"An entry is useful or it is not. It's as simple as that."
Some articles may be useful.
Some articles may be so useless that it's worth our time and energy, as well as the risk of alienating the article's contributor, and the inclusionists, to delete them.
On the other hand, some articles may be useless, but not so useless that the benefit of deleting them outweighs the negatives listed above.
I used to be a deletionist, and in fact I still am, but I've moved my deletionist stance to the eventual "1.0" version. Let's face it - the current Wikipedia is great for compiling a large quantity of pretty good (on average) information. It's lousy at compiling consistently very good information. So why try? Only a more structured project will ever accomplish the latter, so it's pointless to be perfectionist about the current Wikipedia. We should shift our perfectionism to the "1.0" version.
Here's another way to look at it. Some people think that the perfect electronic encyclopedia should include every verifiable fact in the world (let's call them the exhaustivists), while some think that it should only include notable or important or "encyclopedic" facts (the filterists). The nice thing about shifting our perfectionist worries onto the "1.0" version, is that it accomodates both camps when thinking about the current version of Wikipedia. When someone adds an article on a non-notable topic, the exhaustivists will of course be happy, but the filterists (like me) won't mind *too* much, because that article will get weeded out of the great 1.0 version (we hope).
That's the only way I see out of this debate. Otherwise we'll just never have peace between the two camps - the gap is just too wide. (Maybe there'll be two "1.0" versions - an exhaustive version and a filtered version.)
Alex