[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Tag-ID caveat
-----BEGIN PGP SIGNED MESSAGE-----
At 04:11 PM 4/8/97 -0700, Tony Bartoletti wrote:
>I note here that any tag-originator can create "virtual tag-ids". Under the
>assumption, foreinstance, that as primary spending authority for company X,
>I can establish the form for all of my "spend" tags, I could replace the
> (tag (tag-id xxx) (spend (account) (amount)))
> (tag (spend (spendtag-id xxx) (account) (amount)))
>and hope that this still affords some level of increased efficiency, in that
>the generic SPKI-tag-intersector will short-circuit as soon as the spendtag-id
>is determined not to match.
I agree and advocate this approach, if it is ever found necessary.
>The only real argument for elevating the tag-id up to the level of all tags
>would be to assume that the SPKI certware is wired to attempt comparison of
>tag-ids first, and separate from invoking the full intersect-tag-bodies code.
>(Ie, it eliminates one level of processing depth when scanning for tag-match.)
The first level of scan is for subject -- and that should significantly cut
down the population under examination.
-----BEGIN PGP SIGNATURE-----
-----END PGP SIGNATURE-----