[Dbix-class] Using Class::Std under-the-hood

[Darren Duncan]

At 12:05 AM -0700 8/11/05, Dan Kubb wrote:
>The truth is that if we use something other than blessed hash objects
>there is going to be some overhead.  If we are concerned about optimizing
>attribute lookups we should look into an array based object instead.

I think that array-based objects are the worst kind you can have, and 
a maintenance nightmare.  I believe Damian says so too.

>Will there be any appreciable difference in performance either way?  I
>doubt it, especially when compared with the other operations being performed
>by DBIx::Class.

My priority is to do the job right, and not sacrifice this for speed. 
You can gain speed in other ways, such as caching the result of 
expensive operations; eg, caching generated SQL strings (which 
Rosetta::Engine::Generic does).

-- Darren Duncan