Array operator overloading

ArrayAccess interface is an excellent offer from PHP. Though I do not personally like the idea of operator overloading, this interface can be a great tool for developing application faster and neatly by overloading the array access.

The interface is to be implemented by a class and override the accessors methods.

I am giving a simple code that demonstrate its usage.

 class MyArray implements ArrayAccess {       
    private $arr = array();

    function offsetExists($name) {              
        return isset($name);
    }

    function offsetGet($name) {       
        return $arr[$name];
    }

    function offsetSet($key, $val) {       
        $arr[$key]=$val;
    }

    function offsetUnset($name) {       
        unset($arr[$name]);
    }
}
$arr = new MyArray();
$userMap["Hello"]="Worlds";
echo $userMap["Hello"];

This class can use a database for its model. The interface facilitates coding array style with similar __get() __set() functionalities for classes.

Btw, anybody knows how I can show codes like code in an editor with line number ?

Fedex localization with ASP.net-C#

Eventually I have got some free time to relax and to learn. Fortunately this time also turned to a time to earn some

After a long time I worked as free lancer for a week. My brother from US gave me subcontract. The project was mainly to localize Fedex international website as well as build a system for future purpose.

I enjoyed working with ASP.net after about 4 years. Being a java programmer that time I used to code rather use the studio to develop custom component and coding. Though some the localization could be done partially using the VS IDE, many part was was complex and need to modify C#,VB.net and JavaScript code. So, I preferred only coding without visual aid.

In addition to the enjoyment and freedom, end of the week I got quite a good pay

Now, I am really thinking should I continue with free lancing?

PHP with NetBeans

I am always a fan of Java. And like NetBeans as well. Though it used to be slow, but the latest release netbeans 6 is super fast, it may be due to the latest jre. Using the the latest jre, old java swing applications are now super fast!!! Believe me, better else try yourself. I am sure swing will come well into the desktop market in near future (it till then the desktop idea survives )

Anyway, I was searching for an IDE for PHP with auto completion feature. Having downloaded the netbeans6 I was surprise to find a plug-in named PHP! Downloaded it and got exactly what I was looking for

For more information visit Php plug-in for netbeans.

Agile mashup in practice

“Traditional methodologies are a bunch of stick-in-the-muds who’d rather produce flawless documentation than a working system that meets business needs, Lightweight, er, ‘agile’ methodologists are a bunch of glorified hackers who are doing to be in for a heck of a surprise when they try to scale up their toys into enterprise-wide-software.”

– Jim Highsmith, in “Great methodologies Debate: Part 2″, Cutter IT journal

That says, none of the methodologies is perfect alone. Since traditional/formal models like Waterfall, Incremental,RAD model, Evolutionary, Spiral, Component-Based development etc are almost obsolete and impractical. They lost their feasibility in most cases for their time-consuming and expensive nature. Besides, few developers have the necessary background to apply formal methods, they often need extensive training on the model and methods. Furthermore, these models does not support a good communication with technically unsophisticated customers.

Hence come the agile methodologies. To produce the product right we encounter many implementation in this agile era, there are many paths originated from the agile development practice.

Now, which to choose ? Should you choose one and stick to that ? In my experience I found choosing a single methodologies is quite impractical, partly because none can match up with the unique project requirement, partly because there are way too many methodologies defined

Lets look at the core blocks of some major agile practices.

Extreme Programming (XP)
Planning:
Planning goes along with the creation of user stories. Stories are kept in index cards with priority value set by the customer. The xp team assigns a cost in terms of estimate time to complete. If the cost is more than three weeks, the story is broken into smaller ones. Needless to say, as agile goes, stories can be added any time!

Stories which will be included are set with customer’s approval along with the delevary date. The stories with highest value and the most risky ones are implemented first.

Project velocity:
Project velocity is calculated after each release. It is a simple calculation based on how many storeis are implemented in the release. Project velocity helps estimating the next release and understand the estimation accuracy.

Design and Spikes:
XP design provides only the implementation guidelines for a story as it is written. Specifying extra functionality is highly discouraged. Do only the things that is required for the current sotry in next release. If encountered a difficult design problem, create a spike solution i.e. an operational prototype.

