Sunday, April 20, 2008

Compiled versus interpreted languages

During the design of an application, you might need to decide whether to use a compiled language or an interpreted language for the application source code.

Both types of languages have their strengths and weaknesses. Usually, the decision to use an interpreted language is based on time restrictions on development or for ease of future changes to the program. A trade-off is made when using an interpreted language. You trade speed of development for higher execution costs. Because each line of an interpreted program must be translated each time it is executed, there is a higher overhead. Thus, an interpreted language is generally more suited to ad hoc requests than predefined requests.

Advantages of compiled languages
Assembler, COBOL, PL/I, C/C++ are all translated by running the source code through a compiler. This results in very efficient code that can be executed any number of times. The overhead for the translation is incurred just once, when the source is compiled; thereafter, it need only be loaded and executed.

Interpreted languages, in contrast, must be parsed, interpreted, and executed each time the program is run, thereby greatly adding to the cost of running the program. For this reason, interpreted programs are usually less efficient than compiled programs.
Some programming languages, such as REXX™ and Java™, can be either interpreted or compiled.

Advantages of interpreted languages
There are reasons for using languages that are compiled and reasons for using interpreted languages. There is no simple answer as to which language is "better"—it depends on the application. Even within an application we could end up using many different languages. For example, one of the strengths of a language like CLIST is that it is easy to code, test, and change.

However, it is not very efficient. The trade-off is machine resources for programmer time.
Keeping this in mind, we can see that it would make sense to use a compiled language for the intensive parts of an application (heavy resource usage), whereas interfaces (invoking the application) and less-intensive parts could be written in an interpreted language. An interpreted language might also be suited for ad hoc requests or even for prototyping an application.

One of the jobs of a designer is to weigh the strengths and weaknesses of each language and then decide which part of an application is best served by a particular language.


No comments: