EEN Related Literature Survey

Table of Contents

• Electronic Design Notebooks
1. MECE/DICE (Multimedia Engineering Collaborative Environment) project - Enterprise Integration Technologies
   No publications yet. More information (including a demo) on the Web
2. The Electronic Design Notebook (vmacs EDN) - Centre for Design Research, Stanford University and The Performing Graphics Company
   1. [Lakin92] Mapping Design Information.
   2. [Leifer91] Instrumenting the Design Process.
   3. [Lakin89] The electronic design notebooks: performing and processing medium.
   4. [Sivard89] Conservation of Design Knowledge.
3. The Electronic Design Notebook (EDN/DICE) project - General Electric Corporate Research and Development
   1. [Uejio92] Capturing the Corporate Memory of a Product
   2. [Uejio91] An Electronic Project Notebook from the Electronic Design Notebook (EDN)
4. The Virtual Notebook System - Baylor College of Medicine, Rice University
   1. [Fowler94] Experience with the Virtual Notebook System: Abstraction in Hypertext.
   2. [Burger91] The Virtual Notebook System.
• Lightweight Notebooks
1. PENS - Center for Design Research, Stanford University [Hong95] Personal Electronic Notebook with Sharing.
• Product Books
1. Product books - Centre for Intelligent Systems, University of Plymouth, UK
   1. [Culverhouse95] Product Books.
• Pen Computing
  1. [Moran95] Implicit Structures for Pen-Based Systems Within a Freeform Interaction Paradigm.
  2. [Wolf92] Communication and Information Retrieval with a Pen-Based Meeting Support Tool.
• Pen-based technology
  1. [Gessler94] PDAs as Mobile WWW Browsers.
• Information Organization
  1. [Shipman95] Finding and Using Implicit Structure in Human-Organized Spatial Layouts of Information.
• User Interface Design
  1. [Moll95] Articulating a Metaphor Through User-Centered Design.
• Design Rationale
  1. [Gruber96] Generative Design Rationale
• Collaborative Design, Knowledge Sharing, Group Coordination Design Support
1. SHARE project - Stanford University, Enterprise Integration Technoogies
   1. [Kumar94] A SHAREd Web to Support Design Teams.
   2. [Toye93] SHARE: A Methodology and Environment for Collaborative Product Development.
2. ARPA Knowledge Sharing Effort. Including: The Palo Alto Collaborative Testbed (PACT) - Stanford University, Hewlett-Packard, Lockheed, Enterprise Integration Technoogies
   1. [Olsen94] Collaborative Engineering Based on Knowledge Sharing Agreements.
   2. [Cutkosky92] PACT: An Experiment in Integrating Concurrent Engineering Systems.
3. [Ostwald95] Supporting Collaborative Design With Representations for Mutual Understanding.
4. [Rogers95] The Use of an Automatic "To-Do" List to Guide Structured Interaction.
5. [Whittaker95] Back to the Future: Pen and Paper Technology Supports Complex Group Coordination.
6. [Kramer95] Supporting Design Activities in the Written Medium.
7. [Whittaker95] Back to the Future: Pen and Paper Technology Supports Complex Group Coordination.
8. [Reeves92] Supporting Communication between Designers with Artifact-Centered Evolving Information Spaces.
• Design Theory
  1. [Spillers93] Engineering Design, Conceptual Design, and Design Theory: A Report.

[Hong95] PENS

Hong, J., et al., Personal Electronic Notebook with Sharing, Full paper on the Web.

The authors present a lightweight notebook for designers. PENS supports ubiquitos notetaking, agile browsing, and sharing of notes through Internet. The prototype software runs on a Mac PowerBook and is essentially an off-line WWW authoring tool with browsing capabilities. The contents and structure of a PENS notebook can be woven onto a group notebook located on the WWW. At the current stage PENS supports text only notes. The authors see PENS role in recording of "thought-based information", that is, knowledge generated independently of conversational threads. The authors hypothesized that design knowledge is better conserved by providing frequent sharing of work in progress and thus less documentation may be needed. The hypothesis has not been proved or falsified yet.

[Kumar94] A SHAREd Web to Support Design Teams.

Vinay Kumar et al, A SHAREd Web to Support Design Teams, in Proceedings of Third IEEE Workshop on Enabling Technologies: Infrastructure for Collaborative Enterprises, April 1994, pp. 178-182

The authors used WWW as a medium for design information interchange, filtering, access and navigation.

[Toye94] SHARE: A Methodology and Environment for Collaboartive Product Development.

George Toye et al, A Methodology and Environment for Collaboartive Product Development, in Post-Proceedings of the IEEE Infrastructure for Collaborative Enterprises, 1993?

[Olsen94] Collaborative Engineering Based on Knowledge Sharing Agreements.

