I sometimes think a lot of C portability problems are due to lack of namespacing in libc. It's a shame we can't somehow deprecate and move to (eg) ... It would contain familiar functions with a "c_" prefix: c_printf, c_write, etc. This would provide considerably more flexibility to the C implementors and standardizers - for example, I recently ran into a problem with new glibc (patched by a distro) defining "getline" as a macro. (This was in lua's source.. not mine, as it happens)

It's frustrating b/c I very much want libc to innovate, but this disdain for reasonable prefixing policy makes it difficult to know what will compile tomorrow...

why conservative?

You might wonder why I chose to blog as the conservative coder.

It sure has nothing to do with my political leanings!

Instead, it has to do with skepticism. Every year ideas come out that promise to change the face of programming: new languages, paradigms, management techniques.

And indeed, these concepts do heavily influence our thinking about problems. In fact, I think the best ideas maybe don't directly try to change the way we think, but rather try to make our thoughts more cogent. For example, the Mythical Man-Month debunks ideas that can be readily seen in the field. It does not try to argue for new workflow techniques or the like. That is why I think the MMM is much more likely to stand the test of time than ideas like Extreme Programming. Another great example is Design Patterns. This straightforward, unpretentious book outlines a number of approaches to problems that work. It's a way of thinking!

What really motivates me to write this is that most programming blogs are falling all over themselves to be current, even edgy. Speak ill about Haskell at the moment, and you'll be branded a luddite. This blog is for people thinking the Kool-aid tastes a little funny...