Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 1 1/11/2004 Day 2, Part 1 Estimating Software Size Section 2 Calculating the Benefits of Software Reuse

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 2 1/11/2004 The Overall Planning Cycle Analyze Job Manage Risks Execute Generate Detailed Plans Generate Initial Plans Measure, Manage Productivity and Quality

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 3 1/11/2004 Detailed Planning - Processes Estimate Effort and Cost Estimate Schedule Evaluate Source Information Statement of Work Requirements Constraints Standards Processes History etc. WBS Size Effort & Cost ScheduleOK Complete Detailed Planning Revise & Negotiate Not OK Estimate Size

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 4 1/11/2004 Architecture of Spreadsheet Cocomo Based Estimate Other Effort Estimates... Analogy based Size Estimate Software Reuse Analysis Final Effort Estimate Productivity Based Effort Estimate Other Size Estimates... Final Size Estimate Expert Based Size Estimate Size / Reuse EffortSchedules Generic Schedule Effort Schedule

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 5 1/11/2004 Calculating the Benefits of Reuse

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 6 1/11/2004 Dealing with Reuse Many software programs are derived from previous programs This MAY result in savings of cost and/or time It MAY also result in increased quality BUT !...Reuse can also cost more and take longer and yield lower quality

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 7 1/11/2004 Quality = x Reused: Quality = x Developed: Quality = y Old SoftwareNew Software Quality of new software depends on x and y. If old software has poor quality it brings down the quality of the new software. But if it has high quality, it brings quality up. Reuse Quality Diagram

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 8 1/11/2004 What Can be Reused? Just about Anything Code Test code Test cases and procedures Documentation Design specifications Designs Requirements specifications Etc., etc.

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 9 1/11/2004 Reuse is Almost Never “Free” Because you seldom have everything you need –You typically need to create tests or documents or other things And you need to design the software to incorporate the “reused” components And you need to integrate the “reused” components with everything else

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 10 1/11/2004 Reuse Terminology Legacy Software Software developed for a previous application & BELIEVED TO BE OF USE for new application Modified Software Software developed for a previous application that is suitable for a new application after a MODEST AMOUNT OF MODIFICATION Reused Software Software developed for previous application suitable WITHOUT CHANGE OF ANY KIND

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 11 1/11/2004 Legacy Software vs. Reusable Software Legacy Software Designed for a Specific Application Lax Standards for –Documentation –Test Procedures –Test Cases Like looking in a junkyard to find something of use Reusable Software Designed for Several Applications Good Standards for – Documentation – Test Procedures – Test Cases Like looking in a library to find something of use

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 12 1/11/2004 Documenting Reuse Estimates The Total column is total delivered units of Software This could be applied equally to lines of code, function points or other measures ItemNewModifiedReusedTotal Component Component Component Component Component Total

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 13 1/11/2004 How Do You Count How Much of a Component is Modified or Reused? Consider Component #4 and Component #5 on the previous slide A Rule of Thumb: –Go to the smallest level known Unit or module (typically about 100 LOC) –If the unit is not changed, it is “reused” –If the unit is changed, it is “modified” –If more than 50% of the unit is changed, it is “new”

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 14 1/11/2004 Incorporating Several Distinct Kinds of Modified Software ItemNewModified Type 1 Modified Type 2 ReusedTotal Component Component Component Component Component Total

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 15 1/11/2004 Total Delivered Size Conversion Process Equivalent Size Reuse Information Calculating the Benefit of Reuse After estimating size, and before estimating effort, you must convert reused software to “equivalent” new software

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 16 1/11/2004 How Do You Convert? The conversion factor is based on how much less effort you will expend for reused or modified software than for new software Assuming you have historical justification of the conversion factors, you can do a simple calculation

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 17 1/11/2004 Example This can be estimated directly via “reuse factors,” e.g.: –Reused software takes about 30% as much effort as new software –Modified software takes 60% as much effort as new software

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 18 1/11/2004 Example Concluded This says that the total effort to develop these 5121 units of software will be comparable to that of producing 3567 units of new software. The same approach can be used for any size units: lines of code, function points or any other measures of software size.

