Semantic Visualization 
Penny Chase, Ray D'Amore*, Nahum Gershon*, Rod Holland, Rob Hyland, Inderjeet Mani*, 
Mark Maybury, Andy Merlino, Jim Rayson 
The MITRE Corporation 
202 Burlington Road 
Bedford, MA 01730, USA 
* 1820 Dolley Madison Blvd. 
McLean, VA 22102, USA 
{pc, rdamore, gershon, rholland, hyland, imani, maybury, andy, jrayson}@rnitre.org 
ACL-COLING Workshop on Content Visualization and Intermedia Representations 
Abstract 
This paper summarizes several initiatives at MITRE 
that are investigating the visualization of a range of 
content. We present results of our work in relevancy 
visualization, news visualization, world events 
visualization and sensor/battlefield visualization to 
enhance user interaction in information access and 
exploitation tasks. We summarize several initiatives 
we are currently pursuing and enumerate unsolved 
problems. 
1. Visualizing Semantic Content 
Visualization can support effective and efficient 
interaction with a range of information for a 
variety of tasks. As Figure 1 illustrates, 
information (data elements, attributes, relations, 
events) can be encoded in (possibly interactive) 
visual displays which users can exploit for a 
variety of cognitive tasks such as retrieval, 
analysis (e.g., of trends, anomalies, relations), 
summarization, and inference. In this paper we 
consider a range of semantic content, visual 
mechanisms, and cognitive tasks to deepen our 
understanding of the role of interactive 
visualization. 
f21~ .... 
Figure 1. Information Visualization Process 
2. Document Relevancy Visualization 
Today's users are faced with a dizzying array of 
information sources. MITRE's Forager for 
Information on the SuperHighway (FISH) 
(Smotroff, Hirschrnan, and Bayer 1995) was 
developed to enable the rapid evaluation of 
information sources and servers. Figure 2a 
illustrates the application of FISH to three Wide 
Area Information Server (WAIS) TM databases 
containing information on joint ventures from the 
Message Understanding Conference (MUC). 
Figure 2b illustrates the application of FISH to 
visualize e-mail clustered by topic type for a 
moderator supporting a National Performance 
Review electronic town hall. 
52 
Figure 2a. WAIS FISH Figure 2b. NPR FISH 
The traditional WAIS interface of a query box 
and a list of resulting hits is replaced by an 
interface which includes a query box, a historical 
list of queries, and a graphically encoded display 
of resulting hits (an example of which is shown 
in Figure 2a). In WAIS, the relevancy of a 
document to a given keyword query is measured 
on a scale from 1-1000 (where 1000 is the 
highest relevancy) by the frequency and location 
of (stems of) query keywords in documents. 
Motivated by the University of Maryland's 
TreeMap research for hierarchical information 
visualization, FISH encodes the relevance of 
each document to a given query (or set of 
compound queries) using both color saturation 
and size. 
In the example presented in Figure 2a, each 
database is allocated screen size in proportion to 
the number of and degree with which documents 
are relevant to the given query. For example, the 
MEAD database on the left of the output window 
is given more space than the PROMT database in 
the middle because it has many more relevant 
documents. Similarly, individual documents that 
have higher relevancy measures for a given query 
are given proportionally more space and a higher 
color saturation. In this manner, a user can 
rapidly scan several large lists of documents to 
find relevant ones by focusing on those with 
higher color saturation and more space. 
Compound queries can be formulated via the 
"Document Restrictions" menu by selecting the 
union or intersection of previous queries, in 
effect an AND or OR Boolean operator across 
queries. 
