Based on a quick look, it seems like Cyclone provides a slightly nicer way to specify how to handle the various requests than does a plain Twisted web application. Are there any other advantages to using Cyclone as opposed to plain Twisted?
Cyclone provides a much nicer API than Twisted with routing done right. I haven't played around with Twisted much, but I'm liking how Cyclone works. Cyclone does a bunch of stuff like converting a dict to a json and setting the proper response headers automagically(which was a "Aha" moment for me, I'm not sure if Twisted does all this).
Cyclone+Twisted: slightly nicer way to write the web application. More dependencies for Onionoo.
Plain Twisted: not quite as nice a way to write the application. Fewer dependencies for Onionoo.
I'd definitely prefer Cyclone+Twisted. Cyclone ships with a couple of Debian scripts that make it easier to run as well.
Also, the documentation for Cyclone seems...minimal. It might be straightforward for someone used to using Tornado, but that doesn't describe us over here...
Well, that's true. There isn't much documentation, but it uses almost the same API as Tornado so you can pretty much use Tornado's docs. But that isn't much either so I kinda ended up reading Cyclone's source and that worked. It's actually pretty easy to follow. But, I can see how this can become a real pain.
-Sathya