CrystalWeb - A Distributed Authoring Environment for the World
Wide Web
Ralph Peters
Fraunhofer Institute for Computer Graphics,
Wilhelminenstr. 7, Darmstadt, Germany
peters@igd.fhg.de
Christian Neuss
Fraunhofer Institute for Computer Graphics,
Wilhelminenstr. 7, Darmstadt, Germany
neuss@igd.fhg.de
- Abstract:
-
Human cooperation and interaction is highly dependent on
efficient means for communication. While the World Wide Web
provides powerful tools for information retrieval, it still
lacks support for document authoring. Distributed authoring of
documents requires not only the possibility to jointly view and
edit a document, but also the ability to create and share
annotations. This article examines the issues behind a
integrated authoring environment, and introduces a prototype
for information access, telecooperation and distributed
authoring in a wide area network.
- Keywords:
-
Authoring Environment
Today's use of personal computers in office and information
management increasingly involves telecommunications networks.
Back in the 80s, personal computers were mainly used for text
processing and spreadsheet calculation. In the 90s, we
experience a trend towards inter-personal computing, allowing
for communication via electronic mail, file transfer and
information access. There are an estimated 50 million PCs in
use in businesses in Europe, half of which have network access
capabilities.
A large part of computer supported office work involves
documents. Being able to electronically access them in a wide
area network is an important element of telework. By providing
an easy to use interface for information browsing as well as a
flexible mechanism for hypermedia access, the World Wide Web is
the ideal basis for a distributed information retrieval and
authoring system. It allows for transparent transfer of
heterogeneous documents and provides a point and click metaphor
for interlinked hypertext.
Its ease of use and the availability of browsing software
on a variety of platforms make it an interesting candidate
for implementation of a corporate information system.
However, the Web currently lacks a number of features
necessary for distributed authoring and management
of documents and document related meta-information.
If a team of authors are working together on a common set of
documents (e.g. technical manuals), their write access to
these documents has to be coordinated.
Thus, the document repository has to be augmented with
a document database functionality that provides transaction
mechanisms and access control.
One problem problem here is the fact that
structural changes to the documents require manual maintenance
of hyperlinks. On a conventional Web server documents are
created and maintained as physical files, with links based on relative
or absolute pathnames. Whenever a file is being moved or renamed, links
to that document become inconsistent.
Instead of their physical access path, links to
documents could be based on an abstract, persistent
naming scheme. Following this approach, a server
would use a directory service for mapping an abstract
name (which could be encoded into the document URL)
to the actual physical location.
Finally, an integrated environment for distributed information authoring
and retrieval not only has to support remote document access
and resource discovery, but must also allow for making digital
annotations and facilities for holding virtual conferences with co-
workers. In order to create an integrated authoring environment, Web
browsers/editors have to be augmented with sophisticated
teleconference and annotation tools to support distributed
authoring of hypermedia documents.
We will analyze the requirements for distributed authoring in a
Wide Area Network (WAN), and introduce CrystalWeb, a
prototypical implementation of an integrated environment for
distributed authoring.
This section discusses the World Wide Web as a basic element of
an integrated authoring environment as well as existing tools
and applications for collaboration in a wide area network
(WAN).
The World Wide Web (WWW) [LCG92] is an information retrieval
system on the Internet with an estimated 30 million individual
users. It has created the first true global hypermedia network,
with more than 3000 multimedia databases of hyperlinked
documents. The World Wide Web is conceived as a seamless world
in which information from any source can be accessed in a
consistent and simple way. It was originally developed in 1989
at the European Particle Physics Laboratory CERN as a means of
sharing research information through the organization. By the
end of 1990, the first World Wide Web browser software was
introduced on a NeXT machine. The World Wide Web has since then
experienced tremendous growth and widespread popularity. For
example, the World Wide Web has been mentioned by CNN, the Wall
Street Journal, the Economist, Fortune magazine and the New
York Times. The Web is already heavily used: Over 10 Gigabytes
of World Wide Web data are transferred over the NFS backbone
every day.
