Re: [emacs-wiki-discuss] Thoughts about Reverse Linking

[A.J. Rossini]

On 10/28/05, Jim Ottaway <address@hidden> wrote:
> >>>>> Bastien Guerry <address@hidden> writes:
> > Ok guys,
>
> > i'm starting to code something like planner-registry.el.
>
> > * from any document (gnus, w3m, bibtex), display a list of the
> >   relevant planner pages
>
> > * visiting an item of this list opens the page and set the point
> >   properly (at the beginning of the link)
>
> > * implement `planner-registry-initialize'
>
> > * implement `planner-registry-update' (as a hook when saving)
>
> > * store data in alist of alist like:
>
> >    '(urls ([id] ([page] [point]))
> >      messages ([id] ([page] [point]))
> >      bookmarks ([id] ([page] [point]))
> >      bitex-entries ([id] ([page] [point])))
>
> > Does it sound good? Please send me any relevant idea.
>
> I would just use the url [i.e., any kind of muse/planner url] as the
> key: I can't imagine when one would want to have the list organized by
> messages, bookmarks, etc.; if one did, it would be easy to extract
> these from a list of (url . (<page-data>)) elements.

My initial thought was an EDB (Emacs Database), but I think it'd be
simple just to keep an alist of directed edges (i.e. each
document/item is a node, and directed between links).  Updating ought
to be quick -- i.e. you could have
   insert
   validate/recompute
   delete
   find (all documents linked to one)
   find (all documents linked from one)

functionality, which then could be dumped into all sorts of
post-processing link and description tools.  But I'm getting beyond
myself.
(actual entries could be simple URIs/URLs, which ever suffices for
unique ID and resolution).

best,
-tony

address@hidden
Muttenz, Switzerland.
"Commit early,commit often, and commit in a repository from which we can easily
roll-back your mistakes" (AJR, 4Jan05).