Coding Standards

Presentation by Jordan Belone

Coding standards, sometimes referred to as programming styles or coding

conventions, are a very important asset to programmers. Unfortunately, they are
often overlooked by junior as well as some program developers due to the fact
that many of the recommended coding standards do not actually affect the
compilation of the code itself, a concept that we will focus on later. This report
shall try to demonstrate the importance of coding standards, and why they have
already become a common practice for businesses and programming language
developers alike.
I shall begin with a definition of coding standards. In general they are a set of
standards and guidelines which are/should be used when writing the source code
for a program. Standards are rules which a developer is expected to follow.
However these rules, such as the compiled file must have the extension .h are
often apparent are followed anyway. Guidelines however are much less well
followed. They are stylistic measures which have no direct affect on the
compilation of the guide, and exist only to make the source code more humanly
readable. They are optional, but highly recommended.
An important thing to remember about coding standards that many
programmers seem to forget is that coding standards are not an abstract idea. In
fact, for many companies and practically all programming languages, a coding
standard is a physical document, and is vital for producing uniform code.
Coding standards vary in importance between companies and programming
languages. For the COBOL programmer for example, commenting code (an area
of coding standards which will be explained later) would be significantly less
important than in other languages, as it is possible to make COBOL self
documenting. It also may not be important for a company that is not going to be
continually producing software and it may be more time consuming than just
building the software without. However, in open source software, the importance
of coding standards cannot be emphasised enough. The most prominent
example of this is Mozilla Firefox, the open source internet browser which makes
all its code freely available online. This is so other developers can look over the
source code, possibly identifying possible security flaws/bugs in the code as well
as extending the code. It would be extremely difficult for other developers to
identify potential flaws in the program if it did not follow a particular style
convention. Also, Firefox has a huge database of extensions, some of which it
created itself but many of which were created by external developers. External
extensions would be very difficult to understand and integrate into Firefox if they
did not follow a coding standard, and Firefox may not have been such a popular
internet browser as it is today.

Coding standards, when implemented correctly, can provide huge advantages to

the produced code and the companies implementing them. They can make a
great difference when maintaining software. Maintenance is arguably the most
important step in the Software Development Life Cycle (SDLC) and is definitely
the most recurring. When the developers who are responsible for maintaining the
code are different from the ones who produced it, it relies on them understanding
code developed by another developer. With the lack of a coding standard, this
task would be extremely different because the majority of programmers have
their own programming style which is often difficult to follow by anyone with the
exception of the original developer. Having a coding standard however, means
that all software programmers will be easily able to read and understand all the
code for a program, as long as it conforms to the coding standard and the
developer who is maintaining the code is familiar with it, making it easily
maintained and modify if needed.
Software that complies with a pre-existing coding standard has also been proven
to contain fewer bugs and generally be of a higher quality. This is because the
software is much easier to test, and any potential errors or problems with the
code can be fixed in the development process of the program. Since the cost of
fixing a bug in a program increases exponentially over time it is important to
catch such bugs in the early development process to save money. This is,
however, not the only way having a coding standard can save money for a
company. Having a pre-existing coding standard that the developers are already
familiar with can speed up the overall software development process, get the
product to market faster, possibly making more money for the company. Also, as
discussed earlier, software developed to a coding standard contains less bugs
which means less time and money has to be spent maintaining the code.
Complying with a coding standard can also have a huge effect on teamwork.
Often, programmers learn to program in different ways and have varying, often
conflicting programming styles. This can cause conflict within a development
team and such conflicts would certainly make the team less efficient. Having a
pre-existing coding standard however means that there are no discrepancies as
to how the program should be designed. Another way in which teamwork is
affected is the ability to team switch i.e. have other programmers working on
other projects transfer into other development groups. Having a companyencompassing coding standard means that developers can be switched from
different parts of the company and require very little training to begin
programming in other teams. This complements critical path analysis, a business
technique which identifies the areas of a project which, if delayed, would delay
the entire project. If such areas are identified, then developers can be easily
assigned to such areas, ensuring the project is finished on time.
There are two main types of coding standards, coding standards written for a
company, and coding standards written for a programming language. The
company coding standard is what developers are expected to follow within the
company, whereas the coding standard for a programming language is what the
programming language developers recommend all programmers should follow.

Often, both contain a lot of similar points, and for companies that only program
in one language, both are often almost identical. However for companies which
develop software in a variety of languages, they often formulate their own coding
standard. For example, Mozilla Firefox is written in C++, XUL, XBL, and
JavaScript. Instead of inefficiently using the four separate coding standards of
each language, Mozilla has created its own coding standard.

