>>I think you are saying where you are linked from does effect ranking though<<
Not my links, but competitive rankings I've been tracking as I'm analyzing how to rank in the "widget stores" area, which is highly competitive.
I've noticed that several of the top 10 have one thing in common... they're listed in the Regional> Widget Dealers> Location> Directories category... whereas sites with more or less equal optimization (so it seems... it's very difficult to analyze someone else's site quickly) that are listed in other categories aren't doing as well.
It could also be the content of link text (something I haven't fully checked yet) as well as other links. I was fascinated that one of the very high ranking sites only had passable optimization and 4 links (as opposed to 1,200 for the site above it). One of the links was from the Yahoo 'Directories' category... and the other was from a Stanford University site, apparently a student page, that had the exact plural form phrase linking to this site.
Again, I probably haven't made enough observations to generalize about the category listing... Brett's comment prompts me to check out some of the other factors more closely.
The problem is that there is no definitive spec for robots.txt. As far as I know, there is no W3C or RFC spec for it. So it means whatever a search engine wants it to mean. Do search engines use wildcards? Google does. Does an earlier rule override a later one? Looksmart seems to think so. But does Google? Who knows what Google really does with Looksmart's robots.txt.
The problem is not just for spiders. How do you come up with a correct robots.txt, when the spec is ambiguous?