Upload
smartbear
View
218
Download
0
Embed Size (px)
Citation preview
8/6/2019 Debunking 3 Common CMMI Peer Review Myths
1/12
A SmartBear White Paper
This white paper provides straight talk and advice on what the CMMI really requires with respect to peer
reviews, how to get the most value rom your peer reviews, and what to look or in peer review sotware to
improve the perormance o your organization while achieving CMMI compliance. You will learn techniques to
eliminate wasted eort while tuning your teams perormance or optimal competitive advantage.
Debunking 3 Common CMMI Peer
Review Myths
Contents
Executive Summary ......................................................................1CMMI & Peer Review Myths ......................................................2
Myth 1................................................................................................2
Case Study 1....................................................................................4
Myth 2................................................................................................5
Case Study 2....................................................................................6
Myth 3................................................................................................7
Case Study 3....................................................................................8
Summary...........................................................................................9
CMMI Peer Review Sotware and Best Practice
Checklist ......................................................................................... 10
Author Bio...................................................................................... 11
About SmartBear........................................................................ 12
Ensuring Sotware SuccessSM
www.smartbear.com/codecollaborator
Executive Summary
Today there are a lot o misunderstandings when it comesto the Capability Maturity Model Integration
(CMMI)[1] and peer reviews. But why should you care?
First, peer reviews are an expected practice and required
to achieve CMMI compliance. Second, while peer reviews
are commonly held today, the way these reviews are oten
conducted is costing many organizations wasted time and
eort, and ailing to provide the promised higher peror-
mance payback on their investments. In some cases poor
peer review practices may result in delaying, or not achiev-
ing, CMMI compliance. Recent case study data [2] reveals a
major cause o these ailures can be traced to the act that
many organizations dont realize the options and exibility
they have to use peer reviews to gain higher organizational
perormance.
By Paul E. McMahon
http://smartbear.com/products/software-development/code-reviewhttp://smartbear.com/products/software-development/code-review8/6/2019 Debunking 3 Common CMMI Peer Review Myths
2/12
More specifcally you will learn:
Three commonly held myths associated with the CMMI and peer reviews
CMMI key acts related to peer review requirements
Common peer review best practices used by CMMI mature organizations
Three case studies explaining:
What many mistakenly believe the CMMI requires with respect to the way peer reviews must be conducted
How many organizations over-interpret the CMMI requirements with respect to data collection and product
reviews leading to wasted eort
How the right sotware peer review tool set up appropriately can result in higher quality products and more
satisfed customers
Key checklist items that can help you implement a peer review process optimized to meet the needs o yourorganization and customers while maintaining CMMI compliance
CMMI & Peer Review Myths
Historically there has been a tendency or people to read things into the CMMI model that arent really there
creating unnecessary non-value-added work. Understanding three commonly held myths related to the CMMI
and peer reviews along with the real acts o what the CMMI requires can help an organization improve its per-
ormance and achieve CMMI compliance aster and more eectively.
Myth 1: The CMMI specifes how peer reviews must be perormed including who must participate in reviews and
when peer reviews must take place
There is a common myth that peer reviews must be run in a ormal way, at a specifed time, and participants
must be sotware practitioner peers inside your own organization. Organizations that succumb to this myth
expend unnecessary eort, and oten ail to gain the greatest potential benefts rom their peer reviews early
critical deect removal and products validated early by satisfed customers.
2
CMMI Stages and associated Peer Review Activities
http://www.smartbear.com/8/6/2019 Debunking 3 Common CMMI Peer Review Myths
3/12
What the CMMI really requires
One o the goals o the CMMI Verifcation Process Area is to perorm peer reviews. However, the CMMI doesnt
dictate how peer reviews must be conducted, nor does it dictate the degree o ormality. There is also noth-
ing in the CMMI that specifes who the reviewers must be, or when the reviews must take place. A common
best practice in many CMMI mature organizations is to provide guidelines on how-to options to conduct peer
reviews, including recommended stakeholders to involve in the dierent types o peer reviews that make sense
or your organization.