One thing that you must need is the use of CRC cards. These cards also contains the release and story information. They are the only design artifacts produced by XP.

Coding:
Having done with the story and desing, now dive into development ? No, wait ! First write up some test cases and unit tests. Once the unit tests are created, focus only on what must be implemented. XP also encourages pair programing.

Testing:
As mentioned earlier,testing starts before coding. Tests should be done using a framework which will enable regression testing as the iteration goes on.

Adaptive software development(ASD)

It is based on team self-organization and collaboration. It is almost a formal incremental and iterative method which gives high value on learing practice.

Speculation:
First statge of ASD is speculation when the project goal, constraints etc are decided.

Collaboration:
Collaboration in this methodology does not mean simply being a motivated team working together. It is based on trust, the team met must trust each other to
1. Criticize without anmosity
2. Assist without resentment
3. Communicate to solve a problem

Learning:
Learning is practiced by formal technical review of the component built by the team members.

Hello world!

Welcome to WordPress.com. This is your first post. Edit or delete it and start blogging!

The role of a team lead in agile environment

I think pure waterfall model is impractical. It is as impractical as the pendulum in pure physics where the thread is of no weight and no friction exists. And that is what you can never have in the lab.

We, the software developers know there is always ‘friction’ and nothing is ‘massless’. The ever changing nature of the human mind and limitation to perceive what it needs results in modification in the requirement. That why we are agile, we don’t just fall.

Now, trust and sense of responsibility are at the base of agile methodologies among some other. Here, every member in a team is expected to be self driven and self motivated. They are given task and set free to accomplish that at their own choice. Formal managing, let alone macro managing, hardly gets its place. Then what would the team-lead do ? Should s/he dive into developing? or is there indeed any need of a team leader?

Actually, there is an unavoidable need of the team-lead-role. And s/he should never dive into developing unless it is that required. Rather, he(please add an ’s’ before ‘h’, I’m lazy.) role is to keep all his mets connected as well observe aspect of their work.

Team-Lead role is almost similar to the role of the wicket keeper in cricket; cheering up the players and catch all the balls behind. A team lead should also look for the quality and security of the code component. As well, he will also ensure development path is not deviated from the path. As we are talking about agile methodologies, change may come, but that change should be verified by the team lead.

Besides, the team lead should accumulate the knowledge learned by each of the team members. He is to somehow implement the notion of adaptive software development(ASD proposed by Jim Highsmith) i.e. focus on collaboration and learning. He is to arrange an meeting to share the knowledge of the team mets. It can be in lunch time, can be after lunch hour or whenever seems appropriate. Remember, there is no need to prepare presentation as it often time consuming and deviates the concentration from the development. Main idea is to use any wheel if developed already by a team.

One point must be concerned about at this regard is to maintain the motivation and positive team spirit. Say two techniques are developed by two of the team members for similar type of problem. Both the developers are smart and techniques are smart too. Now, to server in future which method is to be used? The team lead should carefully analyze(don’t take too much time, if not resolved drop it till a right time) the two techniques with participation of them and expectedly he should sort out the right one and has to convince both of them positively.

Always keep in concern to maintain the team spirit and motivation.

Books I have read.

Reading is my hobby and forgetting is my nature. In continuation to this, I have just started reading the book “Agile Software Development Qauality Assurance” by Ioannis Stamelos and Panagiotis Sfetsos.

Recently I was very much inspired by Hasin Hayder to read on, since that, I like to share my reading experience.

This post will contain all my read books gradually

1. Agile software development quality assurance
by Ioannis G. Stamelos, Panagiotis Sfetsos

2. Test-Driven Development By Example
by Kent Beck

3. Software engineering A practitioner’s approach
by Roger S. Pressman

4. Object-oriented PHP, concepts, techniques and code
By Peter Lavin

JQuiery

Once it used to be stated that to be a good developer one should learn a language every year. Why? The basic reason was to widen up ones view. By learning new language we can come to know about new techniques to solve a problem. Following that rule, I tasted some languages like C/C++(:P), java, Perl, PHP VB, Bash, VRML, JavaScript, C#, VC++/MFC(it should fall in C++ !), Flash Action script, SQL, Ferite etc.

