I'm not sure either. I can tell you this. A couple of days ago I tried using TorqueBox with JRuby. After struggling a little getting the right JRuby (within TorqueBox) to get started and finding out what to do with Rakefiles, I managed to get a simple "Hello World" kind of Ruby (Rack) app running with it in-loco. I just went to the directory of the config.ru file. Created a Rakefile there. Then ran commands like rake torquebox:deploy and then rake torquebox:run

It started the JBoss process and at the end of it loaded my simple app. I tested it a little on the browser and ran the Apache benchmark tool on it like ab -n 1000 -c 10 http://127.0.0.1:8080/hello

After it had warmed up, further benchmark runs were much faster.

On my weak hardware, going up to 2.2k requests/second or something. I was like "wow! Give me more of that." Not really that it couldn't be matched by other technologies for that kind of stuff, but that it was a single process easily configured.

If we think that the same process could be shared by many other scripts, programs and so on, it's a good way to reuse all of that in innovative ways.

Deep down it's not much more than having a stable multi-threaded environment that you can rely on.

Then I thought: wait! When I started with Ruby, I was running from all of that kind of complexity!

It's as if we were being assimilated by Java again. Only this time we don't have to use so much Java or xml.

I had a similar issue. It'd be worth expanding the description to be more concrete. What I see there now seems more marketing/management focused than focused at the people who are likely to actually be using it. :)

Here's the best answer I could come up with: http://bit.ly/wtfenterprise

Disclaimer: I'm a TorqueBox/Immutant developer

Yeah, they need to simplify their 'messaging'.

Oh, and see here for what an applicaiton server is: http://en.wikipedia.org/wiki/Application_server

TorqueBox's home page explains the use-cases well.