People are constantly changing directions, and constantly trying out new things which may or may not be useful to them in the future. We're all better off if that code gets released. In turn, we can't expect people to support code they may no longer use, or may not have ever really cared about in the first place.
It should be clear that running an open source project is something completely different. If you make an attempt to start a community, you probably should keep up with that community, or at least responsibly exit it. But people usually build communities around things that are really important to them, or that they have an ulterior motive for advocating (not necessarily a bad motive, just one beyond pure open source).
From where I'm sitting, if Gruber isn't adding anything to the community, he should have been made irrelevant by that community.
Click on anything that's not a JavaScript project, and you get a broken link. Would you rather he spend time fixing them up, or would you rather he spend time improving JQuery?
It's great that this guy works on JQuery, but I think the minimum that should be expected is to properly tie up loose ends: Fix broken links, make sure it's clearly stated that the code in question is considered done (feature complete, unmaintained, etc). Officially hand it off to someone else if possible.
There're a bunch of counterexamples to the "all projects that are making an impact are doing so because they are actively backed and developed." Markdown is one obvious one. The Internet Explorer team was disbanded from 2002-2005, even though it was the most popular browser on the market. Several Google projects that are used by millions of people a day are completely de-staffed.
And there are many more projects that are actively backed and developed and yet are not making an impact. We'd be much better off if their developers quit what they were doing and found something else to do.
Besides, the Markdown license explicitly says it's "as is" with no such support promises.
I don't think you can even call Markdown users Gruber's userbase. He created Markdown because it works for him. If you don't like how it works, change it and call it something else (MarkSide? MarkIn? MarkOut?)
Why? Why are there any commitments whatsoever attached to making code publicly available? I could just as easily keep it to myself.
So, yes, it is better to release nothing than something buggy and unsupported. The former doesn't waste everyone's time pretending to be a solution.
The attention and effort others invest in your project needs to be respected. You do have an ethical obligation to follow through on the project, at least in my view.
Feeling no obligation towards your user base just because one of the price dimensions (money) happens to be 0 is an irresponsible attitude.