Summary CMMI Benefts [3]
33% decrease in the average cost to fx a deect -Boeing, Australia
30% increase in sotware productivity - Lockheed Martin M&DS
15% improvement in internal on-time delivery - Bosch Gasoline Systems
20 +/- 5 deects per KLOC - Northrop Grumman
$2 million in savings rom early detection and removal o deects - Sanchez Computer Associates, Inc.
Increased ocus on quality by developers - Northrop Grumman
Peer Review Best Practices o Successul CMMI Companies
Customize their peer review process based on company and customer needs
Careully document alternative review processes & options
Provide guidelines on stakeholders to involve, including outsiders such as customers
Detail the types o peer reviews specifc to each business
3
http://www.smartbear.com/8/6/2019 Debunking 3 Common CMMI Peer Review Myths
4/124
Case Study 1 - I thought the CMMI required that...?
When your goal is to achieve CMMI compliance, start by talking to the practitioners in your organi-
zation. As they tell you how they do their job eventually they will get to the products they produce.
Then ask: Does anyone else look at these products? Anticipate a common response heard today in
many organizations just starting a new CMMI eort: We dont do ormal peer reviews.
A CMMI lead appraiser who I worked with many years ago taught me to--in these cases-- always ask
the intent question. From a tip in the CMMI guidelines we learn the intent o peer reviews is to iden-
tiy deects or removal and to recommend changes that are needed. This leads to a new question
that should be asked: How do you identiy deects or removal?
When I asked this question at BOND, one o my client organizations seeking to become CMMI Level
3, I heard: We demonstrate our products to the customer early and oten, and they give us eedback.
And We meet daily with our team and discuss our products, and We check our code into a library
daily and encourage our teammates to review and give us eedbackand they do.
What we realized was when those practitioners at BOND said, We dont do ormal peer reviews,they meant there wasnt a set time everyone went into a room to provide a list o deects ound in
their work products. But we also realized they were achieving the intent o peer reviews through a
dierent how-to option.
In todays collaborative environments customers oten operate as teammates providing eedback
early through demonstrations as we saw at BOND. Customer eedback can be a great way to fnd
deects early and help validate that your products will meet your customers needs. BOND had a his-
tory o producing high quality low deect products. When BOND explained this to their CMMI LeadAppraiser she had no difculty accepting this alternative peer review process that was achieving
the goal-- or intent-- o the CMMI peer review expected practice. We documented the process and
trained it just like BOND was carrying it out, and it helped the BOND organization achieve their CMMI
Level 3 goal.
http://www.smartbear.com/8/6/2019 Debunking 3 Common CMMI Peer Review Myths
5/125
In the next myth and case study we examine more closely how-to options, and give the reader some guidance
in deciding which ones are right or their organization.
Myth 2: The CMMI specifes what data must be collected in peer reviews and what products must be
peer reviewed
There is a common myth that every deect reported in peer reviews must include inormation related to thetype and cause o the deect. There is also a common myth that certain types o products must always be peer
reviewed on every project. Organizations that succumb to this myth oten fnd themselves expending signifcant
cost and eort without the accompanying perormance payback.
What the CMMI really requires
The CMMI expects you to analyze data rom your peer reviews, but
it doesnt tell you what data you need to collect and analyze, nor
does it tell you what products you need to review.
So how does an organization determine what data should beanalyzed, and what products should be reviewed? To answer these
questions, frst ask yoursel: What are our measurement objec-
tives? Establishing your measurement objectives is an expected practice in the CMMI model. Then ask yoursel:
What data should we collect rom our peer reviews that supports our objectives?
With respect to which products to peer review, ask yoursel, What products does our customer require us to
produce? Then ask, What products do we know we need to produce to ensure we have satisfed customers?
