Comments on: Embedded Programmers Worldwide Earn Failing Grades in C and C++

By: Vaibhav Garg

Vaibhav Garg — Mon, 01 Feb 2010 11:48:24 +0000

With respect to your analysis on Embedded.com on the demographies of results, there is a fundamental problem with the quiz that might render all location data irrelevant. The default location is the USA, which many might not have bothered changing. At least 2 people didn't, and they sit right next to me in India.

By: fileoffset

fileoffset — Thu, 28 Jan 2010 13:21:25 +0000

Anyone that gets more than 5/10 correct is very good at C++, anything more than 7/10 and I dub thee a C++ expert.

By: Anonymous

Anonymous — Mon, 18 Jan 2010 20:54:21 +0000

"You got 10 of 10 correct. Your score: 100% Congratulations! You passed Netrino's Embedded C Quiz."There were two that were tricky and obscure. There were also two that I wouldn't expect non-embedded people to know. It was a good test.

By: William

William — Thu, 14 Jan 2010 17:05:31 +0000

Food for thought on why C skills are lowerConsider Myself Average EE16 Semester Hrs College work in Assembly Language — Have written whole small accounting systems, avionics, other applications(mostly in Assembler)1 Semester Hr in "C"(Struggled trying to make C work before there were 8bit "C" compilers that really were worth anythingnot even anywhere as good as SDCC now)2 week seminar in C++(If you limit what you use in this language — it can work for embedded — but probably could have done the same things in Assembler without learning a whoole new language and done it with an 8051 instead of a PowerPC)

By: Michael Barr

Michael Barr — Wed, 13 Jan 2010 20:32:03 +0000

I recently wrote a more up-to-date and thorough analysis of these quiz results in a column for Embedded Systems Design magazine. The URL for that is:http://www.embedded.com/columns/barrcode/222300586

By: Anonymous

Anonymous — Mon, 14 Dec 2009 21:15:02 +0000

Hello Michael,I am one of the embedded developers who did not do well on your test, and I thought you might be interested in some of the "whys" of this result.To give you some background – I have been programming 8-bit micros as part of my product design efforts (I program the uC's for my projects, butr I'm not a professional programmer, obviously!) since the late 70's. Virtuall all of my programming knowledge has been self-taught (I have not taken any programming classes). Initially, I started with machine code (on an old KIM-1, using a MOS Technology 6502), assembler (Motorola 6800/6802, 6809, HC05, HC11), and finally, "C" (HC05, HC11, PIC, and Atmel AVR). All of my experience has been on "small" systems. None of my projects has used up more than 16K of code, and most use less than 2K. I use "C" primarily as a glorified assembler, as I have to get down to the bit level in most of programs. Almost all of my variables are global, because the values contained therein need to be accessible to the various functions. Most code is predominantly interrupt driven, and is very "real time". Certain interrupts cannot have more than a couple of microseconds of latency, so I have to juggle things around to make sure that this occurs as required.I virtually never use pointers – I just access arrays of char or int as required using the normal indices.In looking at the questions on your quiz, many refer to things such as #define statement syntax or pointer declarations that I just have never used.Also – questions regarding compiler variations are completely outside of my area of experience. I have used one compiler (IAR EWB) for the past 10 years or so, because it works well & I've had no reason to change. Since I almost never use qualifiers such as static, volatile, etc (since virtually all my variables are volatile, I suppose) I really have no idea whether or not a compiler would allow a variable to be declared both ways (such as volatile & constant – it seems a bit contradictory).I don't use "modules" (to tell you the truth, I am not copmpletely sure what you mean by that) my only concern about static variables is what pertains top their use within a function.At any rate, you can see that I probably don't fit your definition of an embedded programmer very well – perhaps I'm better described as an engineer who programs, but I have written some pretty good, stable code over the years.