Approaches to Web Development for Bioinformatics

On this page:

• Java
  • Language Basics
  • Tools for Java Development
  • Java Compared with Other Languages

Java

This section discusses the basics of the Java language with a bioinformatics flavor. The section Web Programming with Java discusses developing web user interfaces with Java and the section BioJava discusses the BioJava open source project. Several other sections give Java examples as well, including the sections AJAX and Accessing a Relational Database with Java.

Language Basics

The Sun Microsystems java web site¹⁶ is a great place to begin studying Java. You can freely download the latest Java Development Kit (JDK) and work your way through the Java Tutorial.

Here is a Java program that does the same thing as our Hello World Perl program.  The file is in Hello.java.

public class Hello {

    public static void main(String[] argv) {
        System.out.println("Hello World!");
    }
}

Compile it by entering the command.  This will generate a .class file in the same directory.

> javac Hello.java

Run the program by entering the command

> java Hello

The first thing you may notice comparing it to our Perl Hello World program is that it takes three lines to write instead of one, ignoring the comments in the Perl program.  All Java programs must be encapsulated in a class (Hello).  Procedural code is usually encapsulated in a method (main).  There are also concepts of class and method visibility (public).  There is the concept of static versus instance in the main method signature.  The method return type (void) must be specified even if it is not used.  Finally, you have to get an object (System) to get an output stream (out) to print the text with the platform method println().

Although it seems more complex than the Perl example, all these concepts have been developed with code reuse in mind. Let's look at an example that is similar to the regular expression example with Perl that was used to validate that all the symbols in a DNA string were valid nucleotides.  The file is Validate.java.

Compile the class with the command

Test the program with similar input to the Perl script:

The output is similar to Perl script:

There are couple of points are immediately obvious:

• The regular expression is not part of the basic syntax of the language but an argument to compile() method of the class Pattern.  It is more verbose than the Perl example.
• Evaluation of the regular expression is a two step process rather than the one step in Perl but the result is still the same.
• The comments start with /**.  This allows generation of HTML describing the class with the JavaDoc tool.  The @param symbol is used to describe arguments to methods and is copied into the generated HTML.
  

The language constructs described above allow programmers to better encapsulate their code within an application programming interface (API).  The exact interfaces that are exposed to users of the API is can be controlled to be the minimum needed.  Users of the API are constrained to use it in a particular way, according the Java types in the interface and the Java platform comes with tools to document the use API's in HTML.  This creates some programming overhead but it has been thought worth it to many large scale software development projects both for the development of products and in-house systems for all kinds of businesses.

Finally, to complete our comparison of Perl with Java, let's look at an example that translates an RNA sequence to an amino acid sequence. The Java source is in Translate.java.

The program can be compiled and run with the commands

Although the program is more verbose that the Perl equivalent there are several notable things:

• Because of the Java type system a variable can be declared as more general type.
• The Map template ensures that the correct things are put into and got out of the map. This is shown by the fragment Map<String, String> TRANSLATION = new HashMap<String, String>().

Tools for Java Development

A number of tools for Java development are included with the basic Java Development Kit (JDK).  These include the compiler javac, the documentation tool javadoc, the Java ARchiving (compression) tool jar, and performance instrumentation.  In addition, there are probably more freely available development tools than any other platform, including

• Integrated development environments, such as Eclipse and Forte (NetBeans), that include graphics debuggers, command line completion, syntax coloring, static code checking, and code review assistants, among a huge number of other features
• The Junit unit testing framework
• Aspect oriented programming tools, such as AspectJ

There are also a number of commercial tools that can be useful for web development and performance testing.

Java Compared with Other Languages

Java has been the language of choice for beginning computing sciences courses at many universities and for many companies in project and product development.  It is a very well designed language and it is a pleasure (for me, at least) to program in.  However, it has a number of disadvantages

• It is not as quick to pick up as Perl and other scripting languages.  This can be a significant barrier to people who are not software engineering professionals or who do not have a team of software engineering professionals at their disposal for programming tasks.  A large proportion of fundamental aspects of the language have been introduced to scale for large programming teams and for ease of maintenance.  However, these language features can be a barrier to more casual programmers.
• Although it often is the choice for large projects and products with web interfaces, the minimum cost and barrier of entry can be higher than other platforms, such as Perl / CGI, PHP, and Microsoft .NET.
• Java is inheretly at least a little slower than C programs and the use of a garbage collector to clean up memory does add some overhead.  A more significant problem is in the startup time.  Java Virtual Machines (JVM's) have become larger and more bloated over the years resulting in slower startup times and larger footprints.   A bigger problem, however, is the programming culture.  Java culture has often focused on style and capability at the expense of performance.  One case in point are Enterprise Java Beans (EJB's), which are notoriously complex, slow, and do not add the value promised.  All that said, Java is an efficient platform when programs are constructed with care.

The availability of open source projects in Java is also an important factor.  There are a number of open source bioinformatics projects, such as BioJava, and there are a huge number of other open source Java projects of all kinds.

See the page Java Resources on this web site for a list of popular and user suggested online resources.

Please send ideas and opinions by email at alexamies@gmail.com.