Now, should I start learning python or ruby? I know they worth knowing, those are suppose to be a pleasant experience. But, what else might be done? I think knowing some api/libs can do the same! So, we learn GTK, XFC, Enlightenment, JFP, Struts, Tapestry etc.

In this way, I have started tasting JQuery. I am not saying learning, as to ‘learn’or to say “I know THIS” not just means I read few books, some articles and a handful of codes to demonstrate, rather it’s a matter of a longer period than 24 hours or 30 days.

Anyway, I am just writing about what I understand while going through the journey with JQuery( JQ ). Btw, I am studying the book, “Learing JQuery” by Jonathan Chaffer and Karl Swedberg.

What we can do with JQ:

1. Easily access to any DOM element
2. Modify the appearance on the fly [using css]
3. Change the content of the page
4. A wide range event handling to make the page more responsive to the user

Besides these using jq we can do animation on the page,

ajax

integration and simplify other javascript tasks.

JQ has an interesting factory function $() and it can refer to any object very easily, like:

1. Elements : $(‘p’) : Gets all the paragraphs
2. A particular element: $(‘#an_id’): gets the element with id ‘an_id’
3. Elements with same class: $(‘.className’)

[to be continued, feeling sleepy..]

Powered by Zoundry

When captain want to score

It is a post in reply to SQABD Yahoo group.

Really a nice statement, there is no ‘I’ in TEAM.

In addition to it, I would like to say, “…though it is consisted of
some I’s”. And everyone, really need to be concerned about all the I’s.

Why I am saying this, is for another truth that you said about team
sport. Yes, surely someone gets more credit. Say in foot ball, may be
the scorers are highlighted. But if the scorers want to play alone it
is not appreciated and can be risky; as we all know it will also
hamper the team effort.

Additionally, think about the case when the captain is always planning
how to score himself and forgets to plan how to defend his back, or
how to employ all the players for winning the match, in software
development context, meet the deadline and keep the project and make
the customer happy

Though, in some organizations and in some teams, it is experienced
that some members irrespective of his/her role often forget their
play.Sometimes they fail to understand the difference between
confidence and (superiority)complex. It is rather detrimental when it
happens to the captains.

Specially the captains who came up to the managerial level by his/her
development talents, which is undoubtedly appreciable. But
occasionally, they are found to give more concentration on
development(not design) rather on managing. As a consequence,
developers kept idle. There are even some situations where developers
were kept idle for 5+ hours for a 2 decision of the skipper(and it was
solved in 10min after that), why ? cause s/he was coding in quiet
mode! And it was not use to happen occasionally, rather it was a
everyday practice! Developers working late nights often had to wait
for hours together for a decision.

Moreover,another problem of captains who were a Ronaldo(I’m his
fan!)once is frequently found in their initial planning and designing
the system. They tends to concentrate(and confined in the box) more on
development details. I had another experience, where the manager was
very talented and designed an system architecture(it was really a
nice, I admit) and made the developers write more than 60 classes! But
the problem was s/he did not know which technology/api will be used.
When the api is got along with the apidoc, all those piles of codes
are kept aside for the dust to fall on and with a hope to use in
future. Developers again started writing code when they come to know
what to do after sitting idle for hours. And again, the captain uses
his/her full talent to develop some roadblocks (definitely those were
very important) but without an integration plan and guess what… even
without a revision control system. So the developers started writing
the drama, which was scattered and scary…..

So I think, though the word TEAM has no ‘I’ in it but it has an ‘M’ at
the lead. S/he must ensure and be concerned that there comes no ‘I’ in
TEAM.

Have a nice day !

NetBean, even more fun

I am not going to raise another holly war between netbeans and the eclipese IDE. But, even though eclipse ide worked great after little tweaking, i think everything work better after they are tweaked, I love netbeans for its simple and fast interface. I know, many of you already frowned as called netbeans fast ! Yes indedd. At least for C/C++ development. Netbeans appeared simply lightening fast. It easily replaced geany as my favorite IDE.

