One nit, though. This is hardly a problem with WWW alone.
Archie/Veronica fail miserably if you do not actually know beforehand
what you are looking for. In other words, if I know the exact book I
require I can search a list for the title and find the exact location in
the library where it resides. But woe is me if I am looking for books on
a particular topic, the librarian only gives me a blank stare and points
me to the same list of titles.
Since the only place where this information exists is in the
administration of the repository itself, it does little good to reap book
titles. Only the administrator of the repository knows what "resources"
she offers to the network at large.
Currently all resource navigation tools on the gI suffer from this
problem (gI := global Internet.) The only effort which I believe comes
close to being a useful solution to this problem is the "topic oriented"
gopher lists. But the failure here is the fact that this information
must be gathered by a human without any help from the source of the
information -- the repository. Sort of like taking down names of
businesses from the phone book and deciding in which category that
business might reside.
I strongly feel that the repository must shoulder the role of the
cataloger. But (before all you repository managers balk at the size of the
problem), we really need to take a very hard look at what it is we need to
catalog.
I believe that the gI is made up of "resources" where that word refers to
collections of information -- not to the information itself. For example:
The Electronic Newsstand is a resource, as are the magazines it
encompasses, but the individual articles are not resources.
Counterpoint's Federal Register is itself, a resource, as are the
various gubermint agencies included, but not the individual rules and
notices.
It has been in my mind for some time to start something which begins to
categorize the value which the gI represents, but does not specifically
index. (And this is the real question, the difference between an index --
which serves one valuable purpose, and a "table of contents" -- which
serves quite another.)
So...
I propose the GRIP -- Global Resource Identification Project -- where
-- managers of internet resources (a concept, to be clearly defined) are
provided with an IAFA-like template of a form along the lines of:
Name: Out Magazine
URL: gopher://enews.com/1/collected/out
news://enews.out
mail:info@enews.com
Location: Palo Alto, CA
Description: The world's leading magazine of gay and lesbian
issues. Published monthly.
Keywords: Gay Lesbian Homosexual Queer Politics Entertainment
Category: Publications/Magazines
Politics/Gay and Lesbian Issues
Culture/Human Sexuality
Abstract: {URL which points to a long description}
-- the filled in template is retrieved on a regular basis from the
resource site and used to create a number of navigation tools:
Under gopher:
GRIP/
Publications/
Magazines/
Out Magazine/
Information --> the text of the template
To connect --> points to the actual resource
Also incorporated in similar ways into databases accessable via
WAIS/WWW/Whois++ etc.
Guilding Principles:
- provide a template format which has clearly identified
ranges of complience, eg. Required/Recommended elements
- maintain some editorial control over the actual categorization
mechanism to enhance the value of the service
- provide the database in as many forms as possible
- provide the raw data (templates) to allow others to create
a rich set of tools to leverage it in different ways
- distribute the load over as many well connected sites as possible
Offer:
If we can come up with a document which clearly identifies what is
and is now a "resource"...
If we can come up with an "approved" template...
And if we can come up with an "official" categorizing method...
The Internet Company will donate programming/MIPS/pipe and Megs
to get this off the ground.
Ultimately, this is an effort which should be distributed over the entire
gI in a reasonable manner. Portions of the gopherspace so generated
could be distributed on various altruistically inclined sites, as well as
the WWW and Whois++ databases. Dunno 'bout the WAIS stuff. I suppose
it's possible to break this out into categories and WAIS up the categories.
I also think that the raw data (templates) retrieved from each site
should be made available to any and all, to develop a richer set of tools
to interpret the data.
Ok? Who wants to play?
-- </rr> Rob Raisch, The Internet Company