Software Engineering Body of Knowledge from IEEE
The 2004 Software Engineering Body of Knowledge is available at no cost from: http://www.computer.org/portal/web/swebok/htmlformat While it doesn’t focus on estimation, planning and control, it is a valuable resource for software engineers. There will be an expanded 2010 version coming out as well.
Thank you for reading “Dan on Estimating”, if you would like more information about Galorath’s estimation models, please visit our contact page or call us at +1 310 414-3222.
Eating the IT Elephant: Moving From Greenfield to Brownfield Development
The book “Eating the IT Elephant“ discusses the issues of moving from developing new systems to continuing evolution of legacy systems and provides practical, step by step methods of getting these under control. It discusses Brooks Law and its fundamental truths and then says “Brooks had it easy.” It points out that Brooks’ original work was developing a new (Greenfield) system with a local team. And that today our teams are global, our systems are scattered and interrelated.
It discusses the demands of global systems, project reporting, change management, induced complexity (the technology itself is rarely the problem but the way the technology is put together becomes the “induced complexity”), requirements definition and, my favorite, ORGANIZATION and PLANNING.
While the book does not reference SEER directly, it discusses the difficulties and probable failure of projects that do not understand the minimum time, the realistic costs and the risks.
This book recommends EVERYONE read Frederick Brooks’ book, “The Mythical Man Month.” I heartily agree. It has been interesting to see others attack Brooks Law , “There is an incremental person when added to a project that makes it take longer, not less time, and adding people to a late project makes it later.” Brooks Law is alive and well today and involved in nearly every Brownfield “legacy innovation system” as well as new system. The key is making sure you are viewing Brooks Law at the correct level.
The book includes a process for “eating the elephant,” improving Brownfield developments. I recommend giving it a read.
Thank you for reading “Dan on Estimating”, if you would like more information about Galorath’s estimation models, please visit our contact page or call us at +1 310 414-3222.
Don’t Hide Uncertainty: Measure It
We all need to deal with estimates / projections for the future. When doing so, it is essential to include analysis of uncertainty and sensitivity of the variables including:
- Most important assumptions in the result
- Sensitivity of those assumptions (what happens if the assumptions change; which assumptions are most important in controlling overall results?)
- Probability of the estimate
- Risk Analysis showing the impacts of potential risks
Without risk, I don’t believe one has a viable estimate.
Thank you for reading “Dan on Estimating”, if you would like more information about Galorath’s estimation models, please visit our contact page or call us at +1 310 414-3222.
Save 5 to 9% of IT Spending By Zero Based Budgeting
According to a survey by CFO magazine , IT organizations can save 5 to 9% of their budgets by starting their budget from scratch (zero-based budgeting) rather than just revising last year’s budget
The CIO Executive Board created a model of best budgeting practices, based on a survey of 200 of its member IT executives about their budget planning for this year. It then compared the model with practices employed by companies at the other end of the efficiency spectrum to determine potential savings.
The group estimated that some companies wasted 5% to 9% of their 2009 IT spending in missed cost-cutting opportunities, largely by revising the previous year’s budget rather than starting from scratch with a zero-based budget. Freshly scrutinizing every line item tends to unearth significant savings, says Andrew Horne, senior research director for the CIO Executive Board.
The article also identified significant cost savings by ensuring that IT is aligned with the business objectives.
Thank you for reading “Dan on Estimating”, if you would like more information about Galorath’s estimation models, please visit our contact page or call us at +1 310 414-3222.
Frank Freiman: Early Parametric Modeler Passes Away
I heard today that Frank Freiman, one of the very early parametric modelers passed away. Our prayers go out to his family.
Frank started modeling with RCA in the late 1960′s and started RCA’s PRICE systems division. He left RCA and started his next generation estimating models (FAST models) sometime in the 1980′s.
While FAST disappeared some years ago (Frank offered it to Galorath at one time but we didn’t see a fit with SEER) I did find an old FAST document on the web
I was also awarded the Freiman lifetime achievement award, named after Frank.
The world has lost a great man.
Thank you for reading “Dan on Estimating”, if you would like more information about Galorath’s estimation models, please visit our contact page or call us at +1 310 414-3222.
Modeling a New Year’s Resolution In Relation to SEER
Another masterpiece by Galorath’s David DeWitt follows. Amazing it is nearly New Years again. And that it has been 4 years since my personal weight loss.
Modeling a New Year’s Resolution
David DeWitt, Senior Consultant, Galorath Incorporated
Part of the problem with becoming a cost estimation aficionado, who thrives on parametric modeling, is that one tends to think they can build a model for just about everything: Software, Hardware, Aircraft wings, IT infrastructures, cattle futures, etc. In fact, the more one builds the estimates and discovers that the answers are accurate – the more insatiable the desire becomes. Certainly, it must be possible to build a parametric model for nearly anything!
With that in mind – and considering the time of year – what a perfect opportunity to finally perfect that nastiest of all New Years Eve resolutions, (and the one that rings the bells of cash registers in health clubs, sporting goods stores, and vitamin shops around the world):
“This year I vow to… LOSE WEIGHT!”
But, how can one use a parametric modeling approach to losing weight, you ask? Ah, let that beguiling high begin as sacred knowledge bases are considered and the plethora of parameters prescribed. This should be quite easy, after all, look how much a software estimate has in common with losing weight:
- “size” is certainly the largest driving factor in software and ones girth,
- an increase in functionality (due to the lost weight) is the desired outcome for both,
- a methodology must be selected (diet, exercise, grapefruit enemas) to provide an approach,
- and the rigor of oversight (a personal trainer) is key to success or failure
It looks to me as if the math necessary to build a software model contains all the same components for losing weight. Let’s give it a try! First I will need to select the appropriate knowledge bases that best reflect what I am trying to model.
| The Human Problem | The Software Equivalent |
| I want to lose weight so I can perform better in my personal and professional life | Platform: Business Mission Critical – General business and management information system host platform environments that perform functions critical to the execution of the enterprise’s primary mission |
| I’d like my body to be slimmer and more attractive | Application: Graphical User Interface – the impacts of developing user interfaces for business and other applications. |
| I’m willing to increase my exercise and significantly modify my diet. | Acquisition Type: Major Modification – Design will change up to 30% and regression testing will need to be done on all code. |
| I prefer to measure my performance in short cycles to ensure I’m making progress | Development Method: Agile Full – The primary criterion for determining the content of each repetition is customer need (and is typically customer driven). The focus of this strategy is for the software to evolve as the customer requirements are interpreted and implemented over time. |
| I will need a nutritionist and personal trainer to keep me motivated and encourage my success | Development Standard: Commercial High – assumes relatively formal processes and reviews to ensure that sufficient documentation is available to maintain and update the software. |
Now then, all that’s left is to decide is how much weight I want to lose! The best approach would be to look at this as a problem of “calories may cometh but more calories must go-eth!” Since, 3500 is the typical number of calories burned per pound of fat, then for every pound I want to lose I multiply that number by 3500. For example, if I want to lose 35 pounds (hypothetically speaking… honestly), then 3500 x 35 pounds = 122,500. This now represents my calorie weight loss goal.
Of course I need to know my current size expressed as calories (rounding down) 200 x 3500 = 700,000
That means my ideal weight in calories would be (700,000 – 122,500) = 577,500 (or 165 pounds). Now there’s a goal!
Using estimation speak – I can say I have 700,000 preexisting calories, plan to delete 122,500 calories so that my effective size will be 577,500 calories. Easy, right?
But, before I go much farther I need to decide what would be a safe and reasonable number of pounds to lose per month. After all, I’m not interested in a starvation diet – (Emmm, Beer!), but I don’t want this to take forever. Why not a goal of one year! That way this time next year I can pick a new resolution. So, my goal is (122,500 / 12 months) or 10,208 calories per month.
Finally, using the SEER for Software cost estimation software from Galorath Incorporated, all I have to do is:
- Create a proxy that reflects my calories per month in effort I’ll need to expend,
- Select my knowledge bases using the table above,
- Plug in my ideal weight into the “design for reuse amount”,
- Enter the amount of weight to be lost as my “preexisting not designed for reuse”
Drum roll please… and voila! The answer is I can lose the weight in Ten Months using a peak staff of 4.4 people (Me, trainer, dietician, and fan club I suppose).
Wow –only ten months for me to lose all this hypothetical weight! And that makes sense, because as the plan progresses I will be building muscle that burns the calories even faster. Using the SEER for Software cost estimation modeling application I’ve proven that you really can build a parametric model for anything! Now all I need is a ten month sabbatical, funding for the full time dietitian and trainer, and a fan club. Any volunteers?
Thank you for reading “Dan on Estimating”, if you would like more information about Galorath’s estimation models, please visit our contact page or call us at +1 310 414-3222.
Project Prediction Infographic
