Wow, i'd disagree with a lot of what was written in these posts

Jorge:
-=> is not neccesary if you know how to handle data
If you know Delphi, think of Firefly as a subset of the VCL data
components available (both visual and non-visual).
In Delphi, sure you can handle data by hand, but after 5+ years
of using Delphi, i've never seen an application which does!
The problem is that Flash (both the player and IDE) is still maturing,
so things have room for improvement. Each to their own I guess...
But I'm not even sure DCK/Firefly will make it a version 2 though,
as it looks like Macromedia may have other things are in store for us...
Justin:
-=> To try to answer your question though, I would guess that the firefly
-=> components are merely wrappers for the existing flash remoting classes.
Would suggest that further reading of the Firefly API is required,
perhaps some of the Remoting components are wrappers, but
I challenge anyone to show me how remoting handles resolver
updates (ie. changes to a dataset) like the Firefly Resolver does!
Don't know what I mean? Run this live demo:
http://flashpash.sshnug.com/documentation.php?menu=TimeTraxExampleOne call to method "doSaveUpdates" on the client builds all changes
to a dataset (not just the last change) and sends to server, and there's
only a couple of lines of code on the server which handles all INSERT/
UPDATE/DELETE SQL queries. (see the tute for code example)
Okay, so i've written a lot of code to enable the framework for this to
happen, but sorry, IMHO AMFPHP and Remoting alone doesn't touch this!
Whilst i'd agree that native AMF will always be faster than XML,
using a proprietary format will limit you for some things (eg. if you
need to change your client-side technology but want leave the server
alone), and I think using an XMLsocket connection will offer a similar
performance level to AMF (my goal for next version of flashPash

Anyway, i'm not having a go at you guys, I just have a different opinion,
each to their own!
peace
si