#'erestimate+ ma% &ause the proje&t to take lo!ger tha! it woul otherwise –
–
"arki!so!s -aw+ (ork epa!s to )ll the time a'ailable Brooks -aw+ "utti!g more people o! a late job makes it better
Basis for Software Estimatio! •
•
/he !ee for histori&al 0ata + base o! past eperie!&e Measure of (ork+ –
S-#C1Sour&e -i!es of Coe2 •
•
•
•
•
–
3o pre&ise e)!itio! 0i4&ult to estimate at start of the proje&t #!l% a &oe measure1!ot eort2 "rogrammer epe!e!t 0oes !ot &o!sier &oe &ompleit%
F"1Fu!&tio! "oi!t2
Software Estimatio! /e&h!i$ues •
•
•
•
•
•
•
lgorithmi& Moels+ 1Eort 2 Epert 6ugme!t !alog% "arki!so!s+ Sta eort a'ailable to o the1S&ope2 proje&t "ri&e to (i!+ "rei&t to 1"ri&e2&ost re$uire. /op ow!+ #'erall estimate is broke! ow! o!to eort re$uire for &ompo!e!t tasks Bottom7up+ Compo!e!t tasks a! i!i'iual estimates are aggregate
pro&eural &oe orie!te approa&h •
•
•
•
•
Bottom up approa&h "rei&t the !umber a! t%pe of software moules i! the )!al s%stem Estimate the S-#C for ea&h moule Estimate the work1&ompleit% a! /e&h!i&al i4&ult%2 Cal&ulate the work a%s eort+ Co!'ert S-#C to eort
"arametri& Moel Eg+ 8!sura!&e to &o'er or rebuil the propert%, &ompa!% gi'es parameters base o! !o of stories, bri&ks b% la%er, &arpe!ter eort et&+ Eort 9 1S%stem Si:e2;1"rou&ti'it% rate2 <-#C1/housa!s of li!es=&oe2
•
•
•
•
•
•
Eg+ is eperie!&e ,>? a%s=@<-#C /herefore he re$. A? a%s for <-#C B is less eperie!&es a! takes 55=<-#C /herefore re$. 55;9@@? a%s to &oe <-#C "rou&ti'it%9 Eort=Si:e Eort 9 Co!st@1Si:e;Co!sta!t2