Re: URL respitrory

[Nat Friedman <ndf ALEPH1 MIT EDU>]

On Thu, Oct 01, 1998 at 11:31:13PM +0100, Daniel Barlow wrote:
> On Wed, Sep 30, 1998 at 06:31:40PM -0400, Nat Friedman wrote:
> > That sounds like a good way of implementing the stuff I was talking
> > about, with the small reservation that I don't know how you would
> > store the hierarchy stuff in the database.
> 
> Usually, by having columns for 'id' and 'parent id' in each row.
> (Like with directories or processes or whatever).  Or was this not
> what you meant?

I guess it can be made to work in a database pretty easily, then.

> But I thought the idea was to use keyword combinations rather than a
> hierarchy, anyway.

Yeah, that's right; that's the main idea.  But I didn't want to make it
impossible to superimpose a hierarchy onto the repository.

> When they get to sharing those indices, suddenly you have a really
> quite cool tool (and one I've thought about writing myself, at that.
> But, CMUCL bindings for Gtk are still first on my list)

Yes, I agree that it'd be cool to have people sharing these indices
with each other, but the main idea here was just to have deep
integration of URL archiving and retrieval with GNOME.  And I wouldn't
want that idea to get lost in a grandiose scheme to fully categorize
the web.

Cheers,
-Nat

> [What does notyop suggest for "respitrory", anyway?]

Depends on how long you let it run.  Note that there are three typos
in this single word: (1) the extra 'r' in "..trory" (2) the inversion
of 's' and 'p' (3) the missing 'o' between the inverted 's' and 'p'.
So If you tell NoTyop to look hard, it will apply all of its
permutations and find the correction.  On a less aggressive setting,
however, it won't find a replacement word.