™~ Software Engineerin 9 2 pO Tr Metric y —h ic ‘stem, compo, e to which a sy: : nen : itativ, -asure of the degre “a netic ange proc = ee a attribute, Simply ae directly obs: en = i of Soe or product, ‘Metrics can : a TV abl nites or ee be derived from one or more directly o peaks i e Wantitic, quanti ine of co nu Examples of raw Metrics include the number of dineloficoc A Mber 0 document pages, etc. cha Examples of derived metrics include number of er! rors found per Perso, months, line of code per staff-hour, etc. Indicator @ term indicator is used to d The Ties of metrics thar Provide enote a representation of Metric or ; Yet Or process ji insight into an ongoing software developmen Indicators are metrics in a for(Motrig ) vehicle nvaleny, coy, Ihina Inqroe tO M0, © quantity nrewmury of The Pw vd, ah MELE ¢ Ne oy \reeionen tintin anes Oe inp tater He any WY pean agen all! i 1 atari pan be directly bag uw rectly obmervabte diay, my y tine ef codes, Hung wharavety NCTOHIHO GF a procean OF PE : wear mone dl is doy WaNtitie, Fn ities or ean toy ehorived ra a AM Sox , wn ¢ Ae MUDIOS of naw mnerien aretice the nul ; THMONE panes, ote f 4 a sprory fou i yhor oF CFTOry Md Por pay 1 by EXAM plos of derived metrics tnetude BUM m OCHS, Tne of core por slafh hour eles Indicator apresentition of me The term indicator fs used to denote @FEPMNT ti of a y At a vol 4 Series of metrics that provide insight into an ongoing, ee ( OP Mey Pajee icators are Mel i Project or process improvement activity. Indicators 8 fo suitable for assessing project behaviour or process impre ud » The Y enaby 5, the project or the product to mai sd ona project include actual Vers Ky 8s anned staffing, ele, the project manager to adjust the proc things better, Examples of indicators ¢ Planned task completions, actual ve! soflw sus ph are engineer t0 gain insight inig th e Process indicators enable ¢ ing, proc efficiency of an exis Project indicators enable a software project manager to 3 assess the status of on going project x Track Potential ri s& Uncover problem areas before they go critical x Adjust workflows 3c Evaluate the project team’s ability to control quality of products me 4 g SOFTWARE MEASUREMENT pes of measurements - Direct measurement ani There are two ty ndirect measurement.vy Measurement ay aneatt net attribute. we 1 A peightey The direct i tn ies cost and eftort aol ™ applied 1 Sot uct execution speed ogues , ‘ ares ree 1, me rect Measurement gree are some attributes of entithe

it should not demant metrics should not be based op 2 Objective and Consistent. The ; eA jecaats it should give dear results. An independent third i jerive the same metric value using the sam: parr sould 5 2 Exsily Available. The metrics should be available at a reasonal ast. 4. Weld. The metric should measure what it is intended to measurpoasuroront and Motrics i | pobus ie metric should be ; able to withstand minor 4 product yr $s ’ , software metric is an altem; pt to me. easure or predict i predict some attribute any oduct uct 0 cess, TI yon ze set Pca Software Ei 42 on a set of metrics uni Engineering c i f en ee i, community ha eed sople have come up aoe accepted by the field plete their own ways to measure th al an asure the different Bye an? files of a software. as tandard product ptrics are : an () Token metrics Function point and extended function point 0 (@ Bang, metrics ; wa code Complexity Metrics” VL, ] SL) 7 Ai 7 V wv @ Cyclomatic complexity 4 a she ot a J n_is_the process of predicting the siz software prod uct. (Accurate size estimation is crtical to effectively manage process) {the Project planet, must understand eneratean estimate of its size the software developmen are to be built and g hoe is only as good the scope of the softw before a project estima Project estimate iecomplished. as ea of the size of pre ork to | 4A 