He's Not Wrong...
Over at Net Meister there's a nice peice entitled Things They Don't Teach You in School". A lot of what Jan has to say is spot-on observation and good advice.

It's definitely worth a read.

[/code] [permanent link]

Looking back...
As I look back at code I wrote a decade ago as an undergrad, I often find lots of little things that can be done better. For instance, in one file I found that reversing the order in which two functions were called would have eliminated a half dozen conditionals from one of the functions and would have resulted in the same expected behavior, but with fewer lines of code and a lower cyclomatic complexity.

[/code] [permanent link]

Thoughts on National Computer Science Education Week
This week is apparently National Computer Science Education Week. Code.org is organizing the "hour of code" to promote teaching of Computer Science and Programming in schools. They're also organizing petitions to make CS courses count as credits in Mathematics or Science for High School graduation requirements.

In High School, my CS courses were by far my favorites, Programming in Pascal, AP Comp Sci in Pascal, Programming in C++, and AP Comp Sci in C++ ( the language for the exam switched my junior year). I learned a lot about structured code, elegant, efficient code. I learned enough about Data Structures and Algorithms that I didn't have to study for my college CS classes until Computational Structures (Discrete Math II with Scheme, essentially) in my third semester. I had an amazing Computer Science teacher who also taught me Calculus and the proper order of precedence in life: God, Family, Math. I wouldn't be where I am today without that educational opportunity I had in High School. I want others to have that opportunity too.

However, this is where I differ with the opinion of the Code.org folks. I do not believe that CS classes should count toward the Math or Science requirements. In this state, CS counts toward the "practical or performing art" requirements, I'm assuming under the "practical" label. I think this is a better place for it at the High School level.

Computer Science is not a hard Science. It's not Physics. It's not Biology. It's not Chemistry. There's a saying that if the subject has science in its name, it's not really a science. That is true with Computer Science. It's not studying the how and why of atoms, of molecules, of living systems, of anything really. It's not science.

Computer Science is really applied mathematics. I am very fortunate that the college program I went through was very strong in mathematics: Calc I and II, Linear (Matrix) Algebra, Discrete Math, Discrete Math II in the guise of Computational Structures, Probability and Statistics, Theory of Computation, Algorithmic Analysis... the list goes on. All of these mathematical foundations were then applied to a machine, to make the machine carry out a task in an efficient manner. It's those mathematical foundations that are the true core of Computer Science.

While mathematics is the core of Computer Science and Computer Science is essentially applied mathematics, I do not believe it should count toward the Math requirements. The CS classes would likely detract from other mathematics courses such as Geometry, Trigonometry, and Calculus. These courses are far too important to an education to be replaced by a Computer Science course. Many, maybe even most, High School Computer Science courses focus more on "programming" than the fundamental mathematical theories. They will pick the language du jour and teach you the syntax and semantics. They'll teach about basic data structures like arrays, and linked lists. The AP exam currently focuses not on implementing lists, trees, stacks, queues, and sorting and searching algorithms, but on arrays and lists using Java library calls. This is not math. This is learning Java syntax.

[/code] [permanent link]

unHide
After a friend's computer became infected with some malware that hides files in your home directory and tries to extort money from you in order to "recover" your data, I was inspired to write this little program.

All it does is search for hidden, non-system files and unhides them. It defaults to the user's home directory (My Documents), but other directories can be chosen.

If anyone wants to try it out, please do. If you do try it, let me know what parts I can do better. I primarily work on embedded systems and Unix daemons professionally; the user only knows my software exists when it isn't working. As such, I have very little experience with GUIs or human-computer interactions. Any feedback would be much appreciated.

The Windows executable can be found at: http://www.skinnymf.com/~mforde/unHide/. Source code is available upon request.

[/code] [permanent link]

Just a Question
Given the following code fragment, I was asked to implement foo such that the program would output "America." How would you do it?

        int main()
        {
          char *p = "Hello";

          foo(         );
          printf("%s",p);

          return 0;
        }

My solution involved allocating new memory from the heap to store the new string, and changing p to point to that buffer. They didn't like that answer. They preferred the method of putting the new string in the data segment as well.

I personally would always avoid that, whenever possible. "Hello" is stored in a read-only area of memory as is "America" in their preferred solution. Any attempt to alter those strings will trigger a segfault. This is an accident waiting to happen.

[/code] [permanent link]

It's not magic, it's C.
I love reading comments like

/* These defined magically in the linker script. */

I found that in the GNU Standard C Library implementation when GCC told me the the variables to which the comment referred were undefined. I guess that linker script isn't magic after all...

[/code] [permanent link]

C#, XNA, and 8 Queens
I spent some time this weekend learning some C# and getting working a bit with the XNA framework. I implemented a solution for the 8 Queens problem in C++, then ported that class to C#. After getting that working in Windows, I started moving it to the Xbox.

It works there with little issue, but, as expected, writing to system.console doesn't produce useful output. My next step is to get some sort of graphical representation of the chessboard displayed with the solution set.

[/code] [permanent link]

XNA
As of this morning, I am now a member of the XNA Creators Club.

[/code] [permanent link]

Self Documenting Code
Despite what you think about your code, it is not self documenting. When writing code add comments, describe what a function does, explain why you're using that convoluted pointer arithmetic and bit shifting. I'm sure it makes sense now, but someday you'll have to go back and look at it and figure out what the hell you were doing. Or worse, I'll have to go back and look at it and figure out what the hell you were doing.

Sometimes it is important to have some sort of separate documentation for the code; maybe some UML or ER diagrams, maybe some English text. Use what ever it takes to explain what you were doing. Do not paste snippets of your code into the text and call it documentation.That is not documentation, that is code.

[/code] [permanent link]