One of the jobs of working professionals in the design community is to attend design reviews. I imagine that many of us have attended design reviews in the past or will in the future. I am sure that those engineers who have been working in the industry for many years have attended so many design reviews they've lost count.
For those of us that have attended many design reviews, I would suggest in some or possibly many of the reviews attended there is a tendency to lose interest after the first 30 minutes. In addition, I would like to suggest that you would be able to testify to the endless slides of data presented in these reviews, much of which may seem relevant to the presenter but in reality are not that important.
If you are fortunate enough to have clear and concise design review meetings, I commend the management and technical leads that have instilled practices that produce crisp outcomes. However, if you are like the rest of the community of analog engineers, then you must wonder if there is a better way? Why does the presenter have a tendency to fill his or her presentations with reams of data that are difficult to see, or plots that are difficult to read? I would like to suggest that this tendency arises in part because the engineer believes in the necessity to present data showing how the simulations and data collection align over variations. Is there a better way to have design reviews? I believe there is and would like to suggest options.
Before I suggest my proposal for an improved design review methodology, I would like to set a precedent for conducting design reviews. First, the attendees in the design review must remember that its purpose is more than a check-off procedure, but to ensure that the product being designed meets the criteria deemed necessary by the goals established. This means that the engineer must feel comfortable exposing his or her weaknesses in the design without the threat of management blasting the engineer in the meeting. (If needed, that can be handled outside the meeting on a one-on-one basis.) Second, the design review attendees list should include a key technical guru to review the data presented. It is important for the design review to be crisp and to the point.
Now, to present my argument for a more concise presentation in the design review, I would like to introduce the idea of a one-pager. Obviously, it would not be possible to put all of the relevant data on one page for a design review, but the concept is important to strive to meet. The items presented as part of the concept of a one-pager should include a concise list of key topics relevant to the design.
For an effective and engaging design review, the one-pager idea should have a concise set of defined outcomes presented in very few slides or pages.
The one-pager should present:
- Design goal
- Architecture (also include difference from prior architecture)
- Pro/cons of design. Highlight weaknesses in design and how you overcame the design weakness
- Specification targets (including power and area targets)
- Design simulation summary in the form of a compliance matrix to design targets
- Special considerations for design (special power routing needs, etc.)
- Additional supporting material in backup as needed
In addition, the design review process outlined should also include the option for less detailed review during a feasibility study or initial stages of the design process. For instance, during an architectural review, a subset of the information may include just the architecture pros/cons and specification targets.
If design reviews are carried out in the manner suggested, the design community can engage in shorter discussions that are synergistic and can accomplish the same goal early in the process. Engaging the design community and management in changing the culture is not easy, but the benefits outweigh the pain associated with expense and time to implement costly design changes because critical design weaknesses were missed during the design review.
Let me know what you think. What methods have you tried to streamline design reviews?
Brandt Braswell is a distinguished member of the technical staff at Freescale.