It is true, eclipse user base is very rich and so is its plug-in base. Almost all the necessary tools along with huge unnecessary ones (for a particular development view) can be found with eclipse. Especially the subversion plug-in… it is simply splendid). No worries.. I will download the svn module for netbeans. Hope it will not dishearten me

Lastly, I also admit, netbeans is very slow for java development. Last time I was developing a non-day-job project using jsp/struts. I choose netbeans and found that it takes all the memory of my computer. Again after tweaking for memory, it came up with much faster speed, even though an honest comment will be . SLOW.

But, if for C++ development, its simply great. Additionally, you can easily create UML design graphically by drag and drop; and then use it in your project ! One interesting feature of the UML module is that, here some of the design pattern templates are built in i.e. just select a pattern, give custom name click ok… you got the classes.

Finally, I am very happy with the new netbean.. I simply love it.

Using Eclipse is Fun !

My current project is about building an embedded software using C++ on a ARM board. Due to some reasons we had to use use Eclipse as we are to use a proprietary tool chain plug-in of the editor. Anyhow, eclipse was too slow to proceed development work fast where deadline is always yelling at the ear and release are very frequent. I started use geany, my favorite editor, for writing codes and just used Eclipse to build the project.

Tonight, I got some relax time(you know 36 degree centigrade is cooler than 40 degree centigrade ! ) and played with eclipse settings. I just stopped the indexer and some other auto completion options — the result was out of expectation ! Eclipse became fast enough to code on Besides, for source code I like courier font, so I did not miss to change that too.

Now, if you are interested, go to Window->Preferences->C/C++ and set off the content assistant and indexer. Bingo… now you have a faster Eclipse.

C++ Member initialization

Many of us may not know that member variables of a class object is initialized before the body of the constructor is executed! The class type members are initialized with the default constructor i.e. constructor with no arguments. Primitive type member variable like int, float etc does not have any constructors, so they are just initialized with an undefined initial value which is often a garbage.

Anyhow, this behavior can be changed ! While defining the constructor to initialize the member variables with desired values put a colon after the constructor’s last first-bracket and initialize the member variable placing a pair of first bracket after each member name; don’t forget to put the value in between the first brackets. Both the class and primitive type members can be initialized in this way. Here is an example:

class A{
string m_name;
int m_id;

public:
A(int id);
A(int id, string& name);
};
/////
A::A(int id):m_id(id),m_name(“No name”)
{}

This is a faster way than initializing the member variables in the constructor with assignment operator. Because for the later case member variables are first initialized with their respective default garbage values than assigned with the desired values; whereas in later case they are simply initialized with the desired values !

Play game here

Just added 2 free games from DynamicDrive. I love this site for its great resources. Besides, I added fading effect to pop up the game div at center. You can now play these games from this page, just click any of the game link from the top left menu !!

Would you please play once and put your scores on comment ?

Thanks for your encouragement.

Macadamian’s Code Review Checklist

At Macadamian, we practice what we preach with peer code reviews. Before we commit any code to source control, we check it for compliance with this list.

We’ve made the checklist public for the use of software development teams implementing code review as part of their process. For more information about code review processes and software development best practices, read check out the Critical Path newsletter – it’s free, too!

General Code Smoke Test
Comments and Coding Conventions
Error Handling
Resource Leaks
Control Structures
Performance
Functions
Bug Fixes
Math

General Code Smoke Test

Does the code build correctly?
No errors should occur when building the source code. No warnings should be introduced by changes made to the code.

Does the code execute as expected?
When executed, the code does what it is supposed to.

Do you understand the code you are reviewing?
As a reviewer, you should understand the code. If you don’t, the review may not be complete, or the code may not be well commented.

Has the developer tested the code?
Insure the developer has unit tested the code before sending it for review. All the limit cases should have been tested.

Comments and Coding Conventions

Does the code respect the project coding conventions?
Check that the coding conventions have been followed. Variable naming, indentation, and bracket style should be used.

Does the source file start with an appropriate header and copyright information?
Each source file should start with an appropriate header and copyright information. All source files should have a comment block describing the functionality provided by the file.

Are variable declarations properly commented?
Comments are required for aspects of variables that the name doesn’t describe. Each global variable should indicate its purpose and why it needs to be global.

Are units of numeric data clearly stated?
Comment the units of numeric data. For example, if a number represents length, indicate if it is in feet or meters.

