200interview questions and answers.docx

200interview questions and answers.docx

《200interview questions and answers.docx》由会员分享，可在线阅读，更多相关《200interview questions and answers.docx（24页珍藏版）》请在冰豆网上搜索。

200interviewquestionsandanswers

Q:

Whatisverification?

A:

Verificationensurestheproductisdesignedtodeliverallfunctionalitytothe

customer;itinvolvesreviewsandmeetingstoevaluatedocuments,plans,code,requirements;thiscanbedonewithchecklists，walkthroughsandinspectionmeetings.

Q:

Whatisvalidation?

A:

Validationensuresthatfunctionalityistheintendedbehavioroftheproduct;validationinvolvesactualtestingandtakesplaecafterverificationsarecompleted.

Q:

Whatisawalk-through?

A:

Awalk-throughisaninformalmeetingforevaluationorinformational

purposes.

Q:

Whatisaninspection?

A:

Aninspectionisaformalmeeting,moreformalizedthanawalk-throughandconsistsof3-10peopleincludingamoderator,readerandarecorder.Thesubjectoftheinspectionisadocument,suchasarequirementsdocumentoratestplan.Thepurposeofaninspectionistofindproblemsandseewhatismissing,nottofixanything.Theresultofthemeetingshouldbedocumentedinawrittenreport.

Q:

Whatisquality?

A:

Qualitysoftwareissoftwarethatisreasonablybug-free,deliveredontimeand

withinbudget,meetsrequirementsandexpectationsandismaintainable.

However,qualityisasubjectiveterm.Qualitydependsonwhothecustomeris

andtheiroverallinfluenceintheschemeofthings.

Eachtypeofcustomerwillhavehisorherownslantonquality.Theaccountingdepartmentmightdefinequalityintermsofprofits,whileanend-usermightdefinequalityasuserfriendlyandbugfree.

Q:

Whyaretheresomanysoftwarebugs?

A:

Generallyspeaking,therearebugsinsoftwarebecauseofunclear

requirements,softwarecomplexity,programmingerrors,changesin

requirements,errorsmadeinbugtracking,timepressure,poorlydocumented

codeand/orbugsintoolsusedinsoftwaredevelopment.

?

Thereareunclearsoftwarerequirementsbecausethereis

miscommunicationastowhatthesoftwareshouldorshouldn'tdo.

?

Softwarecomplexity.Allofthefollowingscontributetotheexponential

growthinsoftwareandsystemcomplexity:

Windowsinterfaces,client-

serveranddistributedapplications,datacommunications,enormous

relationaldatabasesandthesheersizeofapplications.

?

Programmingerrorsoccurbecauseprogrammersandsoftware

engineers,likeeveryoneelse,canmakemistakes.

?

Astochangingrequirements,insomefast-changingbusiness

environments,continuouslymodifiedrequirementsareafactoflife.

Sometimescustomersdonotunderstandtheeffectsofchanges,or

understandthembutrequestthemanyway.Andthechangesrequire

redesignofthesoftware,reschedulingofresourcesandsomeofthe

workalreadycompletedhavetoberedoneordiscardedandhardware

requirementscanbeeffected,too.

?

Bugtrackingcanresultinerrorsbecausethecomplexityofkeepingtrack

ofchangescanresultinerrors,too.

?

Timepressurescancauseproblems,becauseschedulingofsoftware

projectsisnoteasyanditoftenrequiresalotofguessworkandwhen

deadlinesloomandthecrunchcomes,mistakeswillbemade.

?

Codedocumentationistoughtomaintainanditisalsotoughtomodify

codethatispoorlydocumented.Theresultisbugs.Sometimesthereis

noincentiveforprogrammersandsoftwareengineerstodocumenttheir

codeandwriteclearlydocumented,understandablecode.Sometimes

developersgetkudosforquicklyturningoutcode,orprogrammersand

softwareengineersfeeltheyhavejobsecurityifeveryonecan

understandthecodetheywrite,ortheybelieveifthecodewashardto

write,itshouldbehardtoread.

?

Softwaredevelopmenttools,includingvisualtools,classlibraries,

compilers,scriptingtools,canintroducetheirownbugs.Othertimesthe

toolsarepoorlydocumented,whichcancreateadditionalbugs.

Q:

Whatisagoodcode?

A:

Agoodcodeiscodethatworks,isfreeofbugsandisreadableand

maintainable.Organizationsusuallyhavecodingstandardsalldevelopersshould

adhereto,buteveryprogrammerandsoftwareengineerhasdifferentideas

aboutwhatisbestandwhataretoomanyortoofewrules.Weneedtokeepin

mindthatexcessiveuseofrulescanstiflebothproductivityandcreativity.Peer

reviewsandcodeanalysistoolscanbeusedtocheckforproblemsandenforce

standards.

Q:

Whatisagooddesign?

A:

Designcouldmeantomanythings,butoftenreferstofunctionaldesignor

internaldesign.Goodfunctionaldesignisindicatedbysoftwarefunctionalitycan

betracedbacktocustomerandend-userrequirements.Goodinternaldesignis

indicatedbysoftwarecodewhoseoverallstructureisclear,understandable,

easilymodifiableandmaintainable;isrobustwithsufficienterrorhandlingand

statusloggingcapability;andworkscorrectlywhenimplemented.

