Approaches to Web Development for Bioinformatics

C and C++ Compared with Other Languages

C is the ideal language for programming operating systems.  C++ is a good choice for commercial desktop applications.  However, it is probably not the best language for most bioinformatics applications.  Some disadvantages are

• It is not very uniform across platforms.  The same features that make C ideal for programming operating systems and their utilities make it too close the operating system for convenient cross platform programming.  For example, the byte order is different on Windows than on UNIX systems and, although utilities exist to assist with this, programmers must be aware of the issue.  Perl and Java programs work on any operating systems that Perl and Java runtimes are ported to.  However, a C or C++ program must be compiled for each platform that it is to be run. 
• Linking with the right versions of libraries can be an additional issue. This can be a problem with any language. However, on Windows shared library versioning is not supported and install programs just copy the dll files into the windows/system32 (or equivalent) directory.
• An error in a C program usually leads to the programming crashing.  In C++ there is an exception handling mechanism but it is not as consistent as, say, Java.  Application availability is very important to web users so it is important that crashes do not happen.  C programming is not for the faint hearted.
• Programming in C and C++ is hard.  Programmers have to manually allocate and clean up memory.  Lack of vigilance here leads to the common occurrence of long running programs leaking memory.  C++ is a very complex language, which can be a barrier in itself.
• Programmers must be extra careful about security.  This is especially important for web and network programs where buffer overrun attacks have been the source of many successful attacks.  Because the programs are compiled to machine instructions a successful hacker can gain direct access to the operating system and can take control of the machine.  Other languages often have checks for buffer overruns built into the network libraries.

Despite these disadvantages there are valid reasons for using C or C++.  C is the dominant language for application development for the desktop, embedded systems, and server software like databases and web servers.  You may be starting with a program that is in C or C++ and be unwilling to rewrite it in another language.  It is hard to beat C++ on the Windows desktop for responsiveness and flexibility.  Being more familiar with C than with other languages may not be a good reason to use it.  If you can program effectively in C and / or C++ you can easily pick up other languages.

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