The right data to be analyzed rom peer reviews, and the right products to be peer reviewed or an organization
depends on that organizations specifc business domain, their specifc measurement objectives, and their cus-
tomers needs. A best practice in many CMMI mature organizations is to provide documented criteria that helpproject owners make these decisions given their specifc project situation.
Peer Review Best Practices o Successul CMMI Companies
Establish organizational measurement objectives based on your business needs
Collect and analyze data based on your companys established objectives-- not the CMMI
Provide criteria based on your company and customer needs to help your people make better decisions on
what products to peer review
Simpliy the deect reporting process as much as possible to maximize participation
The CMMI leaves the decision up to
each organization to decide what work
products to peer review... create criteria to
help a project leader make more dynamic
real time decisions with respect to peer
reviews [4]
http://www.smartbear.com/8/6/2019 Debunking 3 Common CMMI Peer Review Myths
6/126
Case Study 2 - Dont we need that data for the CMMI appraisal?
LACM is a CMMI level 3 organization that undertook a process improvement eort to streamline its
processes. The peer review process at LACM required a great deal o data to be collected about each
deect identifed and it required certain documents to be produced and reviewed on every project.The process also required periodic analysis o the collected data. When we looked close we ound that
practitioners were entering all the data about each deect because the company peer review sotware
tool required it, but we also ound no one was going back and analyzing this data.
We also discovered that some o the documents that were being produced and peer reviewed on
every project the customer didnt require and no one in the LACM organization was using them. On
urther investigation we ound the reason the data and those documents were required was because
someone inside LACM wrongly thought the CMMI required them. Practitioners also told us that
because the peer review sotware tool was so difcult to use it actually had discouraged practitioners
rom entering real deects ound.
Ater streamlining the process, peer review sotware, and related training the company ound the
CMMI implementation was strengthened because the people were now ollowing the peer review
process with greater discipline. This improvement in process adoption occurred because the peer
review sotware and process were easier to use and the practitioners in the organization
could now see the beneft o their eort.[5]
http://www.smartbear.com/8/6/2019 Debunking 3 Common CMMI Peer Review Myths
7/127
In the next section we examine how sotware tools can simpliy peer reviews and help with CMMI appraisals.
Myth 3: The CMMI doesnt speciy that sotware should be used or peer reviews, so a sotware tool
cant be helpul
It is true that the CMMI doesnt speciy that sotware should be used or peer reviews. But organizations that
succumb to the myth that peer review sotware cant be helpul oten end up making the eort o producing aquality product more difcult or their practitioners and fnd themselves trailing the competition.
What the CMMI really requires
While the CMMI doesnt speciy anything about tools, the model does expect each organization to analyze
data rom peer reviews. A common best practice ound in many CMMI mature organizations is to establish a
standard approach to the data that is collected and analyzed. This can help project practitioners in carrying out
their responsibilities and help organizations determine where they need improvement to stay ahead o their
competition.
Peer Review Best Practices o Successul CMMI Companies
Create a standard approach to data collection and analysis using peer review sotware
Customize peer review sotware to the process not the other way around
Increase adoption by using sotware that makes the actual peer review easy and unobtrusive
Assign and train personnel responsible or data collection and analysis
Centralize reporting across projects to identiy continuous improvement opportunities and simpliy
CMMI appraisals
Peer review software being used for review of software code.
http://www.smartbear.com/8/6/2019 Debunking 3 Common CMMI Peer Review Myths
8/128
Case Study 3 - How can peer review software help?
When we talked to sotware practitioners at one o our clients about peer reviews we heard:
Sure, we do peer reviews. But we werent getting a warm-eeling this was true because we
could fnd no objective evidence to back it up and their customers had been complaining aboutlatent deects. At the time they used no peer review sotware, nor required any related peer
review documentation.
We put a simple peer review tool in place requiring minimal data collection, and periodic
analysis. The tool helped us maintain records that were utilized as evidence during the CMMI
ormal appraisal. During the CMMI appraisal our interviews which included multiple sotware
practitioners revealed that beore the tool was installed and used the practitioners oten
cut corners and skipped peer reviews when pressed or time. They told us this was easy or
them to do since no records were kept.
