[Catalyst] RCF for a Catalyst Wiki Controller

Matt S Trout dbix-class at trout.me.uk
Mon May 22 14:35:29 CEST 2006


Kieren Diment wrote:
> Wiki markup:  Kwid - it's like pod but easier to write.  Should parse to 
> the same kind of tree. It's also not quite fully implememted yet, so 
> there's the possibility to influence the development of the final standard.
> 
> It would be useful also to be able to use different stores other than 
> RDBMS easily.  The filesystem is good quick and dirty modelling of 
> heirachical information (e.g. biological classification schemes is a 
> classic heirachical info store), so making the orm optional or taking 
> some steps to make it easily swapped with a different model out would be 
> good in my opinion.
Absolutely agreed, I'm currently experimenting with a double-layer MVC 
approach where I have one MVC for the domain model and the app itself is 
a second MVC whose model merely interacts with domain view and 
controller objects (for those of you following along the "stealing from 
Jifty" part, this is not unlike the Jifty action approach although Jesse 
and I can probably claim "parallel invention" rights given we've been 
discussing this sort of stuff for close to a year now).

The nice thing about this approach is that provided you present the 
appropriate domain view/controller interfaces you can swap out the 
underlying implementation trivially since your DBIx::Class or similar 
objects never leak outside to the UI/web layer.



More information about the Catalyst mailing list