Re: Scope of step types

On 10/2/07, Norman Walsh <ndw@nwalsh.com> wrote:
> / Norman Walsh <ndw@nwalsh.com> was heard to say:
> | I think the only negative consequence of this decision is that two
> | pipelines in a pipeline library that each import a different
> | definition for a step type will cause an error where one is not,
> | perhaps absolutely necessary. But we have namespaces to make such
> | things unnecessary, right?
>
> Oh, it also makes managing the set of step types visible to a pipeline
> a little trickier.

Isn't the question of the visibility of "a" somewhat independent from
the question of whether you are allowed to have two different "a" in
the system?

> If pipeline "a" imports pipeline "b" and pipeline "b" imports "c";
> should "c" be available directly from within "a"?

It makes sense to make names global. Especially since we have
namespaces. This is similar to Java where you can have only one class
org.example.Foo at one point in time (ignoring class loaders for the
sake of the argument). But of course, it doesn't mean that
org.example.Foo will be visible to everyone. In particular if
org.example.Foo imports org.example.Bar, and you import
org.example.Foo, you should not see org.example.Bar.

Alex
-- 
Orbeon Forms - Web 2.0 Forms, open-source, for the Enterprise
http://www.orbeon.com/

Received on Wednesday, 3 October 2007 01:38:01 UTC