Just a few notes on what the purpose of the Attributes Proof-of-Concept
is, as I see it.
We Have a Problem
-----------------
Webapps have their web.xml, EJBs have their ejb-jar.xml, JBoss has
jboss.xml, EARs have God-knows-what. It seems like every applicaiton or
domain where metadata is needed have invented its own package format.
Avalon is no exception. This is just plainly confusing.
Enter Attributes
----------------
Attributes promise a way to get rid of most - if not all - of these
packaging formats, and in addition, bring the metadata closer to the
objects it describes. The usefulness of this approach is proven by the use
XDoclet has seen, and by the use of custom tags in Avalon - who wants to
go back to writing the metadata by hand?
C# already supports attributes, and Java 1.5 will.
Goals
-----
The goal of this proof of concept could be summed up as:
"How far can we get by *only* using attributes?"
Some things are theoretically possible with attributes, but practically
hard or impossible. Some things are both theoretically and practically
very possible to do. In order for us to incorporate attributes, when they
come in Java 1.5, we must know the tradeoffs. And that's what this is
about.
If we can solve every requirement with attributes, then *all* our
packaging requirements disappear with Java 1.5. No custom build tools
required, just compile as usual, jar as usual, and there you go.
However, I don't think we'll ever reach that stage: How do you express
component configuration as attributes, or a Phoenix block.xml?
But how far can we get? I want to know that, and I want to know *why* we
can't get any further. Once we know the tradeoffs, we can decide how and
what to use attributes for.
Non-Goals
---------
My goal is specifically *not* to get attributes adopted by the Avalon
community. This isn't "my pet idea" that I'll want to get into the core
Avalon architecture at any price. I think attributes are a great idea, and
I think much of our problems can be solved with them, but I'm not going to
do any pushing to get them into Avalon.
Request
-------
I'd like you all to snap your brains into "attribute mode": When all you
have is a hammer, all problems look like nails. Well, now the only tools
you have are attributes. Let's see how far they can get us. Then, once we
have brainstormed enough, let's take a step back and see what the cost of
our attribute-based solutions are in comparison to the non-attribute based
solutions.
/LS
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