In Figure 2a, the user has selected the union of 
documents relevant to the query "japan" and the 
query "automobile", which will return all 
documents which contain the keywords "japan" 
or "automobile". Color coding can be varied on 
these documents, for example, to keep their color 
saturation distinct (e.g., blue vs. red) to enable 
rapid contrast of hits across queries within 
databases (e.g., hits on Japan vs. hits on 
automobile) or to mix their saturation so that 
intersecting keyword hits can be visualized (e.g., 
bright blue-reds could indicate highly relevant 
Japanese automobile documents, dark the 
opposite). In the example in Figure 2a, blue 
encodes Japan, red Automobile; the color coding 
is set for mixed saturation, the union of the 
relevant document sets for those two keywords is 
selected, and the order (from top to bottom in the 
display) is used to encode the WAIS relevancy 
ranking. One issue is just how effectively users 
can discriminate mixed colors. 
fGA.O study hnk$ (hemi<al Oou~le clkk to edit (ol~s: ~rag (o remange 
t -e,ght 
j I~r¢,du~ t -e~.tl 
t og/~ 3/96-T RANSC RtPT: Jmako n=~lt on the tame ~*vli c~ble 
ic .......... ,~..: ,,~0 ~ c.~, L.~. " 
IL,*2U;,°o'Z~%"-;';:' ,~%';;;: ~'0, ,, ............ ,,,J 
,1: ............................. ,Ix 
Figure 2c. J-FISH Multiserver Visualization 
More recently, we have explored multiple server 
evaluation on popular World Wide Web search 
engines. For example, Figure 2c illustrates a 
query across multiple servers. Research issues 
include differences in relevancy ranking 
algorithms, encoding of multiple attributes 
beyond relevancy using color or size (e.g., length, 
53 
quality, cost, source), and document collections 
which are heterogeneous in size, content, and 
format. 
3. Document Structure/Content Visualization 
Figure 3a (Gershon et al. 1995; Gershon 1996) 
illustrates another navigation mechanism in 
which the user is able to view a hierarchy of the 
browse space. The left: hand of Figure 3a 
displays the traditional HTML layout of a web 
page whereas the right hand side illustrates a 
hierarchical, navigable view automatically 
generated from the underlying structure of the 
browsing space. The user can create a personal 
space by interactively and visually modify the 
structure of hyperspace or extracting segments of 
the documents. 
T'h¢ M I T~4.E ~'por md4n k • mudWac~c 
~'o*'idm t~ekn.,e'tJ tnd ma,tepc paidmsee 
Itn~ ¢lr~irau~t i1 i~n~ i 
\u~ ~kH 
Figure 3a. Hyperspace Structure Visualization 
For discovery and analysis of new information 
and relationships in retrieved documents, we 
have developed a method for aggregating 
relevant information and representing it visually 
(Gershon, et al, 1995). The method is based on 
representing correlations of words within a 
document in a table. These tables could be very 
large depending on the size of the document thus 
making it difficult for the user to perceive and 
make sense of all the highly relevant correlations. 
Since the order of the words is not usually based 
on contents, the order of the words is permuted 
until the highly relevant correlations are 
concentrated in one comer. 
Fat 
Fatigue 22 
Aches 10 
Alments 2 
nausea 15 
Smoking 
3 
Snacks 
47 
11 
33 
sedentary 
2 
4 
Fatigue 
Aches 
Nausea 
Ailments 
fat snacks 
22 
10 
15 
smoking 
3 47 
11 3 4 
33 5 1 
2 4 
sedentary 
2 
Figure 3b. Example of Unaggregated (top) and 
Aggregated (bottom) Tables 
Other research at MITRE has focused on 
automatic discovery and visualization of 
semantic relations among individual and groups 
of documents (Mani and Bloedom 1997). Figure 
3c illustrates the results of visualization of a set 
of documents using the NetMap visualization 
software after clustering these into related groups 
which appear around a circle. Outside of each 
cluster on the circle are displayed intracluster 
relations; in the center of the circle are 
intercluster relations (e.g., a shared named entity 
such as a person, place, or thing which appears in 
multiple documents). The user can zoom in any 
part of the graph. This is shown in Figure 3d, 
which shows individual people (green) and 
organizations (aquamarine). 