The great popularity the Web currently enjoys results from the
availability of easy to use "point-and-click"
browsing tools which make access to information a task as
simple the use of a word processor. The most famous among them
is probably Mosaic, which was developed by the National Center
for Supercomputing Applications (NCSA) at the University of
Illinois at Urbana-Champaign. Web browsers have been developed
for a variety of platforms including MacIntosh and Intel PCs,
which makes the Web not only a vendor but also a platform
independent standard. Its ease of use makes the Web an
attractive vehicle for electronic publishing and for conducting
various business transactions. It offers encryption,
authentication and data security for the exchange of sensitive
documents. Tools for client side searches [BrPo94] and creation
of indexes in a server local manner [NeHo94] as well as
centralized [Kos94] establish mechanisms for resource
discovery. Document formatting and hypertext information is
stored in an SGML derived format called HTML (Hypertext Markup
Language). It includes logical structures like headings,
paragraphs, lists and emphasis e.g. by using a different font.
The World Wide Web also allows for seamless integration of
other information services. For example, a gateway program can
provide access to an relational database system which contains
product specifications.
In order to provide computer support for collaboration,
teleservices and tools have been developed which supply an
audio/video connection. These services can basically be seen as
a replacement for communication over telephone, enhancing it
with a number of features like the ability to establish a
conference and provision of a real time video connection.
Systems like the Multi Media Collaboration Service [AD93]
within the BERKOM project, SunSoft`s ShowMe or Intel`s ProShare
provide not only an audio/video conferencing tool, but also
realize a joint viewing functionality. Stand-alone joint
viewing systems like Shared X [Alt90], XMX [Baz90] and XTV
[AWF91] are often X windows based, an overview can be found in
[Jon93]. Their input can be shared via a so called floor
control system, which basically defines from which machine the
application is controlled. By passing the input focus to
another conference member, cooperative work with the
application is possible. Additional support for cooperation is
provided by displaying a so called shared pointer or
telepointer, which makes the mouse movements of one party
visible to others.
However, audio/video connections and joint viewing are not the
only services that are useful in a conference situation. The
ability to make sketches and notes on a whiteboard is an
important part of a conference. This paradigm is being used by
whiteboard systems: The conference participants can make
drawings and notes on a shared drawing area. An interesting
feature of these group sketching tools is the ability to load
an image as a background into the drawing area and thus make
annotations on this image. This approach is taken by systems
like SketchPad [SNR9][NS91], Wscrawl [Mal94] or GroupSketch
[RG92]. Furthermore, some audio/video conferencing systems like
SunSoft`s ShowMe also have an integrated whiteboard.
Using a shared whiteboard system for annotating and discussing
documents can be done by taking the output of a non-shared
document processing system as a background image. The output of
the document system is captured by taking a "snapshot" of the
screen output, and placed as an image in the background of the
conferencing tool. It is then possible to make graphical and
textual annotations, which can be shared with other conference
participants. The problem with whiteboard based systems is that
only a small fraction of the document (the page which is
visible in the snapshot) can be annotated. This view is static;
existing systems do not reflect changes in the document, or
allow for navigation in the document. Furthermore, it is not
possible to have conference members actually edit the document.
This causes the need for annotating a living document which
requires an approach entirely different to current whiteboard
systems [PNB94].
A distributed information system must provide methods for
access and navigation in a remote document repository. For
example, an author can use a hypermedia browser to access
various documents like specification sheets or technical
documentation, go through the minutes of a previous meeting, or
comments and suggestions from co-workers.
Since the World Wide Web allows for seamless integration of
heterogenous information sources, it can be used to provide an
integrated information system. However, the World Wide Web as
it exists today lacks support for authoring. State of the art
servers use the file system as data repository, requests for
documents are mapped directly to file system paths. A server
does not know about a document until it actually serves it,
especially, it has no control over the creation and
modification of documents. Normally, documents are modified by
loading them into a standard text editor and saving them back
to the file system, the server is unaware of this process.
