Mike
This is a repeat of what I've replied to you in other forums, but I'll post it here for the benefit of others following this thread.
While I agree with you that overly long urls can look like an overkill : the point is that the Url conveys important information about what is on the page, and visitors like that. In the days before database driven content, everyone named their static html files according to what was in them. The shift to database driven content dropped this because of the need to use primary keys in the url. The shift back to more meaningful urls has come about partly because the keywords in the url must be carefully chosen (you can only have one url), and partly because people prefer meaningful urls, so Google is just chasing the feedback loop. What people see as normal probably depends on what part of this long cycle they've come into regular web usage.
My first website was a geocities job full of classics like frames and animated gifs : but it sure had descriptive urls, because each 'page' really was a page full of hand-coded html. If I had called these pages 1.html, 2.html and 3.html, people would have been asking me why I used such a silly naming scheme.
Personally I was horrified the first time I saw an ebay-like url without about 1kb of encoded data in it. It's only through use that we've come to accept these types of urls.
I don't see Google as being at fault : blame the competitive nature of the web. No matter who is doing the search indexing, there's got to be a way of sorting the ranking based on suitability, and the Url is one thing that can't be easily spoofed, spammed or otherwise manipulated. John illustrates this point perfectly : in order to optimise for DotNetNuke Performance, he has to de-optimise for other terms. He's forced to choose the most relevant phrase and concentrate on that only.
I get frustrated too : there's an old forum post that shows up for "DotNetNuke seo friendly". It's out of date (2005), full of disinformation and yet is constantly in the top 10 results. The web would be better off without it, but what can you do? Thankfully, over time, people link to more relevant pages and the old post is gradually falling down the list. With poetic justice as well, seeing as it doesn't use a descriptive url.
From a programmers point of view, I'd prefer to just put in simple urls so I don't have to think about it anymore. But programmers no longer own the web : it's now the domain of marketers, tinkerers, content publishers and the rest. Try explaining to a non-technical user why their page should be tabid/25/articleid/246 instead of /my-article-name.aspx - I've tried explaining to them in the past and I don't think I've ever managed successfully.
I've done a recent blog post on this topic in response to the recent PC World article that was also critical of the long urls trend, and that Google was leading us down the garden path : why you should have long Urls. My main point is this : descriptive urls are good for users, so why not?
-Bruce