[Opensource] servlet vs. other potential expresso containers

[Michael Rimov]

At 01:22 PM 6/1/2002 -0700, you wrote:
>hi,
>
>Expresso is currently designed to NOT assume that its containing engine is 
>a servlet engine.
>
>However, in many places, this obscures access to common features.  For 
>example, casting is done to get at the servlet request in order to handle 
>multiple params with same name.
>
>Generality is a nice in a framework, but so is easy access to 
>commonly-used features.
>
>So what is the 80% case here?  How many people use Expresso in a servlet 
>engine versus some other engine?
>
>If we know that the overwhelming majority of users are deploying in a 
>servlet engine, shall we optimize for that, specializing instead of 
>generalizing?
>
>My guess is that the tradeoff favors specialization in servlets.

At this point, I think that I favor non-specialization except where 
necessary... why?  Because I think we have some future capabilities that 
we'll sacrifice otherwise.  Examples are:

1 - ControllerRun command line I think is far more capable than say Jobs 
when combined with UNIX cron.

2 - WebServices is an area where the Controller Paradigm will reign 
king.  Although the API may change somewhat, it's a definite need.

As far as features like accessing multiple parameters... I think a way to 
allow this in the main Controller API is worthwhile.  It's a situation a 
lot of people need that was overlooked.

So on the one hand yes... 80% of the people don't use it in 
Servlets.  However, I'd still like to see if we can manage certain aspects 
without sacrificing that separation.

Just my thoughts! ;)
                                                 -Mike