What is the status on Zend_Model?

[Ralf Eggert]

Hi there,

I remember the discussion about the model part in the Zend Framework
many moons ago. I also remember a few approaches discussed on these
issue, e.g. providing a set of interfaces or abstract classes to build a
model then "Zend Framework way" or adding a chapter to the reference
guide with best practices about writing models for the Zend Framework.
And I also remember that this was planned for the 1.9 release.

So, are there any news about Zend_Model or whatsoever it will be called?

Thanks and best regards,

Ralf

[weierophinney]

-- Ralf Eggert <r.eggert@...> wrote
(on Wednesday, 01 July 2009, 08:43 AM +0200):
> I remember the discussion about the model part in the Zend Framework
> many moons ago. I also remember a few approaches discussed on these
> issue, e.g. providing a set of interfaces or abstract classes to build a
> model then "Zend Framework way" or adding a chapter to the reference
> guide with best practices about writing models for the Zend Framework.
> And I also remember that this was planned for the 1.9 release.
>
> So, are there any news about Zend_Model or whatsoever it will be called?

Not currently, though I'm working currently on some ideas and
initiatives that may help address this in the future (though not the 1.9
timeframe).

The problem with providing interfaces and/or abstract classes for
"models" at this point is that there are *so* many ways to approach
models: domain models, data mappers, gateways, transaction scripts,
service layers, ActiveRecord, table/row data gateways,
subject/observer... the list goes on and on. Many patterns can be used
in combination (e.g., I used a combination of service layer, domain
model, data mapper, and TDG for my ZF workshop at DPC this year), yet
still work effectively in a standalone way. Additionally, for patterns
such as domain models... just having a class is often enough of an
interface. We could have interfaces for each pattern -- but that starts
to look like overkill, and also limits the actual implementations (as
many patterns can have a variety of implementations).

What will most likely happen is that we will include a chapter detailing
several approaches to model creation -- but still leave the actual act
of creating models to the developer.

--
Matthew Weier O'Phinney
Project Lead            | matthew@...
Zend Framework          | http://framework.zend.com/

[funkyfly]

Would be really nice to have this chapter.

--
Vladas Diržys

On Wed, Jul 1, 2009 at 13:41, Matthew Weier O'Phinney <matthew@...> wrote:

-- Ralf Eggert <r.eggert@...> wrote
(on Wednesday, 01 July 2009, 08:43 AM +0200):

> I remember the discussion about the model part in the Zend Framework
> many moons ago. I also remember a few approaches discussed on these
> issue, e.g. providing a set of interfaces or abstract classes to build a
> model then "Zend Framework way" or adding a chapter to the reference
> guide with best practices about writing models for the Zend Framework.
> And I also remember that this was planned for the 1.9 release.
>
> So, are there any news about Zend_Model or whatsoever it will be called?

Not currently, though I'm working currently on some ideas and
initiatives that may help address this in the future (though not the 1.9
timeframe).

The problem with providing interfaces and/or abstract classes for
"models" at this point is that there are *so* many ways to approach
models: domain models, data mappers, gateways, transaction scripts,
service layers, ActiveRecord, table/row data gateways,
subject/observer... the list goes on and on. Many patterns can be used
in combination (e.g., I used a combination of service layer, domain
model, data mapper, and TDG for my ZF workshop at DPC this year), yet
still work effectively in a standalone way. Additionally, for patterns
such as domain models... just having a class is often enough of an
interface. We could have interfaces for each pattern -- but that starts
to look like overkill, and also limits the actual implementations (as
many patterns can have a variety of implementations).

What will most likely happen is that we will include a chapter detailing
several approaches to model creation -- but still leave the actual act
of creating models to the developer.

--
Matthew Weier O'Phinney
Project Lead            | matthew@...
Zend Framework          | http://framework.zend.com/

[Ralf Eggert]

Hi Matthew,

thanks for your reply.

> What will most likely happen is that we will include a chapter detailing
> several approaches to model creation -- but still leave the actual act
> of creating models to the developer.

That what be my favorite solution as well. Just because of the many
solutions for implementing models. This chapter should have some general
information about implementing models and could even discuss some of the
most common design patterns to implement models. It could even give some
hints how to include third-party solutions for implementing models.

Really looking forward to it.

Best regards,

Ralf

[Martin Martinov-2]

2009/7/2 Ralf Eggert <r.eggert@...>:

> Hi Matthew,
>
> thanks for your reply.
>
>> What will most likely happen is that we will include a chapter detailing
>> several approaches to model creation -- but still leave the actual act
>> of creating models to the developer.
>
> That what be my favorite solution as well. Just because of the many
> solutions for implementing models. This chapter should have some general
> information about implementing models and could even discuss some of the
> most common design patterns to implement models. It could even give some
> hints how to include third-party solutions for implementing models.
>
> Really looking forward to it.
>
> Best regards,
>
> Ralf
>
>

And maybe Zend_Tool can help here, by providers generating the class
skeletons for some chosen design patterns?


--
Regards,
Martin Martinov
http://mmartinov.com/