Integrated Solutions
Design Review Process
|
|
Prepared By: |
|
Lead Author, Integrated Solutions |
Date |
Approved By & Effective on Date of Signature Below: |
|
Yufen Chang Yu, Integrated Solutions |
Date |
THIS DOCUMENT IS TO BE REGARDED AS THE OFFICial DESIGN REVIEW PROCESS within INTEGRATED SOLUTIONS. IT IS TO BE MAINTAINED UNDER CHANGE CONTROL BY the DOCUMENT CONTROL COORDINATOR.
|
|||
TITLE: Design Review Process
|
|||
DESCRIPTION OF CHANGE |
|||
ISSUE |
DATE |
SECTION |
DESCRIPTION
|
1.0
|
3/13/97
|
All
|
|
1.1 |
10/3/97 |
6.3 |
|
Table of Contents
CHANGE SHEET
1.0 Introduction
2.0 The Review Process Overview
2.1 Entrance Criteria
2.2 The Meeting
2.3 Exit Criteria
2.4 Defect Ratings
2.5 Metrics
3.0 Configuration of the Review
4.0 The Role of the Author
4.1 Entrance Criteria
4.2 The Meeting
4.3 Exit Criteria
5.0 The Role of the Reviewer
5.1 Reviewer’s Checklist
5.2 Entrance Criteria
5.3 The Meeting
5.4 Exit Criteria
6.0 The Role of the Facilitator
6.1 Entrance Criteria
6.2 The Meeting
6.3 Exit Criteria
Additional References
Appendix A: Design Review Worksheet
This document is the Design Review Process by which design reviews shall be conducted. Studies have shown that such discussions are an essential part of software development as they lead to a better product in a shorter time.
This process is was developed by Integrated Solutions, Code Inspection Process[1]. Also consulted for this paper was Best Area Practice: Requirements Software Inspections[2] and Fagan Defect-Free Process[3].
Design Reviews will be done on all major design efforts. The CEO/President or customer will decide on when they must be done. When they are required, this process must be followed.
2.0 The Review Process Overview
The design review process is a step that occurs before coding. The design is either a High-Level Software Design Specification, a Detailed Software Design Specification, or a combination of the two. The objective of a design review is to make sure that the design is taking the right approach. That is, the design meets the project requirements, and the design is one that will provide customer satisfaction.
The review should not be considered a time consuming event that must tackle all the defects found. Rather, it should be a simple process of systematically going through the design and listing areas of concern. Each area will be rated.
2.1 Entrance Criteria
The author(s) and reviewers must have completed their entrance criteria in order for the session to proceed smoothly and productively.
2.2 The Meeting
The meeting is scheduled well in advance so as to allow all parties to adequately prepare and meet their entrance criteria. The author is present as the author has a vested interest in ensuring that the review finds all possible defects and that the product meets exit criteria without ambiguity. The reviewers will take turns going through the design from start to finish. One of the reviewers will be designated the scribe, who records all issues and their conclusions. A facilitator will be present to ensure review guidelines are kept.
The meeting takes just two hours! If the design cannot be reviewed in two hours, then it must be broken up so that it can. At the very least, a long break should take place after two hours.
Remember that the reviewing takes place, for the most part, during the pre-meeting phase. The meeting should come as close to a listing of problem areas as can be maintained.
2.3
The author will revise the design based upon the documented feedback and will conduct another review if there are any Major Technical(MT) defects.
The review process is an important step not to be taken lightly. For the design is what begins to strictly define the future of the product. All discovered issues are taken outside the review arena to save time by consulting only with the person(s) necessary to solve them. Quite frequently, problems arise that are related to each other. These can then be grouped and thus handled more appropriately.
Once the review process is complete, the design is ready for coding.
These ratings are to be used by the scribe when noting issues as they arise during the meeting:
MC Minor Change, a small fix to the material is required.
TI Technical Issue, the author and another must resolve an issue (other reviewers may be involved in the resolution) before the material can be approved, but another formal review is not required.
MT Major Technical, the material cannot be approved without another formal review.
If the problem found is a Minor Change(MC), discussion of the solution may commence. However, if there exists a Major Technical(MT) problem, or a Technical Issue(TI) that a reviewer wants to take up with the author(s), or the discussion of the MC has revealed that a more developed solution is necessary, then solution proposals should be postponed to a later date.
In a high-level design, a problem that is MT can significantly affect the productivity of a design review. In such a case, it may be better to halt the review and begin discussion of a solution to this problem.
When the inspection is done, the author will report the number of defects found to the project quality coordinator. The project quality coordinator will keep a running total of all defects found during the different stages in the life cycle of the product. Please note that these numbers will be used solely for quality metric purposes and not for personnel performance evaluation.
3.0 Configuration of the Review
Author Reviewer Facilitator
Pre-Review:
Design
is ready for
review
Review:
After Review :
Yes
No
Report
that design
is ready for
coding.
The author’s presence at the review is necessary to answer any question or clear up any ambiguities that reviewers may have. Also, the author is interested in making sure that all defects are found and all design issues are covered. It would be helpful for the all designers to be familiar with The Role of the Reviewer which is also defined in this document. After all, authors are their own personal reviewers.
Input to the design review is either a High-Level Software Design Specification, a Detailed Software Design Specification, or a combination of the two.
4.1 Entrance Criteria
Design review readiness:
Review preparation:
If at any point it is discovered that the design contains a defect that would make a review unproductive, or if an inadequate number of reviewers will be present at the review, or if an inadequate number of reviewers have done their preparation for the meeting, then the session should be postponed until such issues are resolved.
4.2 The Meeting
The author will choose one of the reviewers to be the scribe. The scribe will note each issue raised, its severity, and its recommended action. (See 2.4 Defect Ratings and Appendix A: Design Review Worksheet.)
