[Openinteract-dev] Re: ESPOPS, the future, and CPAN

[Ray Z. <rz...@co...>]

At 10:23 AM -0400 9/13/02, Chris Winters wrote:
>I have no problems with it. I apologize for not including it in the
>eons since you announced it, but I haven't had that particular itch
>yet :-)

I know how that goes ... I have (non-Perl) code that I've been 
"almost ready" to release for several years now.

>As for naming, that's up to you. SPOPS isn't an official namespace in
>the CPAN sense, but it might make sense for people looking around for
>the first time. I don't want to cause you renaming problems, however.
>
>I don't have any ideas for names. IIRC I was going to call this
>SPOPS::Inheritable, but that's not quite there I think.

You mean you don't quite like the name? Or that ESPOPS doesn't quite 
live up to that name yet?

>What does your
>team call it?

We just called it ESPOPS, but I figured I should give the name a bit 
more thought if I'm going to put it on CPAN.  What about doing 
something like was done with DBI, where there is also a DBIx 
namespace for things build on top of DBI. What do you think of 
SPOPSx::Inheritable?

>  > On a related note ... do you have an up-to-date SPOPS to do list somewhere?
>
>Not explicitly. My current plan is:
>
>  - Finish writing tests so that every major class in the system is
>  represented. (Almost done with this.) This is broader than it sounds
>  because various things shake out as a result. (As is happing right
>  now with security stuff.)

Isn't it amazing the number of bugs that can hide in code forever if 
you don't have comprehensive tests?

>  - Convert the rules, and make class generation behaviors, to use
>Class::Observable instead of the built-in scheme.

BTW, this seems like a reasonable idea to me. We've been using this 
design pattern along with ESPOPS objects and the only real stickler 
we ran into was circular references. We had an object Foo which has-a 
field containing a FooState object. When we made Foo and observer of 
FooState we got a circular reference. If you run into the same issue 
and are interested, I'll be happy to show you how we got around it.

>  - After that is working, I want to finally change the
>  linking/relationship declarations and behavior as you outlined.
>
>After that... it's time for 1.0 I think :-)

Yay!

-- 
  Ray Zimmerman  / e-mail: rz...@co... / 428-B Phillips Hall
   Sr Research  /   phone: (607) 255-9645  /  Cornell University
    Associate  /      FAX: (815) 377-3932 /   Ithaca, NY  14853