Examples
In each coding standard, there are many points and rules. For example, the
coding standard for GNUs Not Unix (GNU) (an open source operating system) is
88 pages long. There are many points contained within such coding standards,
some of which I shall discuss here.
E.g. 1
Indentation
Indentation is discussed in all coding standards at some level. In some
languages, indentation is used by the compiler to identify the scope of functions.
In freeform languages, such as Java, it has no actual effect on the compilation of
code itself, and is just used to make the code more human readable, giving an
indication of scope without actually affecting the program directly.
Type 1
if (g < 17 && h < 22 || i < 60) { return true; } else {System.out.println
(incorrect) ; return false; }
Type 2
if (g < 17 && h < 22 || i < 60)
{
return true;
}
else
{

System.out.println(incorrect);

return false;
}

It is important to understand that type one and type two do exactly the same
thing, and since both pieces of code are written in is Java, the indentation is
ignored the compiler. The indentation is meant for the human user. It can be
seen when examining both pieces of code above that the type 2 code is much
more easier to read than the type one code. I have used a simple if-else loop do
demonstrate how even the simplest of code can become confusing when
indentation is not used effectively. The type two code will not only be easier to
read and understand for the person who is developing the code, but will make it
much easier for the developer maintaining the code to do their job correctly and
effectively.
E.g. 2
Commenting code
Commenting code is also an area which is addressed in the majority of coding
standards. Other languages, such a Tex, are self-documenting and are able to
automatically produce documentation describing themselves. In others, such as
COBOL the code itself is self commenting so there is less of an emphasis on such
areas. However, just commenting code is not enough, as some comments can
actually be of little use at all, and can often make the code more confusing.
Comments should clearly demonstrate the function of the code, not only to other
developers but also to you. Often when writing larger programs it can be easy to
lose track of what certain functions do, and it is easier to read a well written
comment that it is to trawl through lines of code trying to remember what the
function of the code you have already written.
Comments should also not be too long. If your code has followed a coding
standard, it should not be too difficult for developers to understand your code.
Long comments are often unnecessary and make your code look messy.
Comments should not be every line. Often lines such as loops and variable
assignments are very simple to understand and do not require comments. Most
coding standards recommend commenting the end of functions and objects,
rather than commenting every line. In functional languages, it is generally
recommended that a comment explaining a function is written before the code
itself.
The comments should also not be too complicated. The point of a comment is
help the reader understand the code. It is important to keep the comments short

and simple. For programs used on a global scale, it is recommended that they
are written in English, as this is the most commonly used language among
developers today.
According to the Ada coding standard in 2005,
Programmers should include comments whenever it is difficult to understand the code
without the comments.

This is very good advice, and the huge majority of other coding standards agree
with this statement.

E.g.3

Variable declarations.
This is another area which is almost always included in coding standards.
Variable declarations are often overlooked when programming but in larger
programs they can be the difference between understanding code and not.
Variable declarations should be long and demonstrate the function of the
variable which they store. In the majority of programming languages, variable
can have any name, excluding a few keywords that are reserved by the language
itself.

e.g.

If(a < h && z <o && t<e)

{
return true;
}
else

Although this code does follow indentation

rules, it still is difficult to understand. This is
because the code had been give generic
names that do not help explain the variables
at all. I have used a basic if loop to
demonstrate that even the most basic
functions can be made confusing without
correct variable declarations. It must be
remembered, however that almost all
programs will be more complex than this and
therefore making the code even less difficult
to read. Having appropriate naming variables
can reduce the need for comments in our
code and make it more self-explanatory.

{
return false;
}

These are only a few of the many details generally included in coding standards.
Others range for the more general avoid big functions to the more specific
Limit the line length to a maximum of 120 characters.

Conclusion
It is vital to remember that coding standards are only as useful as the method
which his used to enforce them. Even if a company has a well thought out,
comprehensive coding standard, if they are not enforced, then they are useless.
Programs such as StyleCop, can be used to check your code against the coding
standard of the Microsoft development centre. Other tools, such as the Enerjy
code analyzer can not only check your code complies with the Java coding
standard but can also detect some bugs in the system. This is however not a
fully comprehensive solution and it is ultimately the responsibility of the
programmer to comply with the standard.

References
GNU Coding standard
http://www.gnu.org/prep/standards/standards.html
Mozilla Firefox Coding Standards
https://developer.mozilla.org/En/Developer_Guide / Coding_Style
Enerjy Code Analyzer
http://www.enerjy.com/index.html
Brian W.Kernighan and Plaugher, The elements of programming style second
edition.
Rob pike and Brian W Kernighan, The Practice of programming