Most coders get into FOSS in order to do what they love on something they care about. And many are professionals who want to avoid the whole code review and approval process they live with at work. Unfortunately, one big side effect of going that route is that the code tends to get a little sloppy and exhibit some sprawl.
The same goes for the relative absence of quality documentation. Most coders HATE to write documentation - and FOSS projects are dominated by coders. So if serious documentation does get written, it's often because somebody (usually a satisfied user with a technical writing background) nominated themselves to do it and put the time in. -40hz
I think there are a lot of issues in this topic, but the freedom of coders in FOSS (especially their own projects is one of them). Firstly, it's what provides the incentive for them to do it in the first place. But, secondly, as 40hz says, it gives the freedom not to do lots of things that have to be done on successful commercial projects: review, quality control, internal documentation, user documentation, making the product easy to use, making it pretty, support, website maintenance, responding to user requests, sustaining a long-term approach to marketing and developing the product.
Many developers are good at some of these, but very few at all. Users are mostly looking for a mostly complete package with maybe only a few areas of significant weakness. That is true for free as well as commercial products, but it is mostly the commercial entities that enforce a discipline to try to do the lot.
By and large coders take inspiration from within themselves (and seem to just want to code). The people who are just wanting to make money take inspiration from those who might pay them the money (and try to give them what they want and will pay most for, at least cost to themselves).
If you want to make enough money to live off from donations, you do have to take a commercial perspective in terms of giving customers what they want. And that is on top of having a good idea and implementing it well. And shareware authors are (nearly?) as likely not to do all the above as freeware authors.
One other thing I would mention. People like to know what the price/donation should be. They also like a discount. Trying to decide how much to donate, even if they're able to come to the point of deciding that they want to, is just too much; no-one wants to seem a skinflint, but no-one wants to overpay either.
I have vast amounts of free software. Most of it I never use, or use extremely rarely. I'm more likely to persist with software I have paid for (or donated to) and less likely to chase around all the alternatives (though there's a fair bit of commercial software I don't use as much as I expected too) - and this initial commitment is as important to freeware as it is to commercial.