Selecting an individual entity from a document 
returns a display such as that in Figure 3e. Figure 
3e illustrates individual entities encoded with 
color and shapes (e.g., people in green stick 
figures, organizations in aquamarine diamonds, 
locations in purple jagged rectangles, documents 
in yellow circles, person-organization relations in 
white squares). Lines and their properties (e.g., 
color, dashed) can encode relations among these 
entities (e.g., co-occurrence in documents). This 
provides a richer mechanism for discovering 
54 
interdocument and interentity relationships 
during analysis. Current research is 
investigating the role of automated text 
summarization, document retrieval and 
navigation and visualization. 
Figure 3d. Zooming in on Document Cluster 
Figure 3c. Document Cluster Visualization 
Figure 3e. Entity Relation Visualization 
55 
4. Named Entity/News Visualization 
MITRE's Broadcast News Navigator (BNN) is a 
system that is investigating analysis of trends in 
news reporting. BNN performs multistream 
(audio, video, text) analysis to eliminate 
commercials, segment stories, extract named 
entities (i.e., people, organization, location) and 
keyframes, and classify and summarize stories 
(Merlino, Morey, and Maybury 1997). BNN's 
intuitive web-based interface gives the user the 
ability to browse, query, extract from and 
customize digitized broadcasts. Figure 4 
illustrates a trend analysis display from BNN that 
shows the most frequently mentioned named 
entities reported on CNN Prime News TM from 
October to November of 1997. "China" spikes in 
the center of the graph, associated with a state 
visit to Washington. Later "Iraq" spikes which is 
correlated with news regarding UN site 
inspections. The user can click on any point on 
the line graphs and be brought to a list of stories 
that mention that named entity. 
Ik'llq~ 11 
~.~ 
Figure 4. Broadcast News Visualization \] 
In contrast, the user can formulate a query 
specifying keywords, named entities or subjects. 
Figure 5a shows the results of executing the 
query: Find me stories which have a topic of 
1 Note in the display the occurrence of the terms "U.S." and 
"United States". BNN performs no co-reference resolution, 
a topic of current research at MITRE. 
"chemicals", the keywords "chemical weapons", 
person "Sadam Hussein", organization 
"Pentagon", and location "Iraq". Each story in 
this "Story Skim" view is represented by a 
keyframe and the three most frequent named 
entities. Selecting one of these stories yields a 
"Story Detail" display, which as shown in Figure 
5b including a keyframe, named entities, subject 
classification and pointers to the closed caption 
and video source. 
Figure 5a. BNN "Story Skim" Visualization 
Summary Closed Source Topics 
Figure 5b. "Story Detail" visualization 
Current research is exploring connecting these 
broadcast news stories with 
visualizing topic frequencies 
mechanisms for low quality 
transcriptions of broadcast 
56 
intemet stones, 
over time, and 
spoken language 
stories. Other 
investigations are focusing on which presentation 
mixes (e.g., keyffames, named entities, one line 
summary, full video source) are most effective 
for story retrieval and fact extraction from news 
(Merlino and Maybury 1998). 
5. Geographic Event Visualization 
The Geospatial News on Demand Environment 
(GeoNODE) initiative at MITRE is a new project 
investigating visualizing geographic aspects of 
news events. This program builds on MITRE's 
BNN, described in the previous section, and 
MSIIA, addressed in the subsequent section. 
GeoNODE is based on the research area of 
Geographic Visualization which investigates 
methods and tools that impact the way scientists 
and others conceptualize and explore 
georeferenced data, make decisions critical to 
society, and learn about the world (MacEachren 
and Ganter 1990, Taylor 1991). Since news 
reports are about events in the world, the reported 
events and trends can be assessed, queried, and 
reviewed effectively by leveraging a person's 
preexisting knowledge of the world's geography. 
