RE: Web site definition
From: Lavoie,Brian (lavoie@oclc.org)
Date: Tue, Jan 05 1999
Message-ID: <72B89459DD2BD211B5CD0000F840094E1074EC@oa3-server.dev.oclc.org>
From: "Lavoie,Brian" <lavoie@oclc.org>
To: "'www-wca@w3.org'" <www-wca@w3.org>
Date: Tue, 5 Jan 1999 17:20:24 -0500
Subject: RE: Web site definition
Joe,
Thank you for the clarification. The points you bring up strike at the heart
of an issue that we have been wrestling with here at OCLC, which is how to
identify "logical sites" without resorting to manual analysis of harvested
Web pages. The definition of a Web site I offered was the only approach we
could formulate that could be applied unambiguously in automated processing.
You are quite right when you talk about a web site as a something like a
"perception" on the part of its viewers, and that the perception may not fit
well with "physical" criteria, such as server location. Another example of
this point that we have encountered is the case of a Web site that can be
accessed through a single domain name, but the actual location of the files
that are accessed is randomized over a series of IP addresses by a load
distributor. From the perception of the viewer, there is only one Web site
(uniquely identified by the domain name), but technically, there are
multiple Web sites according to the "host" definition.
Issues such as this are especially important in developing consistent and
meaningful web metrics; I hope that some kind of consensus emerges from this
Activity with respect to pinning down these terms.
Regards,
Brian
-----Original Message-----
From: meadowsj@nobs.ca.boeing.com [mailto:meadowsj@nobs.ca.boeing.com]
Sent: Tuesday, January 05, 1999 4:46 PM
To: www-wca@w3.org
Subject: Re: Web site definition
>I have some concerns about your definition of the term "Web site". In
>particular, I'm not sure what the phrase "all Web pages and files under a
>common owner" means. If this means all Web pages produced by a particular
>entity (say, the W3C), I would think that this definition would be hard to
>apply in practice, since authorship of Web documents can be difficult to
>determine.
I believe that came from a set of definitions we were using within Boeing,
and reflects the usage of the terms by our own internal customers. While
it might be impossible to externally determine what makes up a "web site"
by this definition, it's clearly how our customers prefer to think of the
term. Generally multiple groups may share a single web server, but they
would be partitioned off from each other via separate directories and
separate access controls, etc. In other cases, a group may maintain a single
logical "site" that happens to span multiple web servers.
Given that there are at least three major groups (information consumers,
information providers, and infrastructure providers), each with slightly
different
perspectives on some things, it might make sense to identify which terms are
used differently and how... The definitions I offerred up were developed by
the
third group, intending to capture the meanings used by the first two
groups..
Cheers,
Joe