_why isn't the best example - he was awesome, but part of his appeal was that he was distant and not on the same plane as the rest of us.. just not in a negative way. Nonetheless, before he disappeared he tweeted:
programming is rather thankless. u see your works become replaced by superior ones in a year. unable to run at all in a few more.
Even though he didn't attract much negativity due to his whimsical, impenetrable persona, it didn't seem to help him feel much better about his excellent work.
I suppose in our profession we are able to perpetuate ourselves by communicating our ideas into the world. Our influence is felt in posterity through the inspiration of thought in our successors. The kernel may evolve over time, but its evolution may be guided by the initial conditions of its initial code base (and subsequent changes).
A civil engineer may be directly remembered by the the bridge that bears his or her name; a software engineer may be indirectly remembered by the smart code he or she has written.
You want a legacy? Move up the food chain and start discovering new data structures and algorithms. Or work on a large project like GMail.
Writing about programming -- techniques, thought processes, libraries, and code -- is highly valuable, at least in my opinion.
On the other hand, posts that are informed by real-world experience and stories that discuss specific implementations are almost always at least somewhat interesting. This is what Joel came up on; he became well-known because his blog detailed hiring practices and how they've worked out for his company, certain strategies and fallacies at Juno, Microsoft, and Fog Creek, and the outcomes of them, and informed analyses of progress at the institutions wherein he has specific knowledge (primarily Fog Creek, I reckon).
Joel's posts were interesting because he had a lot of stories and a lot of anecdotes to back them up. His blog doesn't simply say, "I think if you give programmers offices with good equipment you'll end with some pretty good programmers", but "I thought if you gave programmers offices and good equipment you'd get good programmers, so here at Fog Creek we do that. Here are some pictures. When I was at Microsoft, everyone was sad because middle-management got the glory and the offices. At Fog Creek, we focus on our programmers, and it greatly increases our productivity, morale, etc.". One of these is much more interesting.
"full of himself", "aloof and distant", "For the average joe schmoe programmer (implying only they can appreciate his work)"
For the average joe schmoe programmer, people don't have such expectations, so there is far less negative reaction. Also, personality does play a huge role. Someone who doesn't seem too aloof and distant from other people will be better off (take _why for example).