The objective of GeoNODE is to understand the 
information integration of geospatial/temporal 
visualizations, information retrieval, multimedia, 
and other technologies to support browsing, 
analysis, and rapid inference from broadcast 
news. 
As shown in Figure 6, GoeNODE will analyze 
global and local cooperation and conflict found 
in broadcast news, internet, newswire and radio 
sources as well as broadcast news. Processing 
will include the identification, extraction, and 
summarization of events from national and 
international sources. GeoNODE will consider 
event types (e.g., terrorist acts, narcotrafficking, 
peace accords), frequency, and severity in an 
interactive geo-spatial/temporal context that 
supports browsing, retrieval, analysis and 
inference. 
57 
Figure 6. GeoNODE Architecture 
Although a geographical context can enhance a 
person's understanding of reported events and 
therefore facilitate news retrieval and further 
queries, the same familiar visualization concerns 
apply to geographic presentation that are salient 
in visualizing any data rich multivariate 
information space. The GeoNODE user 
experience is derived from research, experience 
and standard practice in the visual search and 
retrieval domains: Overview first, zoom and 
filter, then details-on-demand (Shneiderman 
1994). During each stage of the visualization 
process, cartographic methods and spatial 
analysis techniques are applied. These can be 
considered as a kind of grammar that allows for 
the optimal design, production and use of maps, 
depending on the application (Kraat 1997). Select 
cartographic generalization operators are applied 
to address key multi-scale and information 
overload problems (Buttenfield 1991). 
GeoNODE addresses Knowledge Representation 
(KR) and information fusion issues that are 
important to the news event presentation. The 
KR activities specific to GeoNODE are 
concerned with discovering and manipulating 
geospatial and temporal information, specifically 
investigating the following: 
improved natural language processing of 
place names that are central to 
understanding a news report 
• news event modeling 
• cartographic generalization rules 
• transformation of news events to visual 
metaphors 
Spatial information management is currently 
growing in its utility to commercial applications, 
and several industries have already begun to 
explicitly rely on GIS systems, although most 
(53%) companies are evaluating while an average 
of only 7% are implementing or using a GIS 
(IDC 1997). Accompanying the growing interest 
in spatial information is a technology trend 
influencing the architecture of GeoNODE, 
mainly, a shift from single-purpose/standalone 
GIS applications to geospatial extensions and 
services for databases, component frameworks, 
data warehouses and data analysis applications. 
By supporting a component-based architecture, 
GeoNODE can more readily take advantage of 
future geospatial services and an expanding 
number of news sources (internet, newswire, 
radio, and other broadcast sources). 
Further research will investigate incorporation of 
summarization, geospatial/temporal KR, and 
other traditional visualization techniques. For 
example, Figure 7 illustrates some of the kinds of 
visualizations that are being explored by other 
researchers, such as the use of color and 
geolocation to encode relations among 
geographic entries. Figure 7 is a geographic 
visualization of early WWW usage available at 
http ://www.cybergeography.org/atlas/atlas.html. 
These and other research threads will shape 
GeoNODE into a visualization component for 
reasoning about news events in geographic space. 
As a long term objective, the system architecture 
should allow for navigation and retrieval from 
topic, conceptual, and web spaces where a user 
can access, update and annotate existing data 
with spatial information. 
