Re: Re: [all] Requirements default publication

Christian Egli wrote:

>>>>a) low entry barrier
>>>>b) best practise showcase (exploring for us and learning for others)
>>>>c) show what's possible
>>>>
>>>>I guess (b) is the main point, isn't it?
>>>
>>>I think it is mainly b but could possibly be made into a.
>>
>>Then I would consider not to use the sitetree framework.
>>I know that it will be a integral part of Lenya, but I'm
>>afraid it makes the sitemaps harder to read and harder
>>to understand what's going on.
> 
> I think we'll have to define what we mean when we say sitetree
> framework. I envision something simple along the forrest site.xml
> mechanism. And I can't see the default publication without it. How can
> you autogenerate the navigation if you don't even have site.xml?
> 
> It might make the sitemaps somewhat harder to understand but the
> intent of the default publication (in the sense of (b)) is not to
> *completely* understand all the intricacies of the publication but to
> learn how to extend and adapt it to fit your needs.

OK, I see. We provide a general entry point (cocoon:// URI) to generate
the navigation so the publication does not have to care about that.
If this is a simple and easy-to-understand interface, it does not
complicate the learning process.

We should try to keep all interfaces (in Java and in the sitemaps)
- as small as possible
- self-explanatory
- well-documented

When we call a cocoon:// URI that leaves the publication sitemap,
I would suggest to add a comment where it goes to, like
<!-- fetches the navigation from /lenya/sitemap.xmap -->

I noticed that tracing sitemap paths can become quite complicated.


>>>>I'm still against the XHTML doctype if this shall be a
>>>>best-practise publication. I agree that this is useful to
>>>>integrate existing content. But maybe we should point to the
>>>>drawbacks of this approach?
>>>
>>>Sure we can. It just seems that the XHTML doctype is a very pragmatic
>>>approach which can simplify things.
>>
>>It only simplifies the initial effort and editing with conventional
>>editors. IMHO everything else is more complicated.
> 
> I think people want the XHTML doctype, and I think if we ever get the
> document describing the xhtml method
> (http://cvs.wyona.org/cgi-bin/bugzilla/show_bug.cgi?id=89) this could
> make for a nice and simple example.

In my eyes, XHTML is a presentation format, not a content format.
I would use it for templates and for "non-standard" pages like
homepages. And the same would I recommend to anyone who is creating
a site consisting of >3 pages.
I'll stay a hardliner regarding this issue :)

>>>No, the intent is to add pages to the navigation as they are added to
>>>the publication. I know this has problems when you start adding
>>>hundreds of pages but again KISS.
>>
>>How can you define the order of navigation items?
> 
> You can't. It's as simple as that :-).

OK - agreed, if it scales (see below). Otherwise, people might
be disappointed soon and we can't show this feature on presentations
etc.

> The basic mechanism of the sitetree mechanism definitively has to
> scale to other methods of editing.

OK, that sounds good.

Andreas