PART 1: Data and Information ManagementARTICLE IS BELOW4 DETAILED PARAGRAPHSSource code management becomes an important tool as a programmer begins developing complex code. As code develops and changes are adopted, a programmer needs to be able to review the changes made to determine how they impacted results of the code.Answer the following question:How can source code management improve the ability of multiple programmers working together on a piece of code?Project Management3 DETAILED PARAGRAPHSConsidering the roles and responsibilities explored in Module One, discuss the project manager’s role in managing charter elements (e.g., initial risks, assumptions, change control) and meeting milestone expectations defined in the project charter.The project manager is typically appointed through approval of the project charter. In responding to your peers, discuss possible actions taken to ensure continuation of project objectives when the project manager is replaced. Your discussion should include how the change is captured in the project charter
ARTICLEMention source-codemanagement (SCM) to mostdevelopers, and there is thestrong possibility that they willthink about version control ofindividual source-code files. Atits most basic, version control isa method of maintainingseparate versions of files, ratherthan simply overwriting the oldones.if SCM is regarded as nothingmore than a version-basedsafety-net for over-enthusiasticedits, or a cumbersome methodthat over-complicates developmentprojects, then managingyour software configuration canbecome costly in time and effort.SCM can go a lot further, and amore comprehensive SCMsystem can lead to happier andmore productive developers.the most up-to-date SCMsystems offer developers andproject managers a way ofimplementing an efficient andproductive workflow: they gobeyond simple version control -they can streamline projectmanagement. SCM should alsobe part of every professionalprogrammer’s ‘skillset’;however, the discipline is stillnot always featured as anintegral part of academicsoftware educationprogrammes. Many – maybemost – programmers becomefamiliar with the SCM’s principlesin their first jobs, wheretrainee developers are introducedto the basics by moreseasoned software engineers. Asolid working knowledge ofSCM usually comes whenprogrammers assume a moreactive role in project delivery, asthat enables them to see the fullreach of SCM in theirorganisation.Like any computer sciencediscipline, SCM has evolved overthe last 50 years (see timeline,pages 52-53), but its objective hasalways been informed by arequirement to enable softwareprojects to be completedproperly to schedule, and toenable multiple contributionsby software specialists who,although working on the sameassignment, may be geographically-dispersed for the durationof completions; as such, thetenets of SCM have played a partin defining the dynamics ofmanaging global project teams. A modern SCM system alsoenhances traditional versioncontrol with capabilities such asworkspaces, changelists,branching, ‘lazy copying’, andcollaborative development. Aworkspace is the area on thedeveloper’s desktop where theSCM renders private copies ofthe specific versioned files thedeveloper requests. This allowsthe developer to work at his orher own pace without inhibitingthe work of colleagues. As thedeveloper has private copies, heor she can also accomplishbuilds and unit tests beforechecking in files.Changelists allow developersto check in their changes into theSCM as easily referenced, transactionalunits of work. If achangelist number is assigned tothe transaction, then the developerknows that his or her work isstored safely within the confinesof the SCM database. A changelistconsists of the set of filesthat have been changed, theirrevision numbers, and a descriptionof the work performed. Basicfile operations common to allSCM systems (such as adding,editing, deleting files and backingout changes) are recorded aschangelists to help answerquestions such as: Which otherfiles were affected by thischange?; Which defects does thischange fix?; Which releasescontain the fix for this defect?;Which releases do not contain thefix for this defect?Working without usingchangelists can be timeconsumingand confusing, asdevelopers have to keep manualrecords of their work, andcannot easily see how differentfile revisions are related. Moreadvanced SCM systems ensurethat if a changelist affects anumber of files then either thechanges for all the files aresubmitted, or none are. If thenetwork connection is interruptedduring changelistsubmission, the entire submitfails, and the files return to a’pending’ state to allow anyissues to be resolved. Thisfeature removes the worry ofpartial check-ins completelyfrom developers’ minds.One use case for a changelist isto create a ‘branch’ by taking acopy of entire areas of the codebase at specific points in time.Branching allows these copies (or’codelines’) to diverge over time,while offering the option ofmerging selected changesbetween them. A commonexample is that of propagating abug fix from a release codelineback into a development codelineso it can be included in a futurerelease. The SCM system recordswhich changes have been propagatedand removes the need torepeat the same developmenteffort in each branch.Branches inherently havehistory, so changes to files,additions and deletions can allbe