Re: Flat Query

Anne & Lynn Wheeler wrote:
> "David Cressey" <david.cressey_at_earthlink.net> writes:
> > Good Summary.
> >
> > About the only place I still see the argument between exposed
> > pointers and indexes is... right here in the comp.databases.theory
> > newsgroup, where our resident gadfly is still trying to persuade us
> > to go back to pointers, and start over!

and most of the arguments i have read seem to degenerate because of mixed definitions of what a pointer _is_. I'd be very interested if someone has a precise definition of the term as used in the database field (including the distinction of a logical pointer).

> i wonder ... do i have on archived post on this topic from last
> decade http://www.garlic.com/~lynn/94.html#26 Misc. more on bidirectional links
>
> some minor historical regression ...
>
> and of course, hyperlink stuff traces back to Nelson's xanadu
> http://www.xanadu.net/
>
> how about: WWW, what went wrong
> http://xanadu.com.au/xanadu/6w-paper.html
>
> and Engelbart's nls/augment
> http://sloan.stanford.edu/MouseSite/dce-bio.htm
>

I have worked extensively with Nelson (and Engelbart) and part of the misconceptions about "pointers" and the web, lie in the bastardization of Nelson's term "hyperlink".

A bi-directional link is of course nonsensensical. A link is, by definition, always bidirectional (as the links of a chain are). To Vannevar Bush the link was no more than a physical mechanism to create an informational trail (this is Nelson's main influence) which could be followed in either direction. Berners-Lee extracted from Xanadu the link concept, but dropped the bi-directional requirement, creating weblinks, a form of "half link" - but no more than a pointer. Berners-Lee (perhaps correctly) reconciles this in terms of the "worse is better" argument, and that the approach more appropriately suits the logistics of the web and its ad-hoc chaotic nature. Unfortunately the word's meaning seems to have devolved because of this.

However, the web's pointers are still a concession. In terms of functionality there is no instance where using a pointer (half-link) instead of a link offers any logical advantage, and rather consistently deleterious, generating administrative turmoil (404's or endless reorganisational housework).

Now in the RM, links exist (in the original sense of the word) and are assumed by much of relatioanl calculus. They are of course implict, encoded into equality comparisons (dynamic links), as opposed to the explicit static links of traditional hypertext - but the concept applies nonetheless. So in terms of a logical informational model, I can see no point whatsoever of removing this bidirectional functionality that took decades to obtain, and that the web still lacks. Received on Sat Oct 15 2005 - 21:24:24 CEST