If you have worked on either Flex or ColdFusion you know that nowadays there is more development frameworks than people know what to do with.
At one point having a consistent pre-defined path to application development on either of these platforms meant you had to invent wheels many times over. Now, the situation has changed to one where you attempting to decide which framework to use; long discussions and passions are aroused.
At some level this has become as counter-productive as the use of no framework was before.
Here are the commonalities I noticed:
a) Frameworks are for developers:
You have to learn the platform, learn the framework and have some sort of knack for writing code.
b) Frameworks only cover one or the other technology.
They do not span technologies, even if your application is expected to. Thus if you have to write an app that uses a ColdFusion back-end and a Flex front-end; you have to put the puzzle together yourself. One may even say we are back where we were several years in the no framework days.
I understand, there are wizards to get you started in both directions, e.g. CF wizards for flex, and flex wizards for CF, but a wizard a framework does not make ;o)
Don't get me wrong. There are some cool things being done with scaffolding to generate fully fleshed out UIs that are generated from a database, a la Ruby. For example, check out Apptacular.
c) Frameworks focus on developer application parts
Frameworks focus on allowing programmers a base structure that can expanded. Most of this centers around the use of the MVC pattern in one form or the other. However, less attention is given on other services needed in order make an application complete or usable.
For example commonly left to the programmers are things like internationalization, user management, rights management, currency and data management, skinning management.
Using frameworks either on ColdFusion or Flex is nothing new. Creating a Framework that spans both is. Thus the idea of Flex on Fusion (FoF) came into my head and just would not let go. I thought, well, now what? Do you really want to create yet another framework even if the idea sounds cool. Who would want to learn it? and Why?
But, for some reason, the idea was persistent and would not easily go away by itself.
More thinking occurred and I do think that this framework has to enable the user (not necessarily a developer) to quickly built full applications. More like the Magic environment, more graphical, less focus on writing code.
And so I am putting it out here to gather feedback and expand on it.
The idea is not to obsolete all other frameworks, rather build a more user centric framework that can be used along with other frameworks. Thus, you should be able to use platform level frameworks to construct sub-pieces. Use the stength of each plattform, CF, for the database and logic, Flex, for advanced visualation and UI, in a combined stronger form.
I am envisioning a startup process that walks the user through the basic steps, e.g. pointing at the database and determining a subdirectory to write the model implementation, etc. This is similar to other scaffolding models; however, the application that is built will also have a user and security model. More importantly, after the basic build is finished, there should be a UI that can be used to define the application further without coding. Adding more data/screens should all be doable via UI. In the extreme thought the UI would expose an IDE for the pieces that require coding.
Once an application is generated, it can extended easily be extended while it is running by users with sufficient permissions.
I will post some more detail describing screens and workflow in a bit.