Both professionally and as a writer I am bumping up against this aphorism in my current projects.
One of the things they tell you as a writer is to keep revising, keep changing, keep editing, keep making the book better. The same is true of software, though unlike writing often the process is keeping up with changes other people are making that affect your work (I’m looking at you Microsoft).
The first thing I learned as a professional programmer was the difference between the ideal perfect solution, and the practical, applicable, “quick and dirty” solution. On deadline you don’t have time to make flawless code. And truthfully flawless computer code is a lot like haiku’s: not very long and can only express a few things.
Writing is much the same way, especially if you want people to actually read your stuff. Eventually you have to reach the point where it is okay to put something out the door. If you’re constantly rewriting based on your evolving standards (changing user requirements) you will never deliver a product. This is not to say you should send something out that is half finished and buggy. Even though ebooks are becoming more like software in that you can push updates out to everyone who purchased them, you still have to deal with initial market impressions of you. If you become known for making crappy software, or writing crappy books, no amount of post-release revision is going to fix that. Then your only solution is re-branding (maybe a pen name).
So how do you know when something is done? Maybe it’s a fixed number of revisions, or even more practically a release deadline. Maybe it’s finding that fresh beta reader who hasn’t read a lick of the draft and hasn’t already formed impressions of it telling you they love your work. Whatever the case, sometimes you have to accept something is good, and will never be perfect.
How do you decide when something is done, and when it needs a few bug-fixes?
As a side note apparently the above phrase is commonly attributed to Voltaire though it has its origins in ideas from Aristotle and Confucious. Thank you Wikipedia.
[BTW] : Ben Trube, Writer 
I like the milestone of a fresh beta reader being able to look over your work and love it. However, I think that “fresh beta reader” should be yourself. You need to be able to set aside your work for a few days or weeks and be able come back to evaluate it as if someone else had written it.
Of course, this can be next to impossible professionally or on a tight deadline.