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

[Robert P. J. Day]

On Mon, 18 Oct 2004 chr at home.se wrote:

> IMHO, a group is an abstract grouping of pages, for instance something
> that defines the scope of a namespace. In other words, there can only be
> one page called 'A' within a group.

yup, that's the way i'd look at it.  a group is *not* a page -- it
*contains* pages.  just like a directory *contains* files, but
wouldn't *itself* be considered a file within the directory.

> From that point of view, a group might be considered the children of
> a page.

hmmm ... not sure what this is trying to say.

> Anyway, I'm just trying to illustrate that 'group' hasn't really been
> defined yet. If we for instance decided that a page called 'A' may not
> have a child page called 'A', then we could define a group as a page
> together with its child pages.

as i mentioned, i don't think a group should be considered a page.  it
should be an abstract identifier (possibly multi-level) that
*contains* pages (or, possibly, other groups).

> I agree that introducing some kind of distance metric is probably a good
> idea.

i don't think a distance metric would be of much value since the
metric itself could change drastically as you decide to move groups to
other parts of the wiki.

i think what i just described might just work as it involves, not a
metric, but the idea of how quickly you can resolve a page name if
you have to work your way up the tree, looking for the first match on
the first component of the group name.

rday