ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "James Cook" <jimc...@iname.com>
Subject Re: Anteater... I'm Baaaack...
Date Thu, 21 Dec 2000 13:52:17 GMT
----- Original Message -----
From: "James Duncan Davidson" <duncan@x180.net>
> That's going down the road of excessive generalization. Projects that do
> that take a *long* time to ship.

It's sometimes difficult to keep my tongue in check. I think given your
position, some people would read this post and actually believe it. I would
appreciate a debate of the technical merits on *any* of the proposals, but so
far all you have contributed is FUD.

Just for a point of comparison, the code that comes with the frANTic proposal
took me about 8 hours to write. In that 8 hours, I defined a Task interface, 5
simple Task implementations, a properties system that is fully scoped at
runtime, a node naming convention similar to typical OS directory naming, full
dependency support, complex parameter (nested parameters, et al.) resolution,
and an efficient execution engine.

I will state the obvious that this doesn't mean Ant core is "ready to ship", but
it should cast a serious shroud of doubt on your over-simplification. I think
you should remove the word "excessive" from "excessive generalization", as my
proposal doesn't attempt to introduce "excessive" generalization. I didn't
intoduce a "command execution engine", but rather a "task execution engine".
Task is not just a simple one-method execute() function, it is a robust
definition in its own right.

jim


Mime
View raw message