They also told us when the peer review sotware started to be required they knew there would
be a record, and it helped discipline them to ollow the process even when they were pressed
or time. Once people started using the tool regularly, they realized that their products had
ewer latent deects, and customer satisaction increased noticeably. Using peer review sot-
ware helped this company more rapidly prepare or the CMMI appraisal process and ultimately
achieve Level 3 compliance.
http://www.smartbear.com/8/6/2019 Debunking 3 Common CMMI Peer Review Myths
9/129
Summary
The CMMI doesnt dictate how-to conduct peer reviews, nor what products you need to peer review, nor what
data you need to collect, or who needs to participate in these reviews. However, it does expect each organization
to make all o these decisions based on its own business needs, document its decisions, and train its people in
their established peer review process.
The CMMI doesnt dictate the use o peer review sotware, but the right peer review sotware set up with the
right data specifc to business needs has proven to help many organizations achieve their CMMI goals o im-
proving the quality o their products and code, increasing their sotware productivity, and saving money by fnd-
ing and fxing deects earlier in the process. This allows CMMI compliant organizations to continuously optimize
their processes, be high perorming and ultimately keeping them ahead o the competition.
Author Bio
Paul E. McMahon, Principal, PEM Systems (www.pemsystems.com), helps large and small organizations as they
improve their technical and management processes and move toward increased agility and process maturity.
He has taught Sotware Engineering at Binghamton University, State University o New York; conducted work-
shops on Engineering Process and Management, and published more than 35 articles on sotware and systems
development and management. His experience, insights, and recommendations he shares with his clients reect
24 years o engineering and management experience working or such companies as Link Simulation and Lock-
heed Martin, and 14 years o independent consulting. Paul is the author o two books, Integrating CMMI and
Agile Development: Case Studies and Proven Techniques or Faster Perormance Improvement (August, 2010)
and a book on collaborative development, Virtual Project Management: Sotware Solutions or Today and the
Future (CRC Press, 2000).
Endnotes[1] CMMI stands or Capability Maturity Model Integration CMMI is a process improvement reerence model
developed by the Sotware Engineering Institute (SEI) and is required or contractors to bid on many U.S.
Department o Deense contracts.
[2] McMahon, Paul E, Integrating CMMI and Agile Development: Case Studies and Proven Techniques or Faster
Perormance Improvement, Addison-Wesley, August 2010
[3] Demonstrating the Impact and Benefts o CMMI: An Update and Preliminary Results
(http://www.sei.cmu.edu/reports/03sr009.pd)
[4] McMahon, Paul E, Integrating CMMI and Agile Development: Case Studies and Proven Techniques or Faster
Perormance Improvement, Addison-Wesley, August 2010
[5] For more detailed inormation on the BOND and LACM Case Studies and Peer Reviews reer to [2].
http://www.pemsystems.com/http://www.sei.cmu.edu/reports/03sr009.pdfhttp://www.sei.cmu.edu/reports/03sr009.pdfhttp://www.pemsystems.com/http://www.smartbear.com/8/6/2019 Debunking 3 Common CMMI Peer Review Myths
10/12
CMMI Peer Review Sotware & Best Practices Checklist
Use the ollowing checklist to help determine the right peer review process, data, and sotware or your
organization.
10
Checklist Item Dont forget...!Yes No ?
People and Processes
Have your practitioners been trained
in your peer review process and sotware?The CMMI does expect you to train your
personnel.
Does your peer review process identiy
criteria or how your projects and person-nel decide what work products must
undergo a peer review?
The CMMI expects you to peer review work
products, but it does not dictate what theproducts need to be. You decide.
Does your peer review process provide
guidance as how to select thestakeholders or a peer review?
The CMMI expects you to involve stakehold-
ers in peer reviews.