In order to support team members with a mechanism for reviewing
documents, an annotation facility has to be provided.
Annotations can include sketches, text or even a digital
recording of a voice remark. Since the annotations are
displayed right on the document, making corrections and
incorporating changes can be done in a very convenient way.
Finally, an audio/video connection to one or more co-workers
can be opened in case a question arises which needs discussion.
To support this discussion, the document being viewed by the
author can simultaneously be displayed on the other
workstations (shared viewing). Finally, after saving the
document to the remote document repository, the new version is
immediately available to the other team members.
In order to support distributed authoring in a wide area
network, an authoring environment must provide the following
functionality:
- navigation through the document space
Through various navigation techniques, access to electronic
documents has to be provided. This includes retrieval through
hyperlink traversal, direct addressing in a unique namespace,
and retrieval through evaluation of a search query (e.g. based
on author name or keywords).
- secure retrieval
When retrieving a document over a wide area network, security
issues have to be considered. A means of authentication is
necessary in order to make sure the person that has requested a
document is authorized to read it. Furthermore, if the
documents contain sensitive data, they have to be encrypted in
order to ensure privacy.
- transaction and versioning mechanism
In an environment where multiple authors have write access to a
document, a transaction mechanism has to ensure that a document
is not simultaneously opened for writing by more then one
author, otherwise changes will be overwritten by another author
and thus get lost.
- support of annotations and other related
information
In order to support the discussion about documents (a necessary
requirement in a multi author scenario), a mechanism for
storing and accessing annotations has to be provided.
Annotations can e.g. be simple text, digital audio recordings,
or graphical sketches made directly on top of a document
- conferencing and shared editing
Audio/Video conferencing tools allow exchange of ideas and
comments in a way that is very close to a telephone call or a
discussion in a meeting. Shared editing allows multiple users
in remote locations to simultaneously view and edit one
document.
Figure 1 shows the software components
of an integrated environment for distributed authoring: A
distributed document database system stores hypertext files,
annotations and secondary data, and makes them accessible in a
wide area network. The World Wide Web serves as distribution
mechanism: A hypertext browser and a information resource
discovery frontend are used for identifying and retrieving
documents. The authoring system component consists of a
Desktop Publishing (DTP)
system, an annotation facility which provides an additional
sketching facility, and an audio/video communication tool.
Including a mechanism for audio/video conferences
and for sharing an application between several users
allows for online discussions without the need for
physical presence. In order to be able to use the DTP system in
a distributed environment, its can be shared among co-workers
through the use of a sharing component. To close the circle of
information access and authoring, a finished document is
published via a hypertext converter and stored in a proper
format in the database.
FIGURE 1. System architecture
Not only allows this approach for keeping a single
source document for both electronic and paper version
of a file, it also offers the features of a professional
document processing system such as built in drawing tools,
import filters for pictures, spell checking, and an integrated thesaurus.
An authoring environment must handle a document as a compound
entity: while the HTML version of a document is adequate for
electronic access, it lacks layout information required for a
printable version. Thus, both version must be available,
either as separate versions or generated dynamically through
a filter. Furthermore, in order to support electronic
annotations, a digital version of the sketch or the voice
remark is also a required part of the document.
FIGURE 2. Synchronous and asynchronous editing
The document database treats the electronic document as
an opaque entity. Depending on the type of action an
author performs, different views to this entity are being
used. Figure 2 shows the information flow
for synchronous and asynchronous work on a document.
On the left side, the asynchronous case is being
shown. The author requests write access to a
document. If write access is granted, the document
then is locked and transferred to the author. The author
can make changes and create or modify annotations.
After saving the document, another author or reviewer
can in turn access and edit it. There is no direct
communication between the authors, all information
transfer is handled via HTTP.
In case of synchronous work, the situation is different.
After retrieving the document, it must be edited in
a cooperative session.
The clients communicate directly with each other via the
sharing component, which multiplexes the in- and output
of the DTP system and provides a floor control mechanism.
This section describes CrystalWeb, an environment for
distributed authoring created at the Fraunhofer Institute for
Computer Graphics. It augments the functionality of the World
Wide Web with a database system to manage hypertext documents,
a professional DTP system which allows for creating both paper
and hypertext versions of a document, and a set of tools for
collaboration and communication. The Fraunhofer Institute for
Computer Graphics has been working on World Wide Web
applications since spring 1993, the focus has been in the area
of information resource discovery and authoring environments. A
prototype for distributed authoring was demonstrated at the
Telework 94 conference in Berlin. It combines a Web
editing tool with a facility for making digital document
annotations [PNB94]. A Web browser is used for information
retrieval and access to the document pool. When a document has
been reviewed by a co-worker, the Web browser not only opens it
in the document processor but also starts up an annotation
viewing facility. In order to be able to discuss questions with
other team members, a conference can be initiated. The
annotated document can then be viewed simultaneously by all
conference participants. For publishing we use the document
processing system WebMaker (for information see
http://www.cern.ch/WebMaker). WebMaker is a configurable
converter of FrameMaker documents to the World Wide Web, which
enables authors to publish simultaneously both the printed and
the WWW versions of a document.
The following sections will show how the CrystalWeb components
form an authoring environment. The elements of the authoring
process include content based retrieval of documents, viewing
and editing of digital annotations, as well as holding
teleconferences with co-workers.
The first step is the authentication process: A username and a
password has to be entered in order to be able to access the
system. Then, a document search query is being initiated (see
Figure 3). The query is being entered
into with the help of a so called forms interface. Forms are an
element of HTML which allow for providing the user with an
interactive interface where values can be entered and simple
checkboxes and selections can be made. After finishing a forms
interaction, the data gets encoded and is then transferred to
the Web server for evaluation. The CrystalWeb search interface
allows for retrieving documents based on title, author, or
through free text searches that can optionally be restricted to
document fields. As result of the query (see Figure 4), a list of hyperlinks which lead
to the corresponding document is being returned.
FIGURE 3. Query forms interface
Selecting the corresponding hyperlink retrieves the document.
For those documents the user has write permissions, a virtual
link is dynamically created and inserted into the hypertext
document. Selecting this link calls up a CGI gateway which
marks the corresponding document as 'checked out for writing',
and returns the documents FrameMaker source. The Web browser
then starts a FrameMaker session with this document.
Finally, after finishing the changes to the document, it has to
be transferred back to the server in order to make the new
version available to the other team members. Since a well
defined interface for the 'publishing' process is being
employed, the document repository can not only make the new
version available (and store the previous version under a
different name), but also initiate an update of the index for
free text searches.
FIGURE 4. Query result
Distributed authoring requires communication between team
members and computer-based systems which provide an 'interface
to a shared environment to support groups of people engaged in
a common task or goal' [EGR92]. An electronic conferencing
system enables the replacement of face-to-face meetings by
computer supported conferences between geographically dispersed
persons. It is responsible for the administration of the
teleconference itself as well as for additional components like
audio/video. In this workstation integrated conference system a
desktop publishing application will be distributed to all users
via a window sharing component to support shared viewing and
editing. Such a joint viewing or editing tool displays the
graphical output of an application simultaneously on multiple
computers in a network, and thus gives all participants the
same view of e.g. a document. The input to the shared
application is managed by a so called floor control system; an
author has to ask for the floor (the right to address an
assembly) in order to perform an action. Figure 5 shows a screendump of CrystalWeb
authoring environment.
