research/CODECOUNT/download/2010/UCC_Release_Notes_v.2010.07.pdf

Related posts:

1. New Code Counting Tool Made Available
2. Software Code Counter Review and Recommendations
3. Dealing With Generated Lines of Code

Related Posts Computer Generated

• Have We Lost Our Ability to Estimate Software Size
• Step Ten: Track Project throughout Development
• Step Four: Software Sizing
• Software Code Counter Review and Recommendations
• New Code Counting Tool Made Available

Aberdeen group published some interesting information regarding the time and cost to recover from business interruptions.  This is the time to recover 90% of functionality.  I recommend getting the complete report.  They found that best in class recovered 6.5 times faster than laggards and had an average cost of $72,000 versus laggards with an average cost of $2,880,000.

Related posts:

1. Cloud Computing Can Save Time and Money: MIT CIO Panel Conclusion
2. The Cost of Cloud in the Sky Computing Part 1
3. The Cost Vs Benefit of Additional Bandwidth

Related Posts Computer Generated

• The Cost Vs Benefit of Additional Bandwidth
• The Cost of Cloud in the Sky Computing Part 1
• SEER for IT: The Making of a New Product and Lessons learned
• Live From UK Williams Formula 1: Ford Motor Company Europe Uses SEER for Software and IT
• IT Cost of Rogue Users

These days software estimation vendors are competing to have the largest repositories of completed software projects, and the customer is encouraging this competition, which is fundamentally good.  However, there is more to insuring the accuracy of an estimation model than just having a lot of data points sitting on the proverbial shelf.

Where Data Comes From

The first question asked of a vendor is, where does your data on completed software projects come from?  Early on, much of it came from Government agencies, who in turn collected from contractors.  Over time, public sources have emerged that contain voluntarily submitted information from private companies worldwide; the prime example of this being the International Software Benchmark Standards Group (ISBSG).  Galorath has obtained software project data over the years through numerous private and public sources.  The data comprises many thousands of total observations that have passed data quality tests.  Most observations contain size and effort information, thousands more do not contain all the desired fields.

What is Done With Data

While plain-vanilla data can reveal a lot, it has its limitations.  For this reason, Galorath maintains extensive surveillance of industry trends, including third-party analyses.  These can reveal insight into changes in modern practices such as Agile development, the productivity gained by the latest IDEs, and many other ongoing evolutions.  The company is a member of numerous industry consortiums – in part to obtain access to the latest research available.

At Galorath, once data is acquired, it is processed into a form that is usable for analysis.  This involves normalization so that the data points are comparable, i.e., include the same activities from early requirements through testing and the same types of labor, including programmers, testers, management, etc. We also try to find and understand “outliers” – those projects that are so different that they are not useful.  At numerous conferences and in webinars, we have described our normalization process and compared our results against other methods.

Using the collected data we update SEER for Software in several ways.  A key method is to run our model against various stratifications (specific subsets such as “Business” and “Client-Server”) that are defined by SEER for Software’s knowledge bases.  Simply put, we compare the model’s estimates to observed outcomes.  Based on these results, knowledge bases are re-calibrated when necessary.  In fact, data sets are not uniform in terms of the information observed.  Some completed project records may include peak staff, development activities, and software language used, while others don’t.  We account for this by performing a separate analysis of various factors: language productivity, development proportions, productivity variation by application or development method, and so on.  These analyses are done first, and the model is adjusted, before gross analysis is begun.

SEER for Software’s core model is configured to a particular circumstance by a set of knowledge bases, and it’s these knowledge bases that are calibrated based on new industry information and trends.  Each knowledge base is defined in terms of a set of parameters, some visible to users, and others normally hidden.  When a knowledge base is updated the visible parameters, such as Modern Development Practices, may be modified and some underlying calibration factors may be adjusted.  These knowledge base adjustments occur every few years as evidence warrants.

ProjectMiner, Analogy Based Estimation, Metrics and Benchmarking Can Use Raw as Well as Processed Data

While SEER for Software includes numerous data driven functions (such as ProjectMiner data driven estimates, Metrics and Benchmarking data driven analysis, and more)  the overall evolution of SEER for Software is best called “innovation-driven.”  While data analysis is a very important part of how we maintain SEER for Software, we also continually enhance the model’s ability to estimate real world projects.  These enhancements have often been industry firsts: flexible project staffing, off-the-shelf (COTS) integration modeling, translation of estimates into detailed project plans having intricate interdependencies, extended schedule and small schedule estimating, cloud computing solutions, to name only a few.  All these innovations, alongside data-driven updates, serve an important role in insuring the model’s precision.  SEER uses data in a number of ways including, but not limited to:

• Refine knowledge bases and Equations
• Provide Trend Lines on Metrics and Benchmarking
• Plot Your Estimate compared to actual completed projects
• Apply to ProjectMiner (AKA SEER Independent Crosscheck & Verification) which provides a completely data driven estimate either as a crosscheck or primary estimate
• Provide sizing for systems

Related posts:

1. Over 4000 Raw Datapoints Available To SEER for Software Via ISBSG
2. Step Three: Collect Data
3. Using the SEER Effective Technology for Productivity Assessment and Benchmarking

Related Posts Computer Generated

• Viable Software Estimation Modeling: A Key Component Of Software Risk Management
• Using the SEER Effective Technology for Productivity Assessment and Benchmarking
• Software Development Metrics At Galorath: Co-located Teams Are Still Most Efficient
• SEER’s 20th Anniversary and Request for Feedback
• SEER for Software Quick Start: Sizing an Entire Corporate Software Portfolio in Less Than One Week

http://www.galorath.com/index.php/library/estimating-united-conference/

Related posts:

1. Dan’s 2 Papers At CMMI Conference In Denver Nov 18 & 19
2. Estimating United Conference Papers Available On the Web
3. SEER User Conferences & Innovation Awards

Shows ROI of inspections, PSP, tsp, and more  http://davidfrico.com/roi-book.htm

Related posts:

1. Interesting, Classic Look at Software Maintenance
2. ROI: Return on Investment or Risk On Investment
3. Why Projects Fail From A Technical and Process Viewpoint

Related Posts Computer Generated

• ROI: Return on Investment or Risk On Investment

 Gender equity in computing has long been a national goal advanced by those concerned with fairness and by those who know that the female point of view improves the design and development of software systems. Unfortunately, the percentage of young women entering computing-related majors keeps falling, and the female dropout rate is higher than the very high male dropout rate.  

The Bureau of Labor Statistics predicts a large increase in the need for B.S. and M.S. computing graduates in the next decade. The largest untapped pool of potential computing majors and, eventually, computing professionals, is science- and math-talented high school students, but only about 10% of entering undergraduate majors in computing majors are female. Despite the many initiatives aimed at attracting young women, the number of female computing majors keeps dropping…

Prof David Klappholz is involved in the Real Projects for Real Clients Courses (RPRCC) initiative, a K-12 and college level ACM-W project aimed at recruiting young women into, and retaining them in, computing-related majors. The initiative’s approach is based upon a 35-year-long psychological study that followed hundreds of mathematically- and scientifically-talented youth from middle school to middle age and elucidates gender differences in career choice.

Galorath’s head of development is female, as is nearly half our development staff.  And so naturally from us, three cheers for RPRCC!

From http://users.drew.edu/ftrees/TECS/Session_Descriptions.htm

Related posts:

1. Cloud Computing Can Save Time and Money: MIT CIO Panel Conclusion
2. The Cost of Cloud in the Sky Computing Part 1