Does your process identiy the minimum
data items that must be entered related
to a peer review?
The CMMI doesnt dictate what data to
collect. Each organization should make this
decision based on business needs.
Does each required peer review data itemsupport a dened measurement objec-
tive in your organization?
All required data collected in the peer reviewprocess/tool should support your dened
measurement objectives.
Has the person or role who is responsible
or data analysis been identied andassigned?
The CMMI does expect you to assign
resources to analyze data.
Peer Review Software
Is your peer review sotware customiz-able to align your data requirements with
your organization/s peer review process?
Sotware tools should be fexible tosupporting an organizations process.
Is your sotware fexible enough to
support peer review comments thatcome rom varied sources, such as,
collaborating teammates & customers?
Peer reviews can be implemented in
dierent ways. The CMMI doesnt dictatethe how to.
Does your peer review sotware & processprovide helpul hints to aid your people
making better peer review decisions?
Practitioners need reminders to help makethe best decision with respect to peer
reviewing products.
Does your peer review sotware & process
help your practitioners collaborate andcommunicate deects in a non-intrusive
way?
The primary intent o peer reviews is to iden-
tiy deects or removal and to recommendchanges that are needed.
Does your peer review sotware & processsupport peer review o appropriate
artiacts such as images, requirements
documents, schematics, and code.
Peer reviews can be applied to multipletypes o artiacts. The CMMI does not
dictate which artiacts need to be reviewed.
Is the peer review data centralized andeasily available or analysis and
reporting?
The CMMI does expect you to analyze peerreview data.
http://www.smartbear.com/8/6/2019 Debunking 3 Common CMMI Peer Review Myths
11/1211
About CodeCollaborator
SmartBears peer review sotware, CodeCollaborator, helps companies successully use peer reviews to reach their
CMMI Level 3 and above goals. CodeCollaborator is an enterprise-level sotware solution that helps standardize
and expedite peer reviews with team members no matter where they are located, capturing the review metrics
required or CMMI audits, and helping to improve overall product quality.
Reach your teams CMMI goals more quickly using CodeCollaborator to:
Quickly implement the standardized and documented peer review processes required or CMMI
Accelerate and simpliy the CMMI audit review process with centralized reporting
Enable continuous process improvements to get the most rom CMMI and peer review
To learn more about how CodeCollaborator can support your teams peer review needs and CMMI goals visit
http://www.smartbear.com and sign up or a ree trial o CodeCollaborator
Try CodeCollaborator Free or 30 Days
Scan and download your ree trial to see why
users choose SmartBear or peer code review.
http://smartbear.com/products/software-development/code-reviewhttp://www.smartbear.com/http://smartbear.com/products/software-development/code-review/free-codecollaborator-trialhttp://www.smartbear.com/http://smartbear.com/products/software-development/code-review/free-codecollaborator-trialhttp://www.smartbear.com/http://smartbear.com/products/software-development/code-review8/6/2019 Debunking 3 Common CMMI Peer Review Myths
12/12
About SmartBear
SmartBear Sotware provides tools or over one
million sotware proessionals to build, test, and
monitor some o the best sotware applications and
websites anywhere on the desktop, mobile and in
the cloud. Our users can be ound worldwide, in small
businesses, Fortune 100 companies, and government
agencies. Learn more about the SmartBear QualityAnywhere Platorm, our award-winning tools, or join
our active user community at www.smartbear.com,
on Facebook, or ollow us on Twitter @smartbear.
SmartBear Sotware, Inc. 100 Cummings Center, Suite 234N Beverly, MA 01915+1 978.236.7900 www.smartbear.com 2012 by SmartBear Sotware Inc.Specifcations subject to change. CC-WP-3CMMI-042012-WE
http://www.smartbear.com/http://www.facebook.com/smartbearhttps://twitter.com/#!/smartbearhttps://twitter.com/#!/smartbearhttp://www.facebook.com/smartbearhttp://www.smartbear.com/