Are all functions, methods and classes documented?
Describe each routine, method, and class in one or two sentences at the top of its definition. If you can’t describe it in a short sentence or two, you may need to reassess its purpose. It might be a sign that the design needs to be improved.

Are function parameters used for input or output clearly identified as such?
Make it clear which parameters are used for input and output.

Are complex algorithms and code optimizations adequately commented?
Complex areas, algorithms, and code optimizations should be sufficiently commented, so other developers can understand the code and walk through it.

Does code that has been commented out have an explanation?
There should be an explanation for any code that is commented out. “Dead Code” should be removed. If it is a temporary hack, it should be identified as such.

Are comments used to identify missing functionality or unresolved issues in the code?
A comment is required for all code not completely implemented. The comment should describe what’s left to do or is missing. You should also use a distinctive marker that you can search for later (For example: “TODO:francis”).

Error Handling

Are assertions used everywhere data is expected to have a valid value or range?
Assertions make it easier to identify potential problems. For example, test if pointers or references are valid.

Are errors properly handled each time a function returns?
An error should be detected and handled if it affects the execution of the rest of a routine. For example, if a resource allocation fails, this affects the rest of the routine if it uses that resource. This should be detected and proper action taken. In some cases, the “proper action” may simply be to log the error.

Are resources and memory released in all error paths?
Make sure all resources and memory allocated are released in the error paths.

Are all thrown exceptions handled properly?
If the source code uses a routine that throws an exception, there should be a function in the call stack that catches it and handles it properly.

Is the function caller notified when an error is detected?
Consider notifying your caller when an error is detected. If the error might affect your caller, the caller should be notified. For example, the “Open” methods of a file class should return error conditions. Even if the class stays in a valid state and other calls to the class will be handled properly, the caller might be interested in doing some error handling of its own.

Has error handling code been tested?
Don’t forget that error handling code that can be defective. It is important to write test cases that exercise it.

Resource Leaks

Is allocated memory (non-garbage collected) freed?
All allocated memory needs to be freed when no longer needed. Make sure memory is released in all code paths, especially in error code paths.

Are all objects (Database connections, Sockets, Files, etc.) freed even when an error occurs?
File, Sockets, Database connections, etc. (basically all objects where a creation and a deletion method exist) should be freed even when an error occurs. For example, whenever you use “new” in C++, there should be a delete somewhere that disposes of the object. Resources that are opened must be closed. For example, when opening a file in most development environments, you need to call a method to close the file when you’re done.

Is the same object released more than once?
Make sure there’s no code path where the same object is released more than once. Check error code paths.

Does the code accurately keep track of reference counting?
Frequently a reference counter is used to keep the reference count on objects (For example, COM objects). The object uses the reference counter to determine when to destroy itself. In most cases, the developer uses methods to increment or decrement the reference count. Make sure the reference count reflects the number of times an object is referred.

Thread Safeness

Are all global variables thread-safe?
If global variables can be accessed by more than one thread, code altering the global variable should be enclosed using a synchronization mechanism such as a mutex. Code accessing the variable should be enclosed with the same mechanism.

Are objects accessed by multiple threads thread-safe?
If some objects can be accessed by more than one thread, make sure member variables are protected by synchronization mechanisms.

Are locks released in the same order they are obtained?
It is important to release the locks in the same order they were acquired to avoid deadlock situations. Check error code paths.

Is there any possible deadlock or lock contention?
Make sure there’s no possibility for acquiring a set of locks (mutex, semaphores, etc.) in different orders. For example, if Thread A acquires Lock #1 and then Lock #2, then Thread B shouldn’t acquire Lock #2 and then Lock #1.

Control Structures

Are loop ending conditions accurate?
Check all loops to make sure they iterate the right number of times. Check the condition that ends the loop; insure it will end out doing the expected number of iterations.

Is the code free of unintended infinite loops?
Check for code paths that can cause infinite loops. Make sure end loop conditions will be met unless otherwise documented.

Performance

Do recursive functions run within a reasonable amount of stack space?
Recursive functions should run with a reasonable amount of stack space. Generally, it is better to code iterative functions.

Are whole objects duplicated when only references are needed?
This happens when objects are passed by value when only references are required. This also applies to algorithms that copy a lot of memory. Consider using algorithm that minimizes the number of object duplications, reducing the data that needs to be transferred in memory.

