Like many (most?) programmers working collaboratively, I have often wished I had the "perfect" issue tracker, but my version of "perfect" maps very tightly to the way I think software "should" be developed, at least among teams of people working full-time for money.
In spite of my natural tendency to be sure, like every good programmer is sure, that my hypothetical solution is vastly superior to all the horrible, slow, confusing, buggy things that actually exists, I am forced to concede that the corporate software world is not following my implied instructions.
So, in all seriousness, I think there are three huge challenges facing the would-be creators of an issue tracker that would be $GREAT:
1. You have to either accommodate a bunch of different workflows and be good at most of them, or commit to One True Workflow and hope you can convince people you're right.
2. Even if you actually do make one that's more $GREAT than anything else, the Enterprise world will not care -- and that's where the only serious issue-tracking money is, so you'd better do this for love and/or have a very long runway. And if you get so far that they ever do care, you'd better have a bunch of data-migration "whitepapers" ready.
3. In order for anyone on a small team or open-source project to switch to your software, you will have to convince them not only that it's better than whatever they're using now, but also that it won't be offensive to people with longstanding allegiances to other software. And making it unoffensive is likely to contradict your definition of $GREAT.
All that said, go for it! You might be your only customer, but if that means you've got the greatest issue tracker in the world then it might be worth it.