Article by Tom Greaves
The two main areas of software inspection are code review and peer reviews, both are unique and have different levels of efficiency. Recent studies suggest peer reviews are more efficient and cost effective, however this has not been fully confirmed. This article takes a look at the two processes in some detail in an attempt to establish a conclusion.
Code review is usually carried out as a special inspection in which a team of people will examine a portion of code and endeavor to remedy any defects within it.
A defect in a code review can simply be classed as a block of code which does not fully meet its requirements or which does not function as the programmer had initially intended, it could even be code that simply needs improving.
These teams, as well as fixing code can also be a great tool for improving the skills of developers and act as an ideal training ground to help junior programmers learn and implement new techniques.
Peer reviews are usually considered the industry best practice for finding and fixing software defects early and learning abut software artifacts. Software walkthroughs and software inspections make up an integral part of the software engineering process. In order to produce the most efficient peer code reviews it is suggested a collection of knowledge, skills, and behaviors are used in corollary to each other.
Software inspections are the most aggressive form of peer code reviews and will always fully implement all elements in finding defects in software. Software walkthroughs differ in nature and draw selectively upon the elements in assisting the producer to obtain the the widest understanding of a block of code and reaching an agreement between all participants.
All research has shown that peer reviews produce the most return on investment due to fast learning and the early detection of defects. The best results of peer review become apparent when reviews are brought forward within an organization through a defined program of preparation, training practitioners and managers, defining metrics and completing a database structure, and fully sustaining the roll out infrastructure.
It seems clear in todays world that on the topic of code review a more modern form of peer review simply out classes the more traditional formal review.
Disclaimer: All runs performed in Yugoslavia. Synopsis: Impromptu racing between a modded 2006 Lamborghini Gallardo and a modded Toyota Supra in the dangerous streets of Yugoslavia. Got a 25-80 pull, 45-130 pull, and a 45-110 pull. However, in the last two runs, I let the Supra spool before taking off. Mods and other information are below: Gallardo – catless exhaust (LOC +Fabspeed combo) – 125 lb weight reduction – one passenger – 6 speed manual – approx. 450awhp Supra – Greddy T78 Turbo @ 16 PSI – Supporting Modifications – 750cc Injectors – No passenger – Auto transmission – approx. 460rwhp Please subscribe if you enjoy cars, exotics, racing, and film-making. Please note: Since this was an impromptu race, I did not bring my HD camera or GoPro HeroHD setups. Therefore, only my intro scence is 1080P HD and the rest is crappy 640×480 iPhone 3GS footage rendered to 1080P HD. Subscribe and this race will go down again with a full camera setup and also the Supra getting more mods! Music – Custom Created in Sony Vegas Pro 9 using VideoCopilot’s Pro Scores Software.
Video Rating: 4 / 5
Related Software Articles