Does the code have an impact on size, speed, or memory use?
Can it be optimized? For instance, if you use data structures with a large number of occurrences, you might want to reduce the size of the structure.

Are you using blocking system calls when performance is involved?
Consider using a different thread for code making a function call that blocks.

Is the code doing busy waits instead of using synchronization mechanisms or timer events?
Doing busy waits takes up CPU time. It is a better practice to use synchronization mechanisms.

Was this optimization really needed?
Optimizations often make code harder to read and more likely to contain bugs. Such optimizations should be avoided unless a need has been identified. Has the code been profiled?

Functions

Are function parameters explicitly verified in the code?
This check is encouraged for functions where you don’t control the whole range of values that are sent to the function. This isn’t the case for helper functions, for instance. Each function should check its parameter for minimum and maximum possible values. Each pointer or reference should be checked to see if it is null. An error or an exception should occur if a parameter is invalid.

Are arrays explicitly checked for out-of-bound indexes?
Make sure an error message is displayed if an index is out-of-bound.

Are functions returning references to objects declared on the stack?
Don’t return references to objects declared on the stack, return references to objects created on the heap.

Are variables initialized before they are used?
Make sure there are no code paths where variables are used prior to being initialized. If an object is used by more than one thread, make sure the object is not in use by another thread when you destroy it. If an object is created by doing a function call, make sure the object was created before using it.

Does the code re-write functionality that could be achieved by using an existing API?
Don’t reinvent the wheel. New code should use existing functionality as much as possible. Don’t rewrite source code that already exists in the project. Code that is replicated in more than one function should be put in a helper function for easier maintenance.

Bug Fixes

Does a fix made to a function change the behavior of caller functions?
Sometimes code expects a function to behave incorrectly. Fixing the function can, in some cases, break the caller. If this happens, either fix the code that depends on the function, or add a comment explaining why the code can’t be changed.

Does the bug fix correct all the occurrences of the bug?
If the code you’re reviewing is fixing a bug, make sure it fixes all the occurrences of the bug.

Math

Is the code doing signed/unsigned conversions?
Check all signed to unsigned conversions: Can sign completion cause problems? Check all unsigned to signed conversions: Can overflow occur? Test with Minimum and Maximum possible values.

Interesting Code Review "Facts"

The benefit of code reviews is undeniable.

Yet, code reviews don’t happen in most software development organizations. And for good reason. The way code reviews are typically handled uses an enormous amount of people and time resources. Even with system quality improvements developers may not feel the effort is worth the cost.

It turns out you don’t need a high overhead code review process to get results. At one company I worked at I created a very low overhead automated code review system that enabled every line of code to be reviewed before it was checked-in.

Here are some ideas of how code reviews can be made practical in your organization.

1. Interesting Code Review “Facts”
2. Code Review Standard
3. Perspective Based Reviews
4. Deciding If Code Reviews Work
5. Are Code Reviews Necessary With Pair Programming?
6. Code Review Resources

---

My initial impression of code reviews is that they were heavy and bloated and a waste of time. I was wrong on all counts. Here are some interesting facts I found in the research:

1. Using two inspectors is as effective as using four or more.
2. Using an asyncronous process is as effective as a meeting based process for code reviews. For design reviews a meeting based approach may be more effective.
3. One review session is as effective as multiple inspection sessions.
4. Scenario or perspective based reviews were more effective than add-hoc and check-list based reviews.
5. Inspections are effective.
6. Inspecting upstream process like requirements and designs is very effective.

The implication is that the complexity and overhead of Fagan like code inspections are not needed and a semi-automated method will yield improved system quality. Requirements and HLDs seem to benefit from a more structured inspection process, which makes a certain amount of sense.

---

Code Review Standard

This is the list of some policies I would use to create a code review standard for my organization. Take what’s here, read the review on code research, and figure out what will work best for your organization.

The key point to keep in mind is that the heavy weight nature of code reviews is not true. You can create a code review system for your group that is light enough that you can review every change in your entire system.

Every Line of Code Must Be Reviewed Before it is Checked-In

Do not allow code to be checked-in unless it has been reviewed, fixed, and rereviewed. If you can’t do this then your review process isn’t fast or light enough. It can be done.

