Design Reviews |
 |
 |
 |
| Written by Graham Stoney | |
|
Reviews are an important part of any efficient engineering process. Design reviews in particular play an important role in ensuring that the item under review has been designed in a manner which is most likely to meet its requirements and do so as efficiently as possible. Design reviews form part of the verification and validation activities of the project. Many engineers think of testing as the sole method of verification and validation, but including design reviews early in the development process provides the most efficient complement to the testing effort. Well run projects have a balance of both design reviews and testing. The key strength of design reviews is that they can be conducted much earlier in the project than testing can, since they can occur before anything has been built that is ready to test. Catching design errors or inefficiencies early in the project saves valuable time and effort since design errors can have a cascade effect and the cost of correcting them often multiplies as each phase of the project progresses. Design reviews are cost effective even though they involve multiple engineers whose time is valuable, because they save this cascading cost of undetected design anomalies. Engineering design is a creative process. It is a combination of science and art, where there is no single correct solution to any problem; engineering is all about trade-offs, constraints and dealing with risk and uncertainty. The goal is to create a design which meets its requirements as efficiently as is practicable given the present organisational and technological constraints. Engineers should be encouraged to take pride in their designs, and treat them as a work of art. Design reviews offer the opportunity for an engineer to exhibit their creation to an audience with the technical ability to appreciate it. The goal of the design review is to ensure that the engineer(s) have considered the options sufficiently and taken all the requirements into account in making their design choices. The aim is not to ensure that the design is “perfect”, but to ensure that it is reasonably close to ideal in terms of meeting its requirements, can be implemented efficiently, and has the level of maturity appropriate for the given point of time in the project lifecycle. The actual material being reviewed in a design review will vary depending on the maturity of the design, the branch of engineering involved and the technology used in its development. Requirements should be reviewed before a design based on those requirements is reviewed. Timely concurrent engineering requires that consideration be given to the design while requirements are being codified, but a design cannot be effectively reviewed until its key requirements are first agreed upon. Many disputes during design reviews are actually about what the design should do (i.e. the requirements) rather than how it should do it (i.e. the design itself). A design that exists only in the designers head is difficult to review, because it is difficult to disseminate the design information to the reviewers. That said, it can still be reviewed with the help of a white board in an interview-style; this is better than allowing undocumented work to proceed without any form of review. Engineers are sometimes reluctant to expose their work to scrutiny from other people, because it “isn't ready yet”. In some cases this may be true, but it's worth remembering that it is only at the very end of the project that all the information necessary to complete the ideal design with absolute certainty is available. Design reviews are a valuable way to elicit information about areas of the design in which there are unresolved technical risks that may need further investigation by way of prototyping, analysis or input from other people experienced in the field. One antidote to the desire for perfection before a design can be scrutinised is to establish a culture where review is a normal part of the engineering process. Treat documents and other design information as evolving entities and develop an evolutionary mindset, where every deliverable progresses through increasing stages of maturity during the project and get reviewed at key points along the way. Effort is likely to be wasted if a large amount of work proceeds on the detailed design of subcomponents in an architecture which then changes because it hadn't been reviewed and agreed upon. It is reasonable for many unanswered questions to remain after a preliminary review; indeed, one of the purposes of a review is to ascertain which questions in particular remain unresolved, so that they may be prioritised. The information from the design review can then flow into the project plan to ensure that all the important issues are given resourcing and attention as appropriate. Social Bookmark this page:         
|