The discussion will begin with an overview of the design and a general question period presented by the author. Then the reviewers take turns going through the design. Design walk through will proceed section by section. After each section of design, there should be a pause to field or express any issues. If a reviewer explains a section that is inconsistent with the author’s intention, then the author should speak up. Perhaps, a note should be made to further clarify that section of the design. The goal here is to make the design universally understandable. That is, a qualified engineer should be able to read the design and understand it enough to be able to code from it.
The design review is an engineering exercise requiring tremendous professional discipline from all involved. The author will not defend the material, for the reviewers will not be attacking it. Any issues raised should be deemed as constructive and honest criticisms. Any comments that are disruptive or hostile should be interrupted promptly. Remember that everyone is working towards a common goal - to produce a high quality product.
At the end of the meeting the scribe will read all the issues aloud.
4.3 Exit Criteria
Meeting is done:
If an issue exists that is MT, another review should be scheduled once this issue is resolved. If an issue exists that is TI, the author(s) should meet with the reviewer before final approval of the material, but another formal review is not necessary. Issues of rate MC do not need another review. (See 2.4 Defect Ratings.)
It should be noted that though this meeting is done, the designing operation may still persist. The author(s) must follow up on these criteria before the design can proceed into the next phase:
Designing operation done:
The next phase is coding.
It would be helpful for the reviewer to be familiar with The Role of the Author which is also defined in this document. This way, reviewers will be aware of the current state of the design and how the meeting will proceed.
These are the items that the reviewer should be considering while going through the material prior to the meeting:
Comments and discussion points should be written directly on the reviewer’s copy of the material. Perhaps, a rating of the issue would be helpful. (See section 2.4 Defect Ratings.)
If significant issues or defects are found in the design, contact the author immediately. If the reviewer is not able to prepare in time for the review, contact the author. In either case, the author should consider postponing the review meeting.
The author will choose one of the reviewers to be the scribe. The scribe will note each issue raised, its severity, and its recommended action. (See Appendix A: Design Review Worksheet.) The reviewers take turns reading the design aloud. Design walk through will proceed section by section. The reviewers should explain their section as they understand it. If the explanation doesn’t fit with what the author had in mind, then this may be a point that should be noted for further clarification by the author. The point of all this is to make the design understandable by any qualified engineer. At the end of the meeting, the scribe will read aloud all the defects found.
Reviewer’s should resist the temptation to point out corrections in spelling, punctuation, grammar, or expression. These minor corrections should be communicated to the author outside the inspection meeting by submitting a hard copy or electronic copy with these corrections noted. However, if one of these minor problems cause information to be misinterpreted and could affect the final product, those defects should be discussed at the inspection meeting.
All issues previously noted during the pre-review process should be raised at the appropriate times. All comments and defects discussed will be constructive critiques. Expressions of personal criticisms should be interrupted promptly. We must remember that we are all working towards a common goal - to produce a high quality product.
Also keep in mind that the review session should not be a forum for finding solutions to problems. Rather, the session should identify problems or potential problems to which solutions should be discussed at a later date with the appropriate people.
In addition, reviewers should be available to discuss any important issues personally identified at the review session.
6.0 The Role of the Facilitator
It is often helpful for a third party member, usually quality assurance, to be present at reviews to act as facilitator. The facilitator is one who is familiar with the design review process but need not be familiar with the design.
The facilitator need only be familiar with the design review process.
During the meeting, the facilitator will strive to keep the group on track by keeping to the following guidelines:
The facilitator should make sure that all design issues have been covered and all issues raised have been recorded with the proper people identified to handle any problems.
Conclusion of the design review will state the design is accepted as is, or accepted when corrections are met, or need another design review.
[1] Best Area Practice: Requirements Reviews
Catherine D’Abadie, et. al.
Sponsored by Area 51/NDM
[2] Productivity Improvement Through Defect-Free Software Development
- Implementing Fagan Defect-Free Process
Michael Fagan Associates
[3] Software Engineering, A Practitioner’s Approach
Roger S. Pressman, Ph.D.
McGraw-Hill, Inc.
Appendix A: Design Review Worksheet
Design Review Worksheet
Date 2/1/97
Page 1 of 1
Author(s) |
Design Name |
Date of Design |
Section Numbers |
Joni Schettino |
Client GUI |
12/11/95 |
all |
Participant Name |
Role |
Participant Name |
Role |
|
John Doe |
Reviewer |
Hillary Clinton |
Scribe/Reviewer |
|
Marilyn M. |
Reviewer |
Jean L. Picard |
Facilitator |
|
E. Presley |
Reviewer |
Jane Smith |
Author |
|
Army Archer |
Architect |
Pacific Bell |
Customer |
|
Issues found: Total number of issues found 6
Sec.# |
Rating |
Problem |
Recommendation |
3.7 |
TI |
interface mismatch |
discuss w/Hillary |
3.9 |
MC |
vague statement |
rewrite |
4.3 |
MC |
comment vague |
rewrite |
5.7 |
TI |
inefficient algorithm |
discuss w/Elvis |
5.8 |
MT |
requirements violation |
rewrite (see Jean, Marilyn) |
7.2 |
MC |
unnecessary class |
reuse is possible (John) |
Design Review Worksheet Date Page of
Author(s) |
Design Name |
Date of Design |
Section Numbers |
Participant Name |
Role |
Participant Name |
Role |
|
Issues found: Total number of issues found
Sec.# |
Rating |
Problem |
Recommendation |
Design Review Worksheet (continued)
Date
Page ____ of
Design Author(s):
Issues found:
Sec.# |
Rating |
Problem |
Recommendation |
Conclusion |
|||
Acceptence condition (select one) Accepted as is Accepted when all issues resolved Required another design review |