Reviewing code after it has checked-in is next to useless as everyone is exposed to the unreviewed code.

Code Must Be Integrated, Compiled, Unit Tested, and System Tested Before Review

Spending time on code that is just going to change or doesn’t work is a massive waste of time. The entrance requirements for a code review are:

• Code should compile without errors or warnings (according to coding standard).
• Code should be already be integrated with its parent branch.
• Code should be fully unit tested.
• Code should already pass the system tests where possible.

Develop Perspective Based Reviews

For the issues that are important in your software consider developing a Perspective Review for each issue. Perhaps special hotspots for you are internationalization, memory corruption, memory usage, threads, semaphores, etc. An example can be found in Semaphore Perspective Code Review.

You can then use the perspectives to assign roles to review team members.

Use Meetingless Asyncronous Reviews

You don’t need to hold a big meeting for a review. Everyone can review the code when they can find time. Simpy have a done-by date when the review must be completed.

Everyone on the review must review the code. If a person can’t perform the review then the team needs to elect someone else to perform a similar role in the review.

Meetingless aysncronous reviews are fast and light, yet they are still effective. With the right tool support you can easily review every line of code in your system.

Use Between 2 and 5 Relevant Reviewers

Too many reviewers wastes everyones time. Keep the number of reviewers small and relevant to the code being reviewed.

Assign Reviewers Roles and Perspectives

It’s almost impossible to review a lot of different issues in a lot of code. People get tired and they stop noticings things.

The way around the getting tired issue is to use perspective reviews. Create a perspective for each important issue category your are concerned about. Assign the perspectives to people on the review team. Because they are only reviewing for issues in their perspective they will do a better job because they can stay more concentrated. This doesn’t mean they can’t find other issues as well, but they responsible for their perspective.

For example, if using semaphores correctly is important in your software and the code has semaphores, then assign someone the role of reviewing semaphore usage. An example can be found in Semaphore Perspective Code Review.

All Review Communication Should Go to the Review List and be Logged

Part of the benefit of a review is that people learn about the system being reviewed. This learning feature is facilitated by broadcasting email communication between the review team and saving all communication so it can be read by other people later.

Do Not Redesign in the Review

Make a note and schedule design issues for for a later time.

Developers always think they can do stuff better. Take these issues off line unless the issue is that requirements are not being met. Requirements not being met is not the same as you could have done it better.

Do Not Cover Coding Standards Violations in the Review.

Send violations via email or in person.

Talking about violations only gets everyone angry and is a waste of time.

Code Is Rereviewed Until it Passes

Code isn’t reviewed once and then forgetten. Any changes made have to be rereviewed. If you think this is too slow then your process isn’t light enough.

No reviewing all changes makes the process useless as people will just ignore suggestions or introduce new bugs in any changes.

All Issues Must be Fixed, Marked as Not an Issue, or Marked as Bug

Any issue brought up to a developer must be handled. A developer just can’t ignore issues because they think it’s stupid. Every issue must:

• Fixed.
• Marked as Not an Issue. If the developer and the reviewer can’t decide between themselves if an issue should be fixed or not, then the review team gets to decide. If there is only one reviewer then bring a manager in or another team memember.

Automate Your Code Review System

You can make your process light enough by building it into your build system. If your process isn’t light enough work on until it is.

Review All Code on Private Branch Before a Merge

Code developed on a private branch doesn’t need to reviewed during development. But before the code is merged into a parent branch all code changes must go through the complete review process. For this reason, development of a large scale feature, may still want to perform reviews on the private branch because that can speed up the merge process.

Of course, try not to have branches separate from the mainline, but for large features that take a long time to develop you will often need separate branches.

Review the Right Scope of Changes

You don’t have to review every line of code in every module that has changed. Certainly if a module is new it must be completely reviewed. Other than that you may be able to just review the changed code. Though just reviewing changed code isn’t always possible. If you are performing a semaphore perspective review, for example, then you will need to go look at the code within in the scope of the sempahore as well.

Stick to Reviewable Issues

Develop your list of what issues can be reviewed and how they are to be reviewed. Usually this is in the form of check lists and perspective reviews. Don’t allow reviews on other items without changing what can be reviewed. Otherwise people spend endless time on off-topic arguments.

