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

(no name) ((no email))
Fri, 11 Jun 1999 14:21:35 -0500


Okay, here are my comments, FWTW, from the fringes.

> after all, namespaces are just mappings (typically, surjections) from one
> set of symbols called names to another set called addresses.

I approve of the use of the term surjection, but only on educational grounds.
It will cause a few people to look up the word. You may get some objections
subsequently from pedants like me, who will complain that such mappings, while
one-to-one, are usually not onto. There are often many aliases not covered.
Ah, but, I see you mention that below.

> WISN'99 is a name to the human, which he/she resolves to the address:
> http://www.ics.uci.edu/IRUS/WISN.jpg which ...

I map it to Rohit, myself. But I like the cascading codomains notion.

> The filenamespace is controlled by the sysadmin, and then the user.

Which obviates the sysadmin's control, until the BOFH/BSAFH emerges.

> It is hierarchically resolved to an inode address, which UNIX uses to
> indirect the "human name" of a file in the face of moves, renames,
> links, and so on.

Once more, the mapping function's image may only be a virtual codomain, with
subtle ambiguities (not to mention hardware/software/people failures). So,
there's no way for most people to know that:

~baisley/html/yelsiab~/vog.lanf.sso-www/:ptth/index.html ->
~baisley/html/palndrom.html

but it makes my life easier, however frivolously. And that's not such a
frivolous point. Terminal drivers are more complicated than network drivers,
because people are more complicated than machines. Machines "like" everything.

> Aliases, hierarchies, and "dangling links" all compromise the
> integrity of a namespace.
> Looking beyond numbers, though, we see issues of trust: who controls
> the registry? How can we operate truly decentralized namespaces?

Perhaps, but if the web weren't dynamic, late-binding, and occasionally
ambiguous, it would never have happened or survive. Once it's decentralized,
it's out of the center's (existentially, my) control. This is still a net win,
isn't it? After that, it's up to you psychohistorians to write appropriate
standards. ;-)

> Are search engines viable, de-facto resolvers for a great-many IScale
> namespaces, such as the "set of CS Departments", public key
> locations, maps and so on?

This problem is akin to writing a dictionary, in that the vast majority of the
resulting output will be of no interest to the vast majority of users. And it
is being done with volunteer labor of vastly variable skill, with extensions
into fields well beyond philology (what does 'integrate' mean, with respect to
such-and-such a function? what's the right word to use in this sentence,
assuming that my target audience is, say, hungry, bilingual Canadians?), and
even previously uncharted (sometimes blessedly so) domains of information.

Dictionaries work well for those with a suitable knowledge base and gumption.
So do search engines. They are expensive, but the investment isn't always a
waste, serendipitously. But, something far more useful and tractable must be
possible. I've no clue (obviously) what that would be. In the past it was
reference librarians, among others. This could wander off into the AI weeds.

> (hint: the Web doesn't have inodes as persistent, intermediate
> addresses for documents in the face of moves, renames, and deletes).

So what, you want a permanent wnode for every web object, the web as a database
in 3rd normal form? I don't see how that would help without considerably more
persistence in the objects themselves.

> Why is there such fear that unrestrained deployment of XML namespaces
> -- which essentially map tag-names to URLs -- could lead to a
> gridlock of hypertext versioning roadblocks, in which no one document
> can be read from the network without an indeterminate number of
> recursive lookups succeeding as well?

Is it because adding more dictionaries won't teach Johnny how to read?
Granted, a pocket dictionary is an improvement on the OED in a number of
practical ways. It won't give you a hernia. Unless you have to carry 1000000
of them around to do useful work.

> who don't even have bedrock addresses like SSNs, or even unique family names?

Or published maiden names. Counter-example: Liz Taylor.

Hope you make some good progress. It's a worthy challenge.

Cheers,
Wayne

"Hey, we can't all *notice* the same subtleties." -Tom Servo (*emphasis* mine)

http://www-oss.fnal.gov/~baisley