[Pmwiki-users] Re: can pmwiki handle hierarchical content?

[chr@home.se]

On Fri, 15 Oct 2004, Stefan Candea wrote:

> I know there was a ton of discussions on this.  However, can someone
> explain simply (what did it come down to?) why having pages as files and
> groups as subdirectories was a no-no from the begining?

I'm at least partly to blame for not doing an analysis as I've promised to 
do (lack of time etc) - all the posts are still in my folder though, so 
they give me a bad conscience now and then.

What follows is straight from my memory, so don't expect it to be 
complete. IIRC, the problem is two-fold:
* Coming up with a reasonable syntax
* Making a good default choice for "page paths"

For instance, assume that you currently looking at a page called 'C' with
the "path" A/B/C, and this page contains a link that only gives you the
name of the page as "D" (i.e. a relative path). Where should this link go
to if you have several pages called "D", as in the following page tree:

	A
	|-- B		'B' is a "parent" of 'C'
	|   |-- C	<- Current page
	|   |	'-- D	'D' is a "child" of 'C'
	|   '-- D	'D' is a "sibling" to 'C'
	'-- D		'D' is a "uncle" of 'C'

And we can make the choice even more difficult if the parent also happens 
to be called 'D', i.e.

	A
	|-- D		'B' is a "parent" of 'C'
	|   |-- C	<- Current page
	|   |	'-- D	'D' is a "child" of 'C'
	|   '-- D	'D' is a "sibling" to 'C'
	'-- D		'D' is a "uncle" of 'C'


A similar problem can happen when a page should be created because it
doesn't exist yet. I.e. the page you're currently looking at contains
NewPage, but no such page can be found, so you have a "create"-link. But
where should this new page be created? In the current group, or should it
be created as a subpage?

> LinuxHardwareSoundCardsNameofcardDriverVersion1 style gets a little long
> to be referenced in a page ...

I agree...

It might be helpful to have some "real" examples of page trees, so that
maybe you could more easily figure out where you would want links to point
to. (when the "path" isn't absolute or explicit enough).

/Christian

-- 
Christian Ridderstr?m, +46-8-768 39 44               http://www.md.kth.se/~chr