As a software developer, I'm generally drawn to complexity, and I generally view "hard" problems as much more worthwhile, from a business perspective, than easy problems.

For instance I never considered many of the "successful" .COM-type ideas as worthwhile, and I naturally feel the same about many of those percolating up via Bubble 2.0, because they either seemed too obvious, or too easily duplicated (aside from the fact that most of them haven't a hope in heck of ever having a sustainable revenue model...but that's beside the point -- If you can get Yahoo or Google to buy you out with big dollars, it doesn't matter if your flighty userbase would never pay a penny, and that they would never tolerate a single ad impression).

As such I discarded a lot of "neat ideas", only to see them bring success to someone else. I was keeping my eye out for something both unique and difficult.

Yet often it's the simplest of things that hold the most value to people.

Just over a year ago, late December 2004, I was working on some custom C++ JPEG parsing logic (scary pointers and all), and noticed that some of my test files were bloated up with a lot of extraneous data. We're talking 4KB JPEGs that had an extra 60KB+ of data appended to them. Apart from EXIF data -- useful at times, but completely irrelevant at other times -- Adobe Photoshop was particularly notorious for stuffing files full of worthless application blocks.

Many of the images you find on the web have all of this extraneous info, slowing transfer times, increasing server loads, hiking bandwidth bills, causing pestilence and suffering (okay I'm going a little overboard).

As such, I gathered up a ridiculously small amount of the parsing code, compiled it, and "released" it as the rashly named PureJPEG. I expected it would see maybe a dozen downloads from nitpicking webmaster looking to ultra-optimize their user's experience.

Over the past year, over 40,000 people have downloaded this utility directly from yafla. It's been mirrored on quite a few sites (particularly in Russia, for some reason), so I have no idea how many worldwide have downloaded it, but presumably a greater number still.

I would have never imagined that something so simple would have filled such a niche, and I'm a little embarrassed about how trivial of a micro-project it really was, but there it is. As far as little utilities go, it's been a stunning success.