I don't have a link to a good paper on "Why you should use Ruby", but I can relate my experience from something I recently had to do. You might find it useful.
When I'm trying to learn a new language, I write a simple application that solves a common problem. Its the same problem for every language (the little peg puzzle you see on tables at the Cracker Barrel restaurant). That gives me a feel for the base language and how to use it. Then, if there's a web framework for the language, I try to implement a web site in (again, the same web site for every language/framework).
A little while ago, I implemented my website using Ruby/Rails. There was a fair amount of learning involved, but everything went fairly smoothly. When I had questions, I figured out what I needed to know and used Google to find the answer. Sure, there were lots of questions, but it was easy to identify what those questions were and to locate good answers via search. Most things just "worked as expected" one the first or second try, and it was obvious why things did/didn't work. It was a very positive experience for me.
A little less time ago, I decided to do the same thing using JSP, Hibernate, and Spring. I'm fairly experienced with JSP (having worked on multiple sites and services that use it, but without frameworks), but Hibernate and Spring were new to me. However, since I've worked with both ORM and MVC before (ex, Rails), I figured it shouldn't be too hard to learn the JSP version....
That, was not a good assumption. Hibernate and Spring are both very powerful. However, there are so many different types and combinations of configurations and annotations, that it's very hard to figure out what it doing what, and why. The error messages tell you what's wrong, but it's extremely difficult to figure out what question to ask to find out how to resolve the problems. Honestly, I could write a dozen pages on various things I found difficult to deal with. There was an awful lot of googling with "yeah, that's the error message, but the solution discussed uses a different type of configuration than I am, or just plain doesn't solve the problem". There was very little "look, sample code... and I can see from it what the writer is doing and why; that helps me".
At the end of the day, the Ruby/Rails implementation was a pleasure to work on. I learned a lot by just jumping in and trying to do what I wanted to, and then asking the questions that cropped up. The JSP/Hibernate/Spring version was... less pleasant. I felt like it would take months of working on such code just to get to the point where I knew what questions to ask when; like I'd need to be experienced on the platform just to be able to know what questions I should have for a given issue.
I only recently started putting my work in SourceForge. As such there's only 2 implementations of the puzzle game (Java, Haskell) and one of the site (JSP/Hibernate/Spring, and it's nowhere near complete). I'll add the Ruby/Rails version of the site later; I have the code, but I want to remove certain parts that I used from other OS libraries and replace them with my own code. That being said, I included the links to them /shrug