|
Indexing
the Web with Geography
How
do you organise a list of your favourite, useful websites? Organising
items of similar information so they can be easily searched or browsed
to find one particular one requires an intuitive indexing system,
and is a problem that has challenged archivists, librarians and
scholars for generations. Clearly there are many possible ways to
index a list of websites - you might order them alphabetically based
on title or domain name (from Amazon.com to ZDNet), or chronologically
(based on when you last accessed the site). Alternatively, you might
group them thematically based on the type of information or services
they provide (with categories like sports, health, finance, games,
etc), or order them by popularity, or even by the size of the site.
The major Web portals do indeed apply many of these methods.
Geography
can also be used as an index. Location is a widely used as powerful
means of organising information. For example, it is used by governments
to collect taxes, elect politicians and undertake censuses. Maps
provide an intuitive visual tool to browse geographically indexed
information. A whole industry has grown up around the power of location
and mapping, known under the banner of Geographic Information Systems
(GIS) [1].
Geography
has also been used to index websites in form of so called sensitive
maps. These maps are website directories, with individual sites
indexed by their latitude and longitude (often just to the nearest
city). Sites are shown on the maps as dots, which act as hotlinks
to the homepage of the website. One of the best examples of a sensitive
map is the UK Academic Map, maintained by the University of Wolverhampton,
UK [2].
This
map provides a visual directory to the websites of the universities
and colleges in the United Kingdom based on their geographic location.
You simply click on the dot or the university name and you are taken
to the appropriate homepage. Map of the Month discussed
the map with its creator Peter Burden in a recent email.
|
The
UK
Academic Map created and maintained by Peter Burden
Copyright: University of Wolverhampton, 2000
|
TheUK
Academic Map went online in July 1994 (ancient history for the Web!),
having being created, according to Burden, for a seminar to local
industry chiefs showing them "the wonders of the Internet."
Burden is a professor in the School of Computing and Information
Technology at the University of Wolverhampton. He is interested
in cartography, as he said: "I've always been fascinated by maps
and possess a complete set of 1" to the mile OS maps, and that's
a fairly small part of the collection."
In the early days of the Web, circa '94-95, sensitive maps like
this one were a popular way of cataloguing websites, with many cities
and countries having one. At that time the Web was orders of magnitude
smaller than today and it was possible for people to maintain these
maps manually. However, the sensitive map approach did not scale
well to keep pace with the tremendous growth in both the number
and type of websites and services available. And most sensitive
maps have either been taken down or abandoned in a forgotten corner
of the Web. The UK Academic Map is one of few examples that is still
maintained and useful, largely because it specialises in indexing
a limited segment of websites - those of UK universities - which
is relatively static in size. The current map shows 204 different
institutions.
According to Burden, the main strength of the sensitive map is that
"... a very large number of links are presented in a natural
way in a small amount of screen real-estate" and that the mapping
approach is most appropriate "... for resources that have a natural
geographical association." The maps are most useful showing
a 'big-picture' overview of a lot of information.
Clearly
there are limitations with sensitive maps, both in terms of cartographic
design and at the conceptual level of information indexing. Burden
wryly notes that these maps, "...for non-geographical resources
can be rather confusing." In the case of university websites
it is appropriate because this type of organisation is strongly
associated with a location. (Indeed many are named after the city
or town they are in.) But for many other websites, a geographic
location is not an appropriate index. For example, where would you
draw the dot on the map to represent Amazon.com or ZDNet? Locating
the site at their headquarters would have little relevance to most
people searching the Web.
Another
important issue with using geography as the index is that a lot
of people have a bad sense of geography in terms of finding cities
or countries on a map. Americans seem to be particularly notorious
for their geographic illiteracy [3]. Burden reports for example,
that he "...occasionally receives messages to the effect of 'where's
Lampeter University?', .... it turn[s] out that the message was
from somebody overseas whose knowledge of UK geography was rather
rudimentary and certainly didn't extend to knowing the whereabouts
of Lampeter." [4] So, for some people, using geography
as the indexing key may make information navigation harder compared
to, say, browsing through an alphabetic list.
In
cartographic terms, producing sensitive maps for the Web has many
of the same design challenges as conventional map-making [5].
Symbols can quickly become cluttered as many websites are obviously
clustered in large cities which cover a relatively small geographic
area in relation to the amount of information they contain. This
is exacerbated in the UK Academic Map because all the symbols need
to be labelled with long university names. Problems are most apparent
with the high density of universities in cities like Manchester
and London, for example, making the designer's job difficult in
placing all the symbols, and particularly their labels, so that
they do not overlap.
In
the last year or so there seems to have been a resurgence in the
use of geography to "geo-enable" the Web. This is partly in response
to the rapid growth of the Web and the well known problems of "information-overload"
[6], the hope being that providing geographic knowledge within
the Web will help people to find more relevant sites and services.
Another factor has been the tremendous growth in mobile devices
like phones and PDAs, which, by enabling communications anywhere,
seem - somewhat perversely - to enhance the importance of location
[7].
A number of companies are developing geographic services for websites
to determine, in real-time, the location of their visitors and thereby
provide them with more relevant, tailored content. Some of the key
players in this field are Digital Island with their TraceWare technology,
Quova's GeoPoint service, and Digital Envoy's NetGeography [8].
They believe there is a market for technologies that can "localise"
the Web by reliably determining people's physical location (which
often really means at the level of territorial jurisdiction). The
need for this localisation has been shown in a number of recent
cases, such as selling Viagra over the Web, re-distributing TV programmes
[9], online gambling, and even selling Nazi memorabilia [10].
There
are also interesting developments in geographic-aware search engines
(such as iAtlas or Somewherenear [11]), as well as a scheme
called Geo Tags [12] to provide location metadata within
individual Web pages. There is even a well developed proposal for
a completely new top-level domain of .geo, which aims to "...
provide an open and scalable infrastructure to index, discover,
and serve any information on the web based upon the latitude-longitude
location of the data being referred to." [13] So it might
soon be time to register your house location as a Web address to
block a new generation of cyber-squatters!
As
for the UK Academic Map, Burden has no great plans for further expansion,
although "... somebody said the other day that there ought to
be a scale on the map, quite right too." Map of the Month
also asked him what his ultimate map of the Web would be, he replied,
Something
that allows automatic zooming with automatic name placement
(rather like the impressive Encarta Atlas) only going down
seamlessly to much finer detail, combined with a geographically
aware search engine, so you could define an area on the map
and then ask for all the bookshops (or whatever) in your selected
area. Must write that up as a research proposal!
[14]
|
So the days of the geography-free
Web may be numbered, as cyberspace is brought back down to Earth.
Map of the Month
next month
January 2001
Fly
Through The Web
(now back at
Mappa.Mundi Magazine)
|
|