After having had a chance to work through the updated tutorial: http://wiki.apache.org/nutch/RunningNutchAndSolr#A4._Setup_Solr_for_search I think the Solr step is a bit awkward as well. One of the reasons Solr has seen great adoption is that the /example app is so well thought out, and easy to get started.
With Nutch being decoupled from Solr, but depending on Solr, I wonder if there is an issue of a user downloading nutch, and then downloading Solr, and the versions being out of whack? Like if Nutch depends on Solr 3.3, but Solr 4 has been released.
I could see taking a version of Solr and checking it in. Strip it down to JUST what Nutch needs, and then you could include in the Nutch version a nice Velocity based UI for just browsing through the data that Nutch returns.