Q:

Whatissoftwarelifecycle?

A:

Softwarelifecyclebeginswhenasoftwareproductisfirstconceivedandends

whenitisnolongerinuse.Itincludesphaseslikeinitialconcept,requirements

analysis,functionaldesign,internaldesign,documentationplanning,test

planning,coding,documentpreparation,integration,testing,maintenance,

updates,re-testingandphase-out.

Q:

HowdoYouIntroduceaNewSoftwareQAProcess?

A:

Itdependsonthesizeoftheorganizationandtherisksinvolved.Forlarge

organizationswithhigh-riskprojects,aseriousmanagementbuy-inisrequired

andaformalizedQAprocessisnecessary.Formediumsizeorganizationswith

lowerriskprojects,managementandorganizationalbuy-inandaslower,step-by-

stepprocessisrequired.Generallyspeaking,QAprocessesshouldbebalanced

withproductivity,inordertokeepanybureaucracyfromgettingoutofhand.For

smallergroupsorprojects,anad-hocprocessismoreappropriate.Alotdepends

onteamleadsandmanagers,feedbacktodevelopersandgoodcommunication

isessentialamongcustomers,managers,developers,testengineersandtesters.

Regardlessthesizeofthecompany,thegreatestvalueforeffortisinmanaging

requirementprocesses,wherethegoalisrequirementsthatareclear,complete

and

testable.

Q:

Givemefivecommonproblemsthatoccurduringsoftwaredevelopment.

A:

Poorlywrittenrequirements,unrealisticschedules,inadequatetesting,adding

newfeaturesafterdevelopmentisunderwayandpoorcommunication.

1.

Requirementsarepoorlywrittenwhenrequirementsareunclear,

incomplete,toogeneral,ornottestable;thereforetherewillbeproblems.

2.

Thescheduleisunrealisticiftoomuchworkiscrammedintoolittletime.

3.

Softwaretestingisinadequateifnoneknowswhetherornotthesoftware

isanygooduntilcustomerscomplainorthesystemcrashes.

4.

It'sextremelycommonthatnewfeaturesareaddedafterdevelopmentis

underway.

5.

Miscommunicationeithermeansthedevelopersdon'tknowwhatis

needed,orcustomershaveunrealisticexpectationsandtherefore

problemsareguaranteed.

Q:

Givemefivesolutionstoproblemsthatoccurduringsoftwaredevelopment.

A:

Solidrequirements,realisticschedules,adequatetesting,firmrequirements

andgoodcommunication.

1.

Ensuretherequirementsaresolid,clear,complete,detailed,cohesive,

attainableandtestable.Allplayersshouldagreetorequirements.Use

prototypestohelpnaildownrequirements.

2.

Haveschedulesthatarerealistic.Allowadequatetimeforplanning,

design,testing,bugfixing,re-testing,changesanddocumentation.

Personnelshouldbeabletocompletetheprojectwithoutburningout.

3.

Dotestingthatisadequate.Starttestingearlyon,re-testafterfixesor

changes,andplanforsufficienttimeforbothtestingandbugfixing.

4.

Avoidnewfeatures.Sticktoinitialrequirementsasmuchaspossible.Be

preparedtodefenddesignagainstchangesandadditions,once

developmenthasbegunandbepreparedtoexplainconsequences.If

changesarenecessary,ensurethey'readequatelyreflectedinrelated

schedulechanges.Useprototypesearlyonsocustomers'expectations

areclarifiedandcustomerscanseewhattoexpect;thiswillminimize

changeslateron.

5.

Communicate.Requirewalk-throughsandinspectionswhenappropriate;

makeextensiveuseofe-mail,networkedbug-trackingtools,toolsof

changemanagement.Ensuredocumentationisavailableandup-to-date.

Usedocumentationthatiselectronic,notpaper.Promoteteamworkand

cooperation.

Q:

Doautomatedtestingtoolsmaketestingeasier?

A:

Yesandno.Forlargerprojects,orongoinglong-termprojects,theycanbe

valuable.Butforsmallprojects,thetimeneededtolearnandimplementthemis

usuallynotworthwhile.Acommontypeofautomatedtoolistherecord/playback

type.Forexample,atestengineerclicksthroughallcombinationsofmenu

choices,dialogboxchoices,buttons,etc.inaGUIandhasanautomatedtesting

toolrecordandlogtheresults.Therecordingistypicallyintheformoftext,

basedonascriptinglanguagethatthetestingtoolcaninterpret.Ifachangeis

made（e.g.newbuttonsareadded,orsomeunderlyingcodeintheapplicationis

changed）,theapplicationisthenre-testedbyjustplayingbacktherecorded

actionsandcomparedtotheloggedresultsinordertocheckeffectsofthe

change.Oneproblemwithsuchtoolsisthatiftherearecontinualchangestothe

productbeingtested,therecordingshavetobechangedsooftenthatitbecomes

averytime-consumingtasktocontinuouslyupdatethescripts.Anotherproblem

withsuchtoolsistheinterpretationoftheresults（screens,data,logs,etc.）that

canbeatime-consumingtask.

Q:

Whatmakesagoodtestengineer?

A:

RobDavisisago