managed within the contextof the branch. Merging changesbetween branches is mucheasier: the SCM system uses theinformation it has about thehistory of the branch, and thepoint at which the files forked.In this way, many merges can beperformed automatically andhuman intervention is onlynecessary when two changeshave affected the same part of afile – a so-called ‘conflict’.’Lazy copying’ ensures that thesize of the code repository doesnot grow unnecessarily whenbranches are created by storing apointer to the source file(s) ratherthan creating a physical copy. Asthe variant is modified, all that isrecorded are the differencesbetween the modified version andthe original.If large code frameworks areused to support developmentand a branch is needed for thepurposes of rapid prototyping (acornerstone of agile developmentmethods), then attemptingto copy that framework in itsentirety is a long-windedexercise. It is not unusual foreven extensive modifications toa large code framework to resultin a divergence of just a few percent between variants. Becauselazy copying only requires spacefor a tiny proportion of theoverall code base, the storagesavings can be immense.Other advantages thatbranching provides becomecommercially significant whenyou consider collaborative developmenteffort going beyond oneoffice. There might be developersworking on the same project indifferent countries, and alsosubcontractors who can be givenaccess only to portions of the codebase. Traditional ways ofhandling this involve locking filesand waiting for others to communicatetheir status.TiME zONESThis can lead to delays,especially if engineers areworking in different time-zones.SCM that goes beyondversioning files makes itpossible to ‘back out’ of a changeeven when other modificationshave been made to the sourcecodemodule. It works by takingthe state of the system back tothe point before the bad changewas made. It is possible to skipthat edit, but then roll forward,accepting later ones so theresulting file has all the correctchanges in place.Sometimes a change is effective,but it is also far from pretty.Consider a product that isalready shipping, but someusers have encountered a bug. Inthese cases, the emphasis isoften on speed rather thanelegance. The expedient fixmight be simply to repair theactions of the bug rather thanseek out its true cause, whichmight be in some other, possiblyquite distant module. This kindof fudge works but is notsomething that should be propagatedto a new version of aproduct. Some SCM systemsoffer the ability to mark these’ugly code’ changes so they donot propagate to later releases.When copying the data across,these ‘ignored’ changes arepassed over by the system,leaving ‘clean’ code ready toreceive the correct remedy.It is possible to label sectionsof code manually to show developerswhich ugly fixes they needto roll back although thisinvolves a lot more intervention.The idea of having differentcodelines – for shippingproducts and those in development- is an important concept.in SCM. Many SCM systemsadvocate the use of ‘labels’ – anamed list of revisions of files -for tracking product releases.The big problem with labellingis that labels can be altered atany time; thus it’s possible to’lose’ the records of an entirerelease, or worse still, to deliverbugs to customers because therecords for a release havebecome inaccurate.There are other issues withthe labelling approach as labelsthemselves have no history.Usually another label is neededto record every new state and soan explosion of labels results -often with a convoluted namingconvention to help people makesome sense of the enormous list.An Intelligent branching modelwill help remove the issues thatcome from relying too heavilyon labels.OPEN-SOURCE CODESCM also quickly handles morecomplex situations such as whensome of the code to beincorporated is derived fromopen-source software. Changesmade to code provided under aGNU licence, for example, willneed to be supplied on demandto other users of that sourcecode. Some commercial sourcecodelicences also demand thatmodifications be provided to thevendor.Traditional methods mightrequire all the code to be kepttogether. Having third-partycode in a separate branch inyour SCM system is critical inensuring all contractual obligationsare met without confidentialinformation being disclosedat the same time.When developing on morethan one platform, the differencesin the line-ending conventionsacross platforms can leadto some annoying problems.SCM systems can resolve thedifferences between developmentplatforms because, byunderstanding the conventionsof the platform, the SCM systemcan automatically translatebetween the required formats.This reduces the chances of abuild failing later if theplatform used for compilationdiffers from that used forediting.SCM need not only be a way tomaintain source code, but canalso provide intelligence on theprogress of entire projects. It isnot a fancy name for versioncontrol, and it offers time andeffort savings over more traditionalmethods. For advancedusers, SCM forms the underpinningfor successful projectmanagement because of thestructure it provides. If you everfind yourself adding commentsor changing file names to addthings like ‘Version 2.1.2 fix’,software authors and softwareproject managers need to takeanother look at SCM and what itcan do for improving the qualityof their workflow.