1 Code (LOCH / LOG: of the eau nd sim lest’metric size of a compuler apersizevor estimating the efort and 5 = te oP : Software size estimatio te can be made} A’ .17Lines program. jowevel,there is no standard definition of what makes 7 are oe D He ; itferent workers for the sante program may obtain ene gitferent wo I weightage is given to each line of code White the p hat sone aa nltage 2 Program are more difticun to cane other) Bespte all this, this metric conti : : bya cale a comparing the tig Ye Popular tool for calcu lating ad oe ‘ring ire Productiy; Y Programmers, which can be measure as ine Of cod le/man-mo, ay Onth) © oy \The LOC is often use esting-and maintentice Phaseg only to specify the Ware product but also it canbe rhe together with other metrics to analyse Other aspects oF its quality and of Aadvantages of Loc — wa Simple to measure 5 Disady, de Yes to A, th, d during the t size-of the Softy, antages of Loc . It is Programming language dependent, um 3 Does not Modate NO Procedural. language: - accom * Poor software design may lead to excessiv OF code. ve and unnecessary ling — 4.11.2, Token Metrics The Major drawback in LOC size measure is alike. In M.H. Halstead, Proposed one of the token metric where the size ofa gram, which Consists of the Number of unique toke! Ms of : a ns can be defined in aNe Nien, At ‘oun of \ LOCCUTLENCE OLoper 8. Dporat ae “oun of ota perands tal Secure fice tof operand t a ( (aware any symbi Pa wt u eywa wg ra consist of ord ina Mlle “0 il ‘ital apis specifies an ae and speci such as 4, pecial sy Mbols such as by é en 2 a ) wt) Ai we Th f op" orl ——15 «fonction Point (FP) , Hyon point (1). was developed 4 at a nt to overcome difficulties associated w ill N and fo a assist in developing ames " e ‘em to ae yn “gee ru oe wih software de it me f development) . jg-an objective and structured ety messi tasedS” wf (p ese quantifying its functionality provided | ys s the , syste ak Wie , wi a and logical design: his technique break: er components so that they’can be better understood and analysed} ies ial 5, thus divides the system into five basic = The number of edges of the graph NN > The number of nodes of the graph 7 ons P > The number of connected componen i the pro ram, ‘P’ can also be Considered as the number of exits from prog 7 led, it « Once the cyclomatic complexity of a code has been com Puted, it can then be Compared to the complexity of other programs and to the standard ‘ange given in the table below : Cyclomatic Complexity lO ve M9 21-50 vA Ve Sreater than 59 J Code Complexity tA simple Program, without much risk More comple x, moderate risk Complex, High risk Untestable very high risk =<. Fig. 4.4, Cyclomatic Complexity, Another alternative Way to determine the metric icv. _ lecision Points (conditianatsi i alll VAIO EN Gh Ball ‘AN, i | Ny 02 He Wi ing ‘Aly 4.12.2, Information row ia also a On r Ae fired Opies } aly it y | mara fw ae vale anette id Kafora’s Vari jy Ang thé COMpLESity of H nofiware modi fenry’o and Bale Katina and Henry, and ta ale knows a ‘anout Complexity metic) ber of calls from a MON ile Their metho SUgenta identifying the inaigneh and Wentitying the (te, the flows of local information entering, “of local Information Min Number of Cally from ay module (1c, the ows fan-out), by | He In determined by The Complexity jy Information flow metrle {s ro ut)? } in % Van-ou Gu (Procedure ly) % (ayn ae ime Where, ra a ' C+ The Complexity of the module- “, Procedure Length » if@ length of the module, Riven INLOC or by ~~ on sing Mccabe's cyclomatic complexity, a , Fan-in The number Of calls to the module, Fan-out + The number of calls from the module, Advantages of Information flow Metric 1. It takes into account data-driven Programs, 2 It can be derived prior £0 coding, during the design Stage, Disadvantages of Information flow metric