FIGURE 5. Distributed authoring in a teleconference
(click figure to view larger screendump)
Besides the direct telecommunication with the other authors, in
a distributed authoring scenario it is important to be able to
communicate ideas by making freehand sketches and annotations.
The shared desktop publishing application should be overlaid by
the annotation facility like a set of transparencies. Each
participant gets a private transparency where he or she can
sketch comments in a unique color. Sound annotation allows an
immediate addition of succinct comments to a text sector. If
the comments consist of longer passages, text annotation should
be offered in form of a simple editor. On demand, these
passages may be integrated into the document later, e.g.
through a clipboard mechanism.
A document which has been reviewed and contains annotations is
retrieved in the same way, the only difference is that not only
FrameMaker is being started, but also the annotation facility
of CrystalWeb which loads the annotations and displays them on
top of the FrameMaker client.
The annotation facility augments the DTP system with a
mechanism for including e.g. sketches and short sections of
text. However, annotating a living document poses some problems
for the annotating system which have to be taken into account.
The problem with the electronic document is the fact that any
navigation in it, or even scrolling up or down a page will
cause the position of an annotation to lose its meaning. Even
worse, the content itself is of course also subject to change.
Whenever a reformatting takes place, the position of an
annotation will have to be moved. This could be triggered by
inserting or deleting of text, but also by changing attributes
like font size or page margins. Figure
6 illustrates the problem:
even scrolling backward or forward in the text causes the
annotation to lose its meaning, since the relation to the text
parts get lost. The same can
happen due to the insertion or deletion of text.
Annotations on dynamically changing documents are possible
through introducing invisible tags in the underlying
system, and using them as anchors for annotation mark-up.
FIGURE 6. Problems with scrolling or modifying text
The author can edit the document and make changes according to
the annotations. For displaying and editing annotations, the
CrystalWeb environment uses CrystalPad [PNB94] which was
developed in a project financed by Siemens. For shared
authoring a desktop publishing (DTP) application is distributed
to all users via a window sharing component. By this CrystalWeb
supports shared viewing and editing in a multi media
teleconferencing environment as well as making annotations by
overlaying an application with a virtual drawing area, which
can be used like a transparency film.
Although the ability to directly write sketches and short
notes onto a document is an important feature in an
editing process, other types of annotations must also
be supported.
Sound annotations allow for an immediate addition of
succinct comments to a text sector, while longer comments
can be made in the form of an attached text file.
On demand, these passages may be integrated into the
document later, e.g. through a clipboard mechanism. This also
applies to image annotations illustrating a document section.
The World Wide already supports
the concept of external viewers: a hyperlink can
point to a sound or an image as well as to another
hypertext file. Upon selecting the corresponding link,
an external viewer (depending on the file type) is
being started to display the object.
CrystalWeb
allows for inclusion of multimedia annotations via
the annotation facility. The annotation facility also
handles the distribution of the multimedia annotations
during a conference session. Upon writing the document to the
distributed database, an HTML version is being generated in
which multimedia annotations are converted to hyperlinks.
Figure 7 shows the design of hypermedia
annotations. Together with a "simple" annotation a link managed
by the annotation facility is created. Internally, this link is
transferred into the proper annotation in the form of sound, text,
images, video, or other media. With the conversion of the document
external files for each hypermedia annotation and corresponding hyperlinks
in the HTML-File are automatically generated
FIGURE 7. Hypermedia Annotations
In order to provide an integrated environment for retrieval and
distributed authoring of documents, teleservices and
information systems have to be combined into an integrated
environment. The World Wide Web provides powerful and easy to
use tools for information retrieval, but lacks support for
document authoring.
An integrated environment for distributed information authoring
and retrieval not only has to support remote document access
and resource discovery, but must also provide digital
annotations and facilities for holding virtual conferences with
co-workers.
We have analyzed the requirements for an integrated environment
and presented a solution for combining electronic access of
information, digital annotations and teleconferencing tools.
