[Zope-dev] Re: Interest in "AdvancedQuery" and/or "ManagableIndex"?

[Philipp von Weitershausen]

Dieter Maurer wrote:
> Whit ("mailto:d.w.morriss at gmail.com") reported that "AdvancedQuery"
> is going to ship with Plone3 and that packaging would be easier for them if
> "AdvancedQuery" were part of the Zope 2 distribution.

I fail to find an explanation *why* that is.

> According to Whit, Alexander Limi seems to be interested to have
> "Managable Index" in the Zope 2 distribution, as well.

Again, why?

> I have no problems to donate "AdvancedQuery" and/or "Managable Index"
> to the Zope Foundation *BUT* I will not modify the code to bring
> it in line with the different style requirements usually applied
> to Zope components: e.g.
> 
>   * my code uses 2 blank indentation rather than the usual 4 blank
>     (to make it more readable and easier to maintain for me)
>   
>   * I much prefer unit tests over doctests; thus, "AdvancedQuery"
>     and "Managable Index" come with extensive unit tests and no
>     doctests
> 
>   * I use camel case also for parameters and local variables
>     and not only for functions and "global" objects.
> 
> 
> Is there interest in "AdvancedQuery" and/or "Managable Index"
> to become part of the Zope 2 distribution under these conditions?

With that many +1 votes already, it might seem kinda pointless, but I'm 
-1. It's not a strong -1 but nevertheless a -1.

Here's *why*:

* I think Dieter is doing an excellent job at maintaining his products, 
why should that maintenance now become a burden of the Zope 2 
maintainers? Why burden, you might ask? Because it's an "alien" product 
that is all of a sudden adopted into Zope. I'd think most of us aren't 
familiar with its code base, let alone being able to do bugfixes. If we 
adopt it into Zope core, we're bound to maintaining it forever, no 
matter what Dieter decides to do...

* As Dieter outlines above, he prefers a few different styles. Code in 
svn.zope.org and *especially* code in Zope core should conform to a 
certain style. I'm not necessarily talking about doctests vs. unit 
tests, but a certain kind of code arrangement (and that includes 
developer-orienteded docs such as doctests nowadays) are appreciated. 
Why should we go thru the hassle of giving AdvancedQuery a do-over? I 
don't think anybody would appreciate all that work nor would Dieter 
probably appreciate his codebase to be changed (well, he *is* willing to 
donate this thing, so I'm sure he's ok with taking that risk, but 
still... is it really necessary?).

* Plone is shipping with lots of products already, I don't see why it 
simply can't ship with another one. Seriously, why? Plus, if they're 
really trying to solve problems for Plone 3, then it seems to be too 
late already. The target platform, Zope 2.10, is feature-frozen since long.

If this is all about Zope 2.11 and future release of Zope, I would like 
to direct your attention to one of my recent proposals, 
http://wiki.zope.org/zope2/EggifyingZopesExtensionMechanismQuotProductsQuot, 
which I have plans for implementing in Zope 2.11. This propsal will 
allow us to package products like Dieter's excellent ones as Python Eggs 
and deploy them like any other package. So, if Plone really didn't want 
to ship with AdvancedQuery et. al., they could simply use the 
setuptools' dependency mechanism to at least depend on that egg. Hanno 
Schlichting has shown how powerful zc.buildout can be in this area with 
'ploneout' already.


It might not be a popular opinion in Zope 2 land where people would like 
to have as much working out of the box as possible, but I think we 
oughta think about making Zope 2 rather smaller than bigger (where 
"smaller" doesn't mean "ship with less stuff" but "have fewer 
inter-dependencies" so that it's better reusable). An Egg-based 
deployment mechanism with explicitly defined dependencies will allow us 
to do so.


-- 
http://worldcookery.com -- Professional Zope documentation and training
Next Zope 3 training at Camp5: http://trizpug.org/boot-camp/camp5