1CS 350, slide set 2M. OverstreetOld Dominion UniversitySpring 2005Readingz PSP text, ch. 4, 5, 6Planningz Period plansz How shall I spend my time?z Schedule orientedz Product plansz What are we going to do over the next several weeks?z Activity orientedz For any project, these 2 plans are coupled2Weekly Activity Summaryz First step in period planning:z Understand how you spend time currentlyz Start with Time Recording Logz Transfer to Weekly Activity SummaryExample: time log1xRead chap. 7Read T454:45 p4:00 p1/311xLect. ch 5 & 6Class757:00 p5:45 p1/291xComp & test p. 1Prog3510:25 p9:50 p1/30Code prog. 1Prog209:50 p9:30 p1/30Test data for p.1Prog3039:20 p8:45 p1/30Code prog. 1Prog5556:35 p5:30 p1/301xRead chap. 6Read T3511:05 p10:30 p1/28Prog 1 designProg3010:30 p10:00 p1/28Read Prog 1 req.Read S953,3,410:20 p8:45 p1/281xLect, chap. 4, 5Class807:05 p5:45 p1/272xRead chap 4, 5Read T12515,54:25 p2:00 p1/26UCcommentsactivitydeltaint. timestopstartdate63595205170155Totals4545SatFri140140Th7575Wed160953530Tu8080Mon125125Sun 1/26TotalStudySpecRead TextQuiz PrepWrite ProgClassTask DateExampleWeekly ActivitySummaryComments: 1. numbers made up2. recall rule-of-thumb: spend 2 hrs out of classfor each hr in class3Example (cont.)4200145250150Min.4200145250150Max.4200145250150Avg.4200145250150TotalPrevious Week’s Time4209514525170150Min.6359520525170155Max.5229517525170152Avg.10559536025170305TotalTask Class Wrt P Quiz Read T Read S Total Current Week’s TimesSemester total, avg,max & min!Ignore 0’s!Assume wehave onlyone weekAssignment:z Pg. 43.z Due next Wed., Jan. 26.z Must mail to [email protected] Planningz Some definitionsz Product: something you produce, usually for someone elsez Project: produces a productz Task: an element of workz Process: the steps to produce a productz Plans: the way a specific project is to be done; how, when, who, at what costsz Job: something you do; either a project or a task4Job number logz Used to record estimated and actual timesz Used for product planningz Goal: to have personal data on how long certain types of activities takez Must be based on sizez Maybe lots of other factors, but we start with sizez Used as key input for future time estimatesJob Number Examplez See detailed example, pg. 50 based on time log, pg. 51z Detailed instructions, pg. 52z See detailed term def’ns, pg. 53Partial example:programming data onlyJob Date Estimated Actual To date # Time Units Time Units Rate Time Units Rate Max Min1 2/9 120 60 168 87 0.52 168 87 0.52 0.52 0.52Description: write program 1 (minutes per program)... (other activities types, so omitted on slide)3 2/15 205 148 245 133 0.54 413 220 0.53 0.53 0.52Description: write prog. 2...6 2/24 150 87 145 59 0.40 558 279 0.50 0.53 0.40Description: write program 310 3/6 175 95 230 104 0.45 ___ __ ___ ___ ___Description: write prog. 45Quick Survey Resultsz What’s best prog. language?z a. C++ 24z b. Java 1z c. Perlz d. Visual Basic 2z e. Other 1zYou prefer to use:z a. C++ 24z b. Java 1z c. Other 3zYour largest program:z a. < 500 loc 1z b. > 500 & < 1 kloc 10z c. > 1 & < 10 kloc 11z d. > 10 & < 100 kloc 5z e. > 100 kloc 1zWhat’s a kloc?z a. comments inc.? 5z b. declarations? 4z c. only exec. stmts? 14z d. number of CRs? 5Mail to [email protected] After a while, you will have data onprogrammingz avg, min, maxz (really need better units than “program”)z Time to read a text chapterz better units are probably pages rather than chaptersz And other activitiesz This can help with time estimates for future similar tasksUnits for readingz Table 6.1 has student data on minutes/page.z Note variancez minutes/page varies from 3.3 to 7.4 for this studentz Still, can be helpful to distinguish between time to read large chapter and time to read small.6Plot of reading timesReading Times0204060801001201400 102030pagestim ePlot of programming time, C++ Objects0501001502002503003504000204060Time (hours)LOCUnits for programmingz LOC looks like a good predictor of timez It is. Not perfect, but lots of studies show a strong correlation between LOC and development timez The problem is we only know LOC when we’re done. We’re looking for something to predict the time required for a task.z So don’t know LOC when we need themz A solution is discussed later7What’s an LOC?z Depends on formattingz See text for examplesz Depends on what you countz Comments?z Only executable lines?z Compiler directives?Basic metricsz For our purposes LOC:z As long as you are consistent, it will workz For CS 350, use "grep ";" *.h *.cpp | wc –l"z Why? it's easy and good enough. then we all do it the same way.z Basic productivity measure is LOC/hrz Usually misleading to compare your numbers and my numbersz Count loc differentlyz Likely depends on problem domainz Depends on what’s included in your hour vs mine:z Testing, Documentation, Code/design/test plan reviews (etc.)?Estimating Backgroundz Estimating models in other fieldsz Large base of historyz In wide usez Generate detailed planning dataz Require a size estimate as inputz Software size estimating experiencez 100%+ errors are normalz Few developers make estimatesz Fewer still use orderly methods8Size Estimating Principles - 1z Estimating is an uncertain process.z No one knows how big the product will bez The earlier the estimate, the less is knownz Estimates can be biased by business and other pressures (sometimes called ‘‘gutless’’ estimations)z Estimating is an intuitive learning process.z Ability improves with experiencez Some people are better at estimatingPlanning Forms, assignment 1z Plan/size est.z How big (LOC)z How long (minutes)z Base this on your analysis of requirementsz And your experience in writing other programsz And your impression of how long those tookDefect Record Log (pg 47) - 1z Assumptions: After writing your code, you type it in, your proofread your typingz You may do assignment in steps, so this may be all of the program or only part.z Any change made after proofing get recorded on formz If you rely on compiler to proofread, then every thing it finds that you must change appears on form9Defects - 2z Datez Number: