RE: [collections][pattern] Predicate, Transformer, Closure/Comman d, Factory

["Jack, Paul" <pj...@sfaf.org>]

Well, in theory I'm not against [collections] depending 
on [pattern].  But I think it'd have to happen after the
2.1 collections release -- unfortunate, as we'd then be
adding an interface to collections we intend to deprecate
later.

But there are projects that are blocking on a new release
of Collections, and since [pattern] is still in the sandbox,
I wouldn't want those projects to wait until [pattern]
goes through all the promotion steps.  I also wouldn't want
to rush the promotion of [pattern], to make sure it has ample
time to gestate.

So my thinking is:

collections 2.1 -- no pattern dependency
collections 2.x -- pattern dependency

-Paul


> -----Original Message-----
> From: Stephen Colebourne [mailto:scolebourne@btopenworld.com]
> Sent: Thursday, August 08, 2002 3:25 PM
> To: Jakarta Commons Developers List
> Subject: [collections][pattern] Predicate, Transformer, 
> Closure/Command,
> Factory
> 
> 
> As some of you are probably aware, I have started the 
> [pattern] sandbox
> project to gather together generic reusable interfaces. Perhaps not
> surprisingly the first ones I chose came from [collections]. 
> Thus [pattern]
> has Predicate, Transformer, Command and Factory (plus Immutable and
> Identifiable).
> 
> The first 4 from directly onto [collections] interfaces Predicate,
> Transformer, Closure and SimpleObjectFactory. However, the versions in
> [pattern] all have large Utils classes associated and a 
> consistent exception
> handling mechanism.
> 
> I would like to see [collections] depend on [patterns] for these four
> interfaces. This could be achieved by deprecating the versions in
> [collections] and making those extend the [pattern] 
> equivalent. If this is
> done before the next release, then SimpleObjectFactory can be deleted
> entirely (as it was added after 2.0). (If the 2.1 release 
> must go out I
> suggest renaming it to Factory)
> 
> But of course this is controversial, [pattern] is new and still in the
> sandbox. And [collections] currently has no dependencies, 
> this would add
> [pattern] and [lang].
> 
> On the plus side, [collections] would gain a much wider set 
> of Utils for all
> the interfaces than is currently available. And it is a more sensible
> location for these interfaces, for example the new 
> [introspect] will use the
> [pattern] version when I get round to writing it ;-)
> 
> This thread is to gather opinion on whether [collections] are 
> happy to use
> [pattern] in principle -  soon, sometime or never ?
> 
> Stephen
> 
> 
> --
> To unsubscribe, e-mail:   
<ma...@jakarta.apache.org>
For additional commands, e-mail:
<ma...@jakarta.apache.org>

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>

Re: [collections][pattern] Predicate, Transformer, Closure/Comman d, Factory

[Stephen Colebourne <sc...@btopenworld.com>]

I am happy with these answers. Its what I expected :-)

What would help for a future integration would be the following changes to
[collections]:

- Rename SimpleObjectFactory to Factory
- Remove PredicateUtils (the part that returns predicates) - although useful
to collections, it is not core collections functionality, better to not have
it at all.

This will minimise problems for the future in my eyes.

Stephen

From: "Michael A. Smith" <ma...@apache.org>
> On Thu, 8 Aug 2002, Jack, Paul wrote:
> > Well, in theory I'm not against [collections] depending
> > on [pattern].  But I think it'd have to happen after the
> > 2.1 collections release -- unfortunate, as we'd then be
> > adding an interface to collections we intend to deprecate
> > later.
> >
> > But there are projects that are blocking on a new release
> > of Collections, and since [pattern] is still in the sandbox,
> > I wouldn't want those projects to wait until [pattern]
> > goes through all the promotion steps.  I also wouldn't want
> > to rush the promotion of [pattern], to make sure it has ample
> > time to gestate.
>
> I agree with Paul here...
>
> > So my thinking is:
> >
> > collections 2.1 -- no pattern dependency
> > collections 2.x -- pattern dependency
>
> collections 3.0 more like it.  Adding a dependency is a non-backwards
> compatible change (at least in my mind).
>
> michael
>
>
> --
> To unsubscribe, e-mail:
<ma...@jakarta.apache.org>
> For additional commands, e-mail:
<ma...@jakarta.apache.org>
>


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>

RE: [collections][pattern] Predicate, Transformer, Closure/Comman d, Factory

["Michael A. Smith" <ma...@apache.org>]

On Thu, 8 Aug 2002, Jack, Paul wrote:
> Well, in theory I'm not against [collections] depending 
> on [pattern].  But I think it'd have to happen after the
> 2.1 collections release -- unfortunate, as we'd then be
> adding an interface to collections we intend to deprecate
> later.
>
> But there are projects that are blocking on a new release
> of Collections, and since [pattern] is still in the sandbox,
> I wouldn't want those projects to wait until [pattern]
> goes through all the promotion steps.  I also wouldn't want
> to rush the promotion of [pattern], to make sure it has ample
> time to gestate.

I agree with Paul here...

> So my thinking is:
> 
> collections 2.1 -- no pattern dependency
> collections 2.x -- pattern dependency

collections 3.0 more like it.  Adding a dependency is a non-backwards 
compatible change (at least in my mind).  

michael


--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>