Review Team Reponsible for Deciding Issues

If there is a conflict on any part of the review then the review team is responsible for handling it. That’s the only way the review process will be light enough to work.

Keep it Cool

Nobody is perfect. The attitude of the review should never be personal, it should always be professional, with the goal of improving the system and the people building the system. Keep your tempers.

Don’t blame people for bugs. Work together to make things better. No finger pointing! Not ever!

Meetingless reviews can help keep the anger down, but it can make it worse too. When people are in the same room anger can ramp up really quick. And we know in email it’s very easy to say something that can be take wrong. Raise the awareness of these issues in your team.

A good rule is to Never Assume An Attack. If you find yourself getting angry, assume it’s a misunderstanding, not an attack.

No Managers

Unless a manager has something to add to the reviews then they shouldn’t be involved. Issues should be decided by the review team. Managers always have to run to a different meeting, they don’t have time slots open for meetings, and they generally don’t add technical input. So you don’t need managers as part of the review process.

If a Bug Was Not Caught in a Review Figure Out Why

If a bug happens after a review then track down why each bug wasn’t found and then change your development process somehow to try and prevent that bug in the future.

This is not always possible as running full tests are often impossible, but it should be mostly possible.

I would create a bug for each bug to track down why it wasn’t caught. Because this issue is more serious than just the review. It means the unit test, the system test, and the review did not work.

Review Upstream Documents Too

Reviewing product requirement documents, specs, standards, etc can provide excellent return on value. Make sure those products are reviewed as well.

Feed Leasons Learned Into the Team and Documentation

If issues come up during the review that everyone in the team would benefit from, then have a way to make wisdom public.

I would recommend a development wiki where you can write documentation on anything useful that people come up with.

Plug Reviews into Your Source Code Control System

I have done this through change check-in comments. Each change has to be reviewed before it is checked-in. The submit comment must contain a review ID that points to some document containing the review status for the change that is about to be submitted. The code is prevented from being submitted without a valid passed review. If you are able to automate your code review system all this works quite quickly and painlessly.

---

Perspective Based Reviews

Before my reseach into code reviews I had never before heard about perspective based reviews . Perspective based reviews are done from the point-of-view of various project stakeholders. This is as apposed to everyone doing an add hoc or checklist based review. A perspective would be a user, requirements person, hardware interface, exception handling, mutex usage, memory usage, etc. The thought is they are more effective because reviewers are concentrating on a particular issue instead of trying to review everything. Do one thing well i guess.

Reviews are driven by scenarios. A scenario tells the insepector how to go about reading the documentation from particular perspective and what to look for. Clearly there’s a lot to the scenarios.

Perhaps people could be assigned in various review roles. Or maybe some people can perform reviews only from a certain perspective.

---

Deciding If Code Reviews Work

How do you know if code reviews are working for you? Look for

• Increased in test execution rate. The tests suites runs through its test suite. Unstable software can’t do this.
• Increased pass rate.
• Reduction of critical bugs in QA.
• Reduction of field issues.
• Subjective improvements in quality.

Try to measure the issues to verify if code reviews are working for you.

---

Are Code Reviews Necessary With Pair Programming?

An issue that comes up is the relationship between code reviews and pair programming . Do you still need code reviews if you pair program?

If you don’t pair program then you definitely should perform code reviews. That one is easy. Unit tests and system tests simply aren’t good enough to replace code reviews.

After that it depends.

It depends on the product you are developing, it’s importance, it’s complexity, and how critical it is that there be no errors.

Pair programming is an excellent form of active code review, but there are many products where I don’t think two eyes are enough. Some systems are complex enough, some domains are complex enough, and some conseuqneces are severe enough that I want more eyeballs on them before the code is let loose in a real code running system.

When pair programming is used you could decide what code should be reviewed by subsystem or by the type of change involved.

For example, code in a key algorithm may need to be reviewed by the impacted parties. Perhaps a drive change needs to be reviewed. Perhaps when a piece of code is written new or completely rewritten, it should be reviewed. Perhaps if semaphores are added to code that would trigger a review because it changes the system a great deal.

Your mileage may vary of course.