[pmwiki-users] Group & subgroup behavior

Patrick R. Michaud pmichaud at pobox.com
Fri Aug 18 12:33:26 CDT 2006


On Fri, Aug 18, 2006 at 04:39:57PM +0200, Joachim Durchholz wrote:
> Patrick R. Michaud schrieb:
> > The biggest negative I see with the reiser4-inspired proposal
> > is the need for the colon to a sibling link.
> 
> One downside is unavoidable here.
> 
> Either the current page is also the current group, in which case 
> accessing child pages will require a prefix.
> Or the parent page is also the current group, in which case accessing 
> sibling pages will require a prefix.

Or, we don't have hierarchical groups.  :-)

> Pick your evil ;-)

So far, my choice of "least evil" has been "no hierarchical groups".  :-)

But if we did have hierarchical groups, then I definitely come down
on the side of "accessing child pages requires a prefix".

>  > It seems to me
> > that sibling links are by far the most common, and so requiring
> > a colon on each one optimizes markup the wrong way.
> 
> Hmm... Links to child pages could well become common. Particularly if 
> group pages are to contain some sort of directory of their subpages.

Yes, but group pages are probably the minority.  Most pages are leaves --
i.e., they don't have children.

> > It also goes against the natural inclination of most filesystems,
> > in which a bare name "B" within file "A" refers to a file named
> > "B" in the same folder as "A", and not to a child of "A".
> 
> OK, now *that's* an argument that helps sort out intuition.
> 
> Seems like we have a consensus of what's to be considered the "current 
> directory/group": it's the parent of the current page.
> 
> Unless I read any serious objections, I'll update the Proposals page to 
> nail that down and give link syntax designers a firm ground to start from.

Sounds good.

Pm




More information about the pmwiki-users mailing list