Re: wisn99 (8/19-20)-- very early draft -- comments?

(no name) ((no email))
Fri, 11 Jun 1999 16:56:08 -0500

Rohit Khare circled Rodeo Drive with:

> Ah, but it *shouldn't* -- indeed, as you said, the whole virtue of
> the Web is that there is no central naming of even, say message-IDs.
> Anyone with a domain name can conjure up a web space, and put
> shape-shifting format negotiators behind any so-named Hollywood
> Square.

Doh. I misread the hint as a suggestion instead of an explanation.

> The hint is an explication of the fundamental gap that make the task
> DAV AColl placed before themselves so mind-slappingly difficult,
> 'sall.

The world may be my oyster, but I won't share its bathwater, er, inodes?


BTW, I think that hint and explication have incompatible rank, like seed-
thought and dissertation. A good deal of fertilizer/BS is usually involved
in the latter. From Webster's Seventh New Collegiate Dictionary (c) 1971:

explicate vt 1: to give a detailed explanation of
2: to develop the implications of


> [Anyway, Wayne, my point was that "onto" is the *perfect* word to be
> using; it's the definition of surjection in the first place...]

I understand that surjection === onto. I'm not sure yet about the hypothesis.
Is there not a requirement at some level for bijective namespace, for practical
reasons, if not for something more abstract? Let's say I have a large number
of directory entries for huge files, but most of the entries are merely
symbolic links to a relatively small number of real files. This system lacks
injectivity, and would certainly be bad for replication. Is that outside the
concern of the web-scale directory models, or will they require such metadata?
Or, take an even sillier example, is
niljective (abjective?), or more like a shape-shifter. Not that I expect web
directories to provide much help against spambots, mind you.

Somehow, I'm amused by this:

File Not Found
The requested URL /IRUS/WISN.jpg was not found on this server.