Figure 7. Visualization of Geospatial Relationships 
6. Sensor Visualization 
The Multisource Integrated Information Analysis 
(MSIIA) project, led by Steve Hansen at MITRE, 
is exploring effective mechanisms for sensor and 
battlefield visualization. For example, national 
and military intelligence analysts are charged 
with monitoring and exploiting dozens of sources 
of information in real time. These range from 
sensors which capture images (infrared, electro- 
optical, multispectral) to moving target indicators 
characterized by a point and some features (e.g., 
tracked vs. wheeled vehicle) to signals 
intelligence characterized by centroids and error 
elipses. Knowing which source to select and 
which sensors to task is paramount to successful 
situation assessment. An integrated view into 
what sensors are where when, as well as a fused 
picture of their disparate information types and 
outputs, would be invaluable. Figure 8 illustrates 
one such visualization. The x-y dimension of the 
upper display captures the coordinates of a 
geospatial area whereas the y coordinate displays 
time. This enables the user to view which areas 
are being sensed by which type of sensor 
(encoded by color or implicitly by the resultant 
characteristic shape). For example, a large 
purple cylinder represents the area over time 
imaged by a geosychronous satellite, the green 
cylinders are images taken over time of spots on 
the surface of the earth, whereas the wavy blue 
line is the ground track of a sensor flying across 
an area (e.g., characteristic of a unmanned air 
vehicle such as predator). If we take a slice at a 
particular time of the upper display in Figure 8 
we get the coverage of particular areas from a 
specific time. If we project all sensor coverages 
58 
over an area downward to the surface, we obtain 
the image shown in the lower display of Figure 8. 
military for planning and training. The sand can 
be sculpted to match the terrain in a specific 
geographic region. People standing around the 
table can place plastic or metal models of 
vehicles and other assets over this terrain model 
to indicate force deployment and move them 
around the terrain to indicate and/or rehearse 
force movements. 
Figure 8. Sensor Coverage Visualization 
A user can utilize this display to determine what 
material is available for a given time and space, 
analyze unattended coverage areas, and plan 
future collections. MSIIA is also investigating 
georegistration and display of the results of 
collections in an integrated, synthetic view of the 
world (e.g., fusing maps with images with radar 
returns). We consider next another example of 
synthetic views of the world. 
7. Collaboration and Battlefield Visualization 
Just as visualization plays an important role in 
information space visualization for MSIIA, 
MITRE's research on the Collaborative 
Omniscient Sandtable Metaphor (COSM) seeks 
to define a new generation of human-machine 
interfaces for military Command and Control 
(C2). The "sandtable" underlying COSM is a 
physical table whose top is rimmed with short 
walls and filled with sand. It is used in the 
59 
In defining COSM, we expanded the 
functionality of a sandtable and moved it into an 
electronic domain. It now taps into global 
gigabyte databases of C2 information which 
range from static data on airfield locations, to 
real-time feeds from hundreds ground, air, and 
space based sensors. This data is used to 
synthesize macroscopic or microscopic views of 
the world that form the foundation of a 
collaborative visualization system. Manipulating 
these views leads not only to modifying data, but 
also directing the actions of the underlying 
physical assets (e.g., moving an icon causes an 
aircraft to be redirected from point A to point B). 
A conceptual view of COSM is shown in Figure 
9, where participants at air, land, and sea 
locations collaborate over an electronic 
sandtable. Some users are physically present, 
while others are represented by their avatars. 
The key elements of COSM are geographic 
independence (transparent access to people, data, 
software, or assets regardless of location), a 
multimodal, direct manipulation interface with an 
initial emphasis is on the visual modality, 
heterogeneous platform support (enabling users 
to tailor data depictions to a range of platform 
capabilities), and data linkage (maintaining all 
parent, child, and peer relationships in the data). 
Figure 9. Conceptual View of COSM 
Figure 10. Virtual Reality Instantiation 
A first instantiation of COSM was implemented 
using Virtual Reality (VR) technology, as 
illustrated in Figure 10. The table is a 
stereoscopic projection system driven by a 
graphics workstation. It uses a horizontal display 
surface approximately 6 feet wide and 4 feet deep 
to display maps, imagery, and models of the 
terrain and objects upon or above the terrain. 
Since it is stereoscopic, objects above the terrain, 
such as airbome aircraft, appear to be above the 
surface of the table. The vertical screen behind 
the table is a rear-projection display used 
primarily used for collaboration support. At the 
top, we see a panel of faces representing all the 
remote users who have similar systems and are 
currently connected to this one with audio, video, 
and data links. The table serves as a shared 
whiteboard that is visible to all the users and can 
be manipulated by them. The larger faces at the 
bottom of the vertical screen are two users who 
have "stepped up to the podium" and currently 
have control of what it being seen on the table. 
