[Grok-dev] Re: newbie question: megrok.five status

[Philipp von Weitershausen]

Please stay on the list... CC'ing grok-dev again.

On 13 Nov 2007, at 11:59 , kevin gill wrote:
>>> I had a look in http://svn.zope.org, and it is not listed (it is  
>>> listed
>>> if you give a revision: http://svn.zope.org/?rev=78965). I assume  
>>> that this
>>> means it is either moved, merged or deleted ( I am not a  
>>> subversion user ).
>>
>> I've moved it to my Sandbox since it was an experiment and never
>> considered for anything real.
>>
>
> Out of interest, does the megrok.five approach that you experimented  
> with
> require the programmer to write a Zope2 product that uses the Grok
> features, or does it work as a clean Grok style package. From the  
> readme,
> it seems to be very clean, but the Five examples that I have found  
> (and
> there are very few) include the Zope2 product stuff and the Zope3  
> stuff.

It doesn't really make a practical difference whether you write an old- 
school Product or a Python package, except that with a Product, you  
get a few things for free:

* configure.zcml, meta.zcml, overrides.zcml are automatically loaded
* your package is automatically registered in the "Products" registry

You can achieve the same things with regular Python packages, you'll  
just have to do it manually in two lines of ZCML.

>>> I also read in the IRC logs (21/9/2007):
>>>
>>> philiKON	timte: megro.five is dead	16:59
>>> philiKON	timte: but i have plans for something else	16:59
>>>
>
> I assume that philiKON is referring to the same strategy as Phillip
> outlines below.

I would think so, since that's me :)

>>> There are a number of posts on zope-dev recommending using Grok  
>>> instead
>>> of ZClasses. Am I correct to assume that this is still the  
>>> position of Grok
>>> project or is Grok?
>>
>> Well, I think back then we were suggesting to use Grok instead of
>> Zope2+ZClasses. We weren't suggesting to use "Zope2+Grok" (if such a
>> thing existed).
>>
>
> Ahh.. that is an unfortunate distinction. ZClasses have an bad  
> educational
> effect; they are very easy to learn and use, they build on Zope2 TTW
> concepts very cleanly, there is a tool to convert them to Zope2  
> Products,

I'm not a ware of such a tool.

> but they push you away from Zope3 Component technologies. Grok could  
> fill
> a similar niche, but they lead into Zope3 Component Technologies.

I'm not sure what the point is you're trying to make. We don't need  
Grok bolted on Zope 2 to fill that niche. If people want Grok for what  
it is, they shouldn't care that it uses Zope3 underneath and not Zope 2.

>>> Should I continue to investigate Grok for a Zope2 site?
>>
>> I think the Grok mechanism and approach would be great for Zope 2
>> applications. However, I don't think it should and will look exactly
>> like Grok. That said, we're working towards making it possible to use
>> Grok's mechanisms for registering various types of components
>> independently of the Grok web application framework itself. Then  
>> people
>> can work on integrating those into a "Grok for Zope 2" or "Grok for
>> Plone" or whatever (they should have different names, obviously).
>>
>
> That would be really useful. Is there any timescales?

Nope.