On Two Part Responsibility in Programming

Its interesting, I was reading an article recently about how the new Solid State Drives have often underperformed standard hard drives in tests because of the way Windows Vista is optimized for writing data to disks. This is certainly not the fault of Microsoft per se. I do not mean to bash anyone here. It just intrigues me that a technology heralded to surpass magnetic disks could be doing poorly just because of the way the software is written. Certainly if an operating system were optimized for random access storage like SSDs, they would be light years ahead of magnetic storage.

(For those who do not know, Solid State Drives use flash memory for their storage, much like the memory cards in digital cameras or new iPods. Magnetic disks are metal platters that look something like small CDs that spin and have data encoded magnetically on them.)

What’s happening is that manufacturers of these SSDs are redesigning the inner workings so that they will work better with current operating systems and pretend they are more like magnetic disks. This all strikes me as odd, and it poses the question of whos responsibility it is to do the changing. Normally I would say that the software should be rewritten, but the fact that it is much easier to redesign the drives than to ship a whole new operating system makes the market work the way it is now.

I’ve often come across something like this dilemma when programming. When there are two classes or methods interacting with one another and I need the two to perform a task once, its often an odd decision of picking which of the two I put the action in. Practically it could be either one, just as practically it makes little difference to consumers whether Microsoft or Samsung does the work of redesigning their product. Theoretically though, its a choice of which is more elegant, more perfect. Its a process of writing cleaner code that encapsulates my ideas and commands in the best possible package.