The figure shows the user interacting with the 
table through the use of two magnetic position 
trackers. The first is attached to a pair of 
stereoscopic glasses, and as the user moves his 
head and walks around the table the computer 
determines his eyepoint location from the tracker 
and recomputes his view accordingly. The 
second tracker is attached to a glove that serves 
as an input device. The user's gloved hand 
becomes a cursor and he can use his fingers to 
touch an object to indicate selection or grab and 
move an object to indicate an action. 
Several different kinds of information can be 
displayed on the table. Figure 11 illustrates a 
display of current air and ground information. 
There are several aircraft depicted as realistic 
models, with the relative scale of the models 
representing the relative sizes of the respective 
aircraft. They move in real-time, with the 
stereoscopic display making them appear to be 
flying above the table. Conceptually, the 
positions of the aircraft are provided in real-time 
by a radar system and the user has the option of 
displaying them as symbols or models. Remote 
users worldwide have real-time access to the 
data. The hemisphere in the upper left is a 
simple, unclassified representation of the threat 
dome of a Surface to Air Missile (SAM) 
emplacement. The large arrow is a cursor that is 
controlled by a remote user who is collaborating 
over this display. The amorphous blob in the 
lower left is a depiction of a small storm cell that 
is also moving through the region. This weather 
data is visually integrated in real-time with the 
current air picture data. The aircraft position, 
weather, and threat information are all provided 
by different sensor systems. However, they share 
a common spatiotemporal reference that allows 
them to be fused in this real-time synthetic view 
of the world. Every object in this synthetic view 
also serves as a visual index into the underlying 
global C2 database. Selecting an aircraft would 
let us determine its current status (airborne with a 
certain speed and heading) and plans (origin, 
60 
destination, and mission), as well as associated 
information such as logistics at its base of origin. 
Figure 11. Synthetic View of the World 
Our current research is focused on the use of 
aggregation and deaggregation of data within 
visual depictions, in order to support a wide 
range of users. A weaponeer wants to study the 
details of a target (e.g., construction material, 
distance below ground) that is only a few 
hundred feet by a few hundred feet in size. A 
commander wants an overview of all airborne 
assets, targets, etc. for a region that is several 
hundred by several hundred miles in size. 
However, those examining an overview will 
frequently wish to "drill down" for maximum 
detail in certain areas, while those examining a 
detailed area may wish to examine a more global 
view to retain context. Allowing the 
visualization of data with this wide range of 
geographic scopes, as well as iterative travel 
between detail and overview, poses challenges in 
both data depiction, data simplification, and 
intuitive navigation techniques. 
8. Conclusion and Research Areas 
The above varied and rich application spaces - 
e.g., visualizing search results, topics, relations 
and events in news broadcasts, battlefield 
activities - provide a number of challenges for 
visualization research. Fundamental issues 
include: 
1. What are effective information encoding/ 
visualization techniques for static and 
dynamic information visualization, including 
complex semantic objects such as properties, 
relations, and events? 
2. What are the most effective methods for 
utilizing geospatial, temporal, and other 
contexts in synthetic displays of real world 
events that facilitate interface tasks (e.g., 
location, navigation), comprehension, 
analysis and inference? 
3. What kinds of interactive devices (e.g., visual 
and spatial query) are most effective for 
which kinds of tasks (e.g., anomaly detection, 
trend analysis, comparative analysis). 
4. What new evaluation methods, metrics, and 
measures are necessary for these new 
visualization methods? 
