From: James Britt <jamesUNDERBARb@...>
Date: 2004-05-15T04:58:11+09:00
Subject: Re: Please revisit  Ruby's Best and vote for additions

Curt Hibbs wrote:
> 
> Another part of Why Ruby is provide advice to people who are advocating
> Ruby in their workplace. So, what I'd like to know is when you get the
> pushback "Nobody knows it.  I can't get people to support it.", what is
> your response?

I don't have any good answer, because the short answer leads to more 
questions, for which I don't know what to say.  If I tell management 
that any developer with more than half a brain and some OO experience 
can pick up Ruby in a few days (at least learn enough to read code, 
debug, maintain, etc.), the question turns to cost.  Why spend any money 
to retrain people when J2EE/JSP script-monkeys are a dime a dozen?  Is 
Ruby going to be cheaper in the long run? Will the resulting code be as 
or more: stable ,fast, scalable, secure??  Where's the competitive 
advantage?  Is it better to hire a local team of fewer-but-smarter 
developers rather than outsource to someplace 12 timezones away for half 
the cost?

For many IT managers, software and programmers are a commodity. 
Management would rather pay a small fortune on licensing fees for 
something like WebLogic or iPlanet, and hire off-the-rack Javadroids, 
the sort of folks who can code EJB remote interfaces in their sleep, but 
can't tell you how to pick a sort routine or a data structure best 
suited for a given task.

The "standardization" on J2EE means that developers don't have to know 
any actual computer science or software engineering, they just have to 
know a particular API.  It's like paint-by-numbers, or a filling in a 
coloring book; Java has become the Crayola of programming languages.

It's a tough sell.  When I was helping interview developers, myself and 
another developer were more interested in how a candidate thought, if 
they at least knew about certain things and had an aptitude for 
learning.  But management was very concerned about actual experience; 
they didn't want anyone to spend time *learning* Ant or JUnit, they 
wanted someone plug-and-play.


>>We need to demonstrate that Ruby easier to develop, easier to 
>>maintain, and solves the same business needs (including concerns over 
>>speed, security, scalability, and reliability).
>>
>>Perhaps Ruby needs the equivalent of the "Pet Store" example.
> 
> 
> Actually, what we really need is a simple spec. for something like a
> "Pet Store" example, and then implement that in (say) Ruby, Java, C#,
> and Python, each by someone skilled in that language and its development
> environment, and keep track of some basic product metrics in a uniform
> way.

Well, there's the J2EE Pet Store as the canonical example.  I believe 
Microsoft did a version in C#/.net.  So consider that as the baseline; 
The implementation is the spec.

It helps if a Ruby version goes toe-to-toe with Java or C# version, but 
I think the key is a demonstration that you can build the same sort of 
common business apps in Ruby, and show that they're just as fast and 
scalable but easier to develop and maintain.

James