Maybe they're not pressing for an unreasonable amount of perfection. Maybe they have a different view of what perfection would be, of when it would be reached.

Part of the discontent I read in this thread could come from the fact that people have very different perceptions about the purpose of this program. (Just a hypothesis... it's how I read some of the negative posts here.)

I'm not really sure myself, I have to admit.

I come from the old N810 days when we read this announcment:
and this blog post:
(emphasis added)

So for me, ever since these days in Oct 2007, it was crystal clear that whatever device program comes next, it would not be a developers program (unless stated otherwise, as for example with the Fremantle Stars). This was my expectation, and whatever I read since, I read and understood in this context. (BTW, I had also expected since Oct 2007 that karma would be the indicator for the next device program.)

I was 100% sure about this and couldn't understand why some opposed karma as a valid (and agreed upon) indicator... It wasn't a developer-program anymore, was it?

Re-reading this thread and more recent material now, especially this:
... plus the title of this thread (where the word "developer" suddenly is back), I'm no longer sure. Was I wrong in the first place? Was I the only one who had wrongly read the recent announcement in the context of the old device programs?

The bottom line is: Some may see imperfection only because they thought the whole thing was created to reach a different goal. Coming from the "Note that we have removed the word “developer” to make it clear..."-past, I don't see imperfection in the way it's handled. Not at all. Quite on the contrary, I see it successfully addressing a lot of the issues people had with all previous programs because those were tied to the Nokia online stores.

It all boils down to... communication.

And yes, karma isn't 100% perfect. May I quote Quim:
(emphasis added)

So I don't see much use telling the Maemo team @Nokia that karma could be re-calibrated... They know, we know they know, and solutions are already being discussed.
(Discussing it, though, automatically means evaluating tasks people do for/in the community. What's more important: A bug report or a blog post? How many thanked posts should be equal to one ported application? ...? It should be clear from the start there'll be never consensus on this.)