18-02-2008, 08:46 AM
Doesn't this whole issue show up the weakness of the MISRA-2004 (and earlier) approach?
MISRA argues that many problems arise because of \"misunderstanding and difficulty in the C language\"/\"Misconceptions among programmers\". Instead of attacking the root cause, it creates an unwieldy set of rules designed to \"neutralize [the effect of integral promotion] by taking no advantage of the widening that occurs with small integer operands\". The number of casts required obfuscate the code, and weaken the compiler's static type checking: the compiler cannot detect an improperly placed cast.
One idea to eliminate \"misconceptions among programmers\" is a training & certification program. The issues are limited -- they represent only a handful of pages in the standard -- so such a program could be largely automated, e.g., via the Web.
Regards,
Konrad Schwarz
MISRA argues that many problems arise because of \"misunderstanding and difficulty in the C language\"/\"Misconceptions among programmers\". Instead of attacking the root cause, it creates an unwieldy set of rules designed to \"neutralize [the effect of integral promotion] by taking no advantage of the widening that occurs with small integer operands\". The number of casts required obfuscate the code, and weaken the compiler's static type checking: the compiler cannot detect an improperly placed cast.
One idea to eliminate \"misconceptions among programmers\" is a training & certification program. The issues are limited -- they represent only a handful of pages in the standard -- so such a program could be largely automated, e.g., via the Web.
Regards,
Konrad Schwarz
<t>Konrad Schwarz<br/>
Siemens AG<br/>
CT SE 2 (Corporate Technology, Software & Engineering - Embedded Systems)</t>
Siemens AG<br/>
CT SE 2 (Corporate Technology, Software & Engineering - Embedded Systems)</t>