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 19 1/11/2004 Where Did Those “Reuse Factors” Come From? Experience! –Over many hundreds of projects –But these are averages, and they vary a lot Your experience may vary from mine –You must keep track in order to know Different kinds of modified software may have different factors

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 20 1/11/2004 Typical Reuse Factors (*) See later slides for information on concurrent reuse

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 21 1/11/2004 We Can Get More Accurate If we are willing to look more closely at the –Process & –Reuse characteristics, Then we can gain a deeper understanding of the reuse impact We can also use this information to calculate the reuse factors used in the previous example

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 22 1/11/2004 First Examine the Process List the steps of the process Then determine the % of the total effort expended in each step, when developing new software Note that this is effort, not time Example:

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 23 1/11/2004 Next Develop a Factor to Represent the Reuse Benefit in Each Process Step Require- ments DesignCode and Unit Test Integrat ion Relative Effort 18%25% 32% Reuse Factor 10%15%5%75% Modified Factor 40%35%55%80%

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 24 1/11/2004 For Example … Let RM be the % of the requirements effort for this reused software –We must analyze requirements to see if reused software can be used Let DM be the % of the design effort for this reused software –We may have to plan to test the software and maybe design the rest of the software in a special way Continued...

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 25 1/11/2004 Example (continued) Let CM be the % of the coding and unit testing effort required for this reused software –Normally 0 for “pure reused” software, but what about new test code? –Typically the % modified for modified software Let IM be the % of the integration effort required for this software –often % even for pure reused

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 26 1/11/2004 Applying the More Accurate Method For each phase of the process –Determine % effort for each process phase –Determine the effect of reuse

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 27 1/11/2004 Why is it “Modified” if You Change it Only a Little Bit Totally reused software has: –Identical documentation –Identical test procedures, code, etc. –One master copy to maintain in the configuration management library –One part number for record keeping, inventory, etc. Continued...

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 28 1/11/2004 Even if you Change only 1 Comment Line You need to maintain two copies in the CM library (of code, test software, etc.) And if you change 1 line of executable code –You need to change tests and documentation

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 29 1/11/2004 What Is Concurrent Reuse? “Concurrent” reuse is reusing something multiple times within the same software product –For example, if the same subroutine is used in each of several system components It has very low cost, but it is not “free” –You have to integrate and test each component –And if you find a problem, you must fix it multiple times

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 30 1/11/2004 Reuse Factor To Use for Concurrent Reuse Typically, a concurrent reuse situation calls for a factor no larger than 5% –If the initial code is new or heavily modified, the second and subsequent uses (concurrent reuse) tend to need factors of about 3-5% –If the initial code is reused or lightly modified, the proper factor tends to run from 0.5% to 2%

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 31 1/11/2004 Concurrent Reuse Example ItemNewModifiedReusedConcurrent Reuse Total 2,1381,7821,2012,0007,121 Factor100%60%30%1% Net2,1381, ,588

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 32 1/11/2004 Maintenance vs Reuse? If you have a large amount of code and are modifying only a part of it, as in a maintenance situation, is this a form of reuse? If so, what reuse factor is appropriate? –Typically the factor would be very small –Often, people estimate maintenance with a bottom-up technique rather than as a form of “reuse”

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 33 1/11/2004 More Notes on Reuse The earlier in the process you reuse, the more leverage you get –Reusing an architecture or a design will support multiple target machines, languages, etc. Continued... Requirements Design Code Test

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 34 1/11/2004 A Final Note on Reuse The most important factor in planning for reuse is the application domain –A series of products in the same domain will get more reuse than a series of unrelated products –And it will be easier to find what you need when you need it –And the staff are more likely to be familiar with the old software and how it can be used again

Copyright © , Dennis J. Frailey, All Rights Reserved Day 2, Part 1, Page 35 1/11/2004 Architecture of Spreadsheet Cocomo Based Estimate Other Effort Estimates... Analogy based Size Estimate Software Reuse Analysis Final Effort Estimate Productivity Based Effort Estimate Other Size Estimates... Final Size Estimate Expert Based Size Estimate Size / Reuse EffortSchedules Generic Schedule Effort Schedule