[Grok-dev] Re: an outline for the tutorial

[Martijn Faassen]

Hey,

Fernando Correa Neto wrote:
> On 3/21/07, Martijn Faassen <faassen at startifact.com> wrote:
[snip]
>> [zodb needs its own zodb.zope.org - agreed, good idea. Now we only need
>> someone to write the text and help put it up :)]
> 
> Yup. Give me the names that I will extract technical info from the
> developer's head and come back with some content.
> After that, we will need to put it in a very good shape and then put
> the site up.
> Very good shape meaning to do almost the same marketing you wrote in
> the about page. And since ZODB was already rocking for some time, I
> believe we have the use cases to show people how reliable, scalable it
> is and the companies that are using it ;). All the ones that uses zope
> :D.
> Do you think it is possible?

It's certainly possible, if someone is willing to write a bit of text 
and do the work of turning it into a small site. We can put it on the 
same site as grok.zope.org, the infrastructure is already in place. I 
can easily get a zodb.zope.org created.

Best person to start talking about this to start out would be Christian 
Theune. He can also point you to other people you might want to talk to. 
What I will do:

* help review the text.

* help you get a login into the same server that runs grok.zope.org.

* talk to the people to get the DNS set up.

[snip]
> Yup, this is the point I would say:
> How to go thru the "Convention over Configuration" process to
> integrate z3c.sqlalchemy?
> That's why I was trying to extract this info from the developers heads
> so we can just post a doctest on the list so that the core developers
> when looking at it would say:
> 
> Ok, this is compliant with grok standards and should be registered as
> a extension.
> 
> In this case, I am asking for the "Convention over Integration"
> process that would be generic enough so any average python developer
> could at least start and ask for help to finish.

This is an interesting question. I'm not sure whether any of us right 
now can formulate the answer - I'll need to think about this. There are 
technical answers that are relatively easy: use the grokker 
infrastructure of grok. To determine how things should be integrated 
into Grok is a creative process, and based on experience.

One of the problems with sqlalchemy integration is that we perhaps, as a 
group, have too little experience to know what we would really want. One 
thing we should do is talk to Brian Sutherland and some other people who 
helped develop sqlos for their opinion. Another approach is to play with 
z3c.sqlalchemy and zalchemy (to see which features each offer) and see 
whether we can write some prototypes that uses this with Grok. I imagine 
some form of integration between SQLAlchemy schemas and Zope 3 schemas 
will be important. I also imagine we need SQLAlchemy-backed containers 
(and its contents).

Anyway, just some thoughts on how you'd get this process started.

> But I know this is another story, I just want to be around when a new
> feature like zc.sourcefactory would be integrated in grok so I can try
> to identify the process and check if this process repeats for another
> packages too.

That would be very helpful. I hope Christian Theune, who knows 
zc.sourcefactory best, will sketch out some of his ideas on this.

Regards,

Martijn