In visualization, we tend to deal with complexity 
through methodologies involving abstraction, 
aggregation, filtering, and focusing. Insights 
from natural language processing promise to help 
extract semantic information from text channels, 
to provide a richer, task-relevant characterization 
of the information space. Visualization can 
certainly benefit from other aspects natural 
language processing in achieving economy of 
interaction such as notions of context in 
reference (e.g., "fast_forward <the next week>") 
or relation (e.g., move "<enemy_icon> behind 
<Bunker Hill_icon>" in the currently focused 
display). An investigation of many applications, 
tasks, and interaction methods will be required to 
make progress in better understanding and 
answering these and other ~ndamental 
questions. 
61 

References 
MacEachren, A. M. Department of Geography, 
Pennsylvania State University, USA. Chair, ICA 
Commission on Visualization to be published in the 
Proceedings of the Polish Spatial Information 
Association Conference, May, 1998, Warsaw Poland 
(http://www.geog.psu.edu/ica/icavis/polandl .html) 
Buttenfield, B. and McMaster, R. (1991). Map 
Generalization: Making Rules for Knowledge 
Representation, Longman Scientific Technical, 
England. 
Exploratory Cartography: Maps as tools for discovery by 
Menno-Jan Kraat http://www.itc.nl/~arto/kraak/ 
Gershon, N., LeVasseur, Winstead, J., Croall, J., Pemick, 
A., and Rue, W. (1995). "Visualizing Intemet 
Resources," In Gershon, N. & Eick, S.G. (eds), 
Proceedings for Information Visualization T95 
Symposium, (pp. 122-128) IEEE Computer Society 
Press. 
Gershon, N. (1996). "Moving Happily through the World 
Wide Web." IEEE Computer Graphics and 
Applications, March 1966 (pp. 72-75). 
Gershon, N. and Eick, S. G. (eds.) 1997. Introduction: 
Information Visualization. IEEE Computer Graphics 
and Applications. 17 (4). July/August. 
MacEachren, A. M. and Ganter, J. H. (1990). A pattern 
identification approach to cartographic visualization. 
Cartographica, 27(2): 64-81 
Mani, I. and Bloedom, E. (1997). Summarizing Similarities 
and Differences among Related Documents. In 
Proceedings of the Fourteenth National Conference on 
Artificial Intelligence, (pp. 622-628). AAAI Press, 
Menlo Park, CA. 
Merlino, A., Morey, D. and Maybury, M. (1997). 
"Broadcast News Navigation using Story Segments", 
ACM International Multimedia Conference, Seattle, 
WA, November 8-14, (pp. 381-391). 
(http://www.mitre.org/resources/centers/ 
advanced_info/g04 ffonn/mmhomeext.html) 
Merlino, A. and Maybury, M. to appear. An Empirical 
Study of the Optimal Presentation of Multimedia 
Summaries of Broadcast News. In Mani, I. and 
Maybury, M. (eds.) Automated Text Summarization. 
Mitchell, R., Day, D., and Hirschman, L. (1995). "Fishing 
for Information on the Intemet." In Gershon, N. & 
Eick, S.G. (eds), Proceedings for Information 
Visualization T95 Symposium, (pp. 105-111) IEEE 
Computer Society Press. October. 
Morris, H. and Sonnen, D. (1997). Spatial Information in 
Data Warehouses: Business Potential and Risks, 
Bulletin #14753, International data Corporation, 
November 1997. http://www.itresearch.com/ 
Shneiderman, B., (1994) Dynamic queries for visual 
information seeking, IEEE Software 11(6): 70-77. 
Smotroff, I., Hirschman, L., and Bayer, S. (1995). 
"Integrating Natural Language with Large Dataspace 
Visualization," In Adam, N. and Bhargava, B. (eds), 
Advances in Digital Libraries, Lecture Notes in 
Computer Science, (pp. 209-224) Springer Verlag. 
Taylor, D. R. F., Ed. (1991). Geographic information 
systems: The microcomputer and modern cartography. 
Oxford, UK: Pergamon. 
