From: "Hal E. Fulton" <hal9000@...>
Date: 2001-11-02T05:04:04+09:00
Subject: [ruby-talk:24111] Modularity question

Opinion question for you all.

Not *purely* Ruby-related, but I am coding
in Ruby and will especially appreciate any
Ruby-like tricks that I may not have thought of...

I'm about to start coding an app. Will let you
know more when it's finished.  :)

Besides the "main logic" it will have a component 
that manipulates external data and a component
that handles the interface.

I want these to be as discrete from each other as
possible. It's a "separation of concerns" issue (and
a relatively easy one, no AOP involved, I think).

For one thing, I want to allow at least two different
UIs (graphical (Tk?) and text-based), and I want the 
possibility of changing the graphical part later
without too much difficulty -- maybe even adding 
a web interface or some other.

As for the database part, I'm uncertain as to whether
I want to use simple marshaling or DB or even MySQL.
But I don't want to be "tied" to however I first implement
this.

My first thought is to have a very high-level interface that
simply handles things in a way utterly transparent to the
main logic -- using calls that mean "ask the user for a 
yes or no response" or "get the object that corresponds
to this name or key."

I could then (for the UI) do a different "require" based on
a command line parameter or whatever.

But I can't help feeling that I'm overlooking a chance to
Do Something the Right Way.

Anyone have opinions on this?

Hal