The ability to make annotations onto a living document while it
is being modified provides a natural way of discussing changes
with others while working them into a new version. Since these
annotations can be stored together with the document, they can
also be used outside a conference situation.
Both document and annotations can be distributed via the World
Wide Web and are accessible by any team member. Future work
will aim at support of workflow control and use of a database
management system for the document repository.
The work on the annotation facility CrystalPad was performed in
a project financed and supported by Siemens ICK / Siemens ZFE
München and Siemens ZFE Saarbrücken.
-
[AWF91]
- Hussein M. Abdel-Wahab and Mark A. Feit. XTV:
A
Framework for Sharing X Window Clients in Remote
Synchronous
Collaboration. X11R5 contrib distribution,
contrib/clients/xtv,
August 1991.
- [AD93]
- Altenhofen, M.; Dittrich, J.;
Hammerschmidt, R.; Käppner, T.; Kruschel, C.;
Kückes, A;
Steinig, T.: The BERKOM Multimedia Collaboration Service,
Proceedings of the ACM Symposium on MultiMedia, The
Association
for Computing Machinery, New York, 1993.
- [Alt90]
- Michael
P. Altenhofen. Shared X. Technical report, NESTOR Project,
Digital
Equipment Corporation, CEC, Karlsruhe, 1990.
- [Baz90]
- John Bazik. XMX: An X Protocol Multiplexor,
Version
1.0. FTP: wilma.cs.brown.edu: pub/xmx.tar.Z, Dezember
1990.
- [BrPo94]
- De Bra, P.M.E.; Post, R.D.J.: Information
Retrieval in the World Wide Web: Making Client Based
searching
feasible, Proceedings of the First International World-
Wide Web
Conference, CERN, Switzerland, May 1994
- [EGR92]
- Ellis, C.A.; Gibbs, S.J.; Rein, G.L.:
Groupware: Some issues and experiences. Communications of
the ACM,
34(1): 39-58, January 1992.
- [Jon93]
- Oliver Jones.
Multidisplay Software in X. In The X Resource, Issue Six.
O'Reilly
and Associates, Inc., 1993.
- [Kos94]
- Koster, Martijn:
ALIWEB --- Archie--Like Indexing in the WEB, Proceedings
of the
First International World-Wide Web Conference, CERN,
Switzerland,
May 1994
- [LCG92 ]
- Berners-Lee, T.; Cailliau, R.; Groff,
J.; Pollermann,B.: World-Wide Web: The Information
Universe, in
Electronic Networking: Research, Applications and Policy
Vol.1
No.2, Meckler, Westport CT, Spring 1992
- [Mal94]
- Pal S.
Malm. The unOfficial Yellow Pages of CSCW. Technical
report,
University of Tromso, January 1994.
- [Kos94]
- Koster,
Martijn: ALIWEB --- Archie--Like Indexing in the WEB,
Proceedings
of the First International World-Wide Web Conference,
CERN,
Switzerland, May 1994
- [NS91]
- S. Noll and M.G. Schendel. Cooperative
sketching in a network environment for the automotive
industry in
europe. In Proceedings of the Eurographics 1991, Technical
Report
Series, Vienna, 1991.
- [PNB94]
- Peters, R; Neuss, C.;
Bernhard, M.: CrystalPad: Shared Authoring with
Annotations,
Postersession Proceedings IWACA '94 Workshop, Heidelberg,
1994.
- [RG92]
- Mark Roseman and Saul Greenberg. GroupKit: A
groupware toolkit for building real--time. In Jon Turner
and
Robert Kraut, editors, Proceedings of the Conference on
Computer--Supported Cooperative Work, Toronto, pages 43 --
50. The
Association for Computing Machinery, New York,
1992.
- [SNR91]
- M.G. Schendel, S. Noll, and J. Rix.
Distributed SketchPad System: A tool for cooperative
sketching in
a network environment. In Contribution to COMICS--
Workshop,
Toulouse, June 1991.
.
