T h u s it was natural to elease to release

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: dying o , s ar a r bu e p o c er ribut shown by the dashed line in Figure 4. T h u s it was natural to elease to release. individual a h ocorren t i n t x happly uni- and multivariate regression and putenomelaaion eecquisite detail. Similarly, a system, a REND VERAGE pro and t r a p eries models to niques to fit appropriate regressioncess, oime-shenomenon may be viewed from the outside, by d a general upward trend in the size, acts of observing; clarifying; and by measuring and modeling represent the process for purposes of planning, forecasting, and e systemmproving et in intenanrcas apwhole. A s the studtriprutgrs,spedteevi-, and trends. From such activities i and th ima part o e ro- identifiable at y b o e e s at , rns onents, dence aec, matlemdnhs,t instruc- ighteconsiider ancreasing knowledge and understanding, based on modulcs ust u atee t t a one m on obta ns i software maintein Figure ceanh envarnoempnramepers tha s a ehavior of both tore anistmm and its subsystems, the process t roject e b self-regulating h g sy s e , nan 1 . T de ha i c us ea t nds. Whencthte aapparentldatra were shand s,ts ut-bproallsseobeying its subje t o veraged y andom ock i b su over ce - s. e 2 , the wn specific cernaAVERAGEioELEASEI d and internal dynamics. t vat R aws o previously o rseircdatan lha SEQUENCE NUMBER LEGEND: Starting with the initial release of os/360 as a base, we have 0 SIZE IN MODULES I M s t L E S se L E D T h uODUheHANDfirst observations encstudieddtthe isearchVERAGE RELEASEI ourage h e nte aA tion between management and the evolution c for models +RELEASE INTERVAL SEQUENCE NUMBER data were tpepresentsedhown in tFatgoverofedLts/ed60amiusing veorain inofependent variablesof the improvetha r lotted a s laws higure n 0oEGEND:yODULES cbeha ci rt h IN d RELEASE SEQUENCE NUMBE Sl2E 3 M n by X MODULE of nonltihe metaossitblm exponential -on, people,S HANDLEDprogram material near-p sys e y of organizati m+ RELEASEdINTERVAL ncement (i. e . , maintenance) process. We cannot ent an and a enh trapolatiinvolved in the cuethiongand maintenance process, i n the evoluon suggested f r r at er rowth say at this time that we have used all the key independent varitly at odtds wifth rogrammingrsyntems. the then cu re s t proion o p ables. T h e r e is undoubtedly much more to be learned about the also highly erratic with major, but riz t programming proed, fluctuationsrhapswn ciessairy reo4eby varhabeewhy wethallege thattcnu-ractea w seofhe sho ne n F gu t xplain i er l s and e data con i ha I t is pe l cess. Our method of study has been that of regression-outside ase to rels acre.atievemaieltessanlm,ont enhancgeme4 t S e raa lvor a gprogramming Fi ure n i n de f e o u e s e N on, r t h n e n , a ce a s d p ogr in-which weohtavetrends eo f “structuredranaamsis.” Starting term d a ly with gr w h tTuesday, Septemberms.13t veis theacttraleds eaience ofall who have been ino the dsyste 10,oI ry cl r u en xp r s isplay f evolution Growth Trends . X studied the programming process:’ as it perpment of os/360, and now give a preliminary roject statistics of this programming system, survived a number of versions o r releases gan. T h e data for each release included meaf the system, the number of modules added, , the release d a t e , information on manpower used and costs involved i n each release. In large, apparently stochlonger ariations in the It takes astic, v and s from releaseonger ase.release the l to rele to the programming process What can you deduce from these graphs? Figure 3 • A v e r a gg r o w t h e trends of system attributes compared with p l a n n e d growth next release indicated a general upward trend in the size, ost of the system and the maintenance proThe dule of tatements by components, mosize s, sincreases, instrucover time handled in Figure 1 . T h e various parameters xpose trends. When the averaged data were in Figure 2 , tIterequires modifying ta had h previously erraticda smooth. more and more modules • • per plotted as shown in Figure ditional data were each release spicions of nonlinear-possibly exponential xity. Extrapolation suggested further growth gnificantly at odds with the then current proa were also highly erratic with major, but Tuesday, elated, fluctuations shown in Figure 4 by corr September 10, 13 I LEGEND: 0 Sl2E I N MODULES X M O D U L E S HANDLED RELEASE INTERVAL + I AVERAGE RELEASE SEQUENCE NUMBER 0 Sl2E I N MODULES X M O D U L E S HANDLED of nonlinear-possibly exponential + RELEASE INTERVAL rapolation suggested further growth ly at odds...
View Full Document

This document was uploaded on 03/16/2014 for the course EE 360f at University of Texas at Austin.

Ask a homework question - tutors are online