Gregory R. Olsen et al, Collaborative Engineering Based on Knowledge Sharing Agreements, in Engineering Data Management: Integrating the Engineering Enterprise. ASME 1994, pp.1-12

Authors propose modeling engineering collaboration as a network of interacting agents that encapsulate the capabilities of both users and their tools. Inter-agent communication includes both formal exchanges and informal communication among people. The contents of messages exchanges among agents are defined by adopting formal information sharing agreements (ontologies).

[Cutkosky92] PACT: An Experiment in Integrating Concurrent Engineering Systems.

Mark Cutkosky et al, PACT: An Experiment in Integrating Concurrent Engineering Systems, On the Web: http://www.eit.com/papers/pact/pact.html, 1992?

[Lakin92] Mapping Design Information.

Fred Lakin et al, Mapping Design Information, On the Web: click here, AAAI 92 Workshop on Design Rationale Capture and Use, June 21, 1992

Using computer-based tools (like vmacs-EDN) during the conceptual design phase can easily generate thousands of pages of electronic documents. The need thus arises for organization and navigation tools. The design map-maker system created by authors deals with this issue by employing map of design information. The map-maker system comes into play in two stages.
During the design information capture stage designer tags expressing key ideas parts on the notebook pages as she generates them. Tagging is performed with visual markers. This activity costs 20-30sec. per hour. Idea Tag Table is a machine and human readable outline containing tags denoting the current key ideas in the designers view. It takes about 20min. per week to update the table. To allow for translation between the designer-centered view (expressed by idea tags) and the organizational view (requirements) of a design, a translation table is created by a knowldege engineer (or by a designer). This translation table is based on the particular designer's notebook habits and design ideaas. The cost of creating table is a couple of hours monthly.
In the design information use stage, the map-maker translates user queries expressed in terms of requirements into a text-graphic object base query which returns a map containing the following information:

• a list of the relevant pages,
• a thumbnail summary of each page,
• a graphic hyper link to each page,

and also:

• the filter used to process user's requirement query
• the key features of the processed requirements,
• the text-graphic features inferred to be relevant to the key requirements,
• the text-graphic object-base query submitted to the login system.

Presenting the additional information makes it easy for the user to further refine the query.
In addition to the benefits of the present solution described above, the authors state in conclusion that their approach creates in the long term a knowledge base composed of large amount of informally structured text-graphics which is not fully interpretable by machines right now, but which in the future will provide designers with more insightful maps as machine interpretation becomes better.

[Leifer91] Instrumenting the Design Process.

Larry Leifer, Instrumenting The Design Process, International Conference on Engineering Design, ICED 91 Zurich, August 1991

Results of research and their implementation:

• To support design knowledge capture (which paper media is lacking) computer media should:
  • preserve the temporal structure of design activity which provides some ques about the design rationale. Design activities are thus recorded. The design acitivity record can be replayed, used as a database, or interpreted as a "language" and parsed to extract semantic features. A promising model for design rationale representation is provided by tracing co-evolution of two pages, one containing the current state of design problem and another containing the current state of design solution.
  • provide design capture agents (?) interacting with designers to annotate key elements and events. Designer labels text-graphics in EDN with "design idea tags". An agent (like the design summarizer - see below) can then gather and organize design material
  • provide paper-like agility in addition to advantages of computer-based media. The user must be free to express herself without reference to domain conventions (thus the importance of generic text-based summarizer).
• Designers must participate in knowledge navigation. The following tools facilitate the participation:
  • Summarizer - organizes a large set of text-graphic pages into a conceptual hierarchy by employing:
    • spatial parsing to discover important pieces of text (eg. titles) entered by the designer (hence designer participation) as a natural part of the design process
    • rules (?) defining important text in this domain
    • "design idea tags" entered by the designer
    The summarizer can be called "Designer aided summarizer". A fully automated summarizer might miss the designers "real intent".
  • A domain specific summarizer was developed in addition to the generic text-graphis one.

[Lakin89] The electronic design notebooks: performing and processing medium.

Fred Lakin et al, The electronic design notebooks: performing and processing medium, The Visual Computer, Springer Verlag 1989, 5:214-226

• Paper notebook provides freedom and agility to designers
• Designers during conceptualization process are performers. EDN has to facilitate this performance before it provides any further processing of captured data
• Domain specific text-graphics are processed according to "visual languages". The system automatically selects a visual language for a given object and then spatially parses the arrangement of the pieces according to the syntax of the language. General purpose sketches are included without any processing.

[Sivard89] Conservation of Design Knowledge.

Cecilia Sivard et al, Conservation of Design Knowledge, 27th Aerospace Meeting, AIAA'89

• Goal: creation and maintanance of an accurate record of a design throughout its lifecycle. Electronic notebook as a design knowledge acquisition system.
• EN provides designers with interaction methods enabling them to communicate information.
• Implementation is based on Vmacs which is processing visual objects, performing analysis and calculations on text or graphical objects and annotations of objects with the time of their generation.
• Two ways of capturing design knowledge are explored.
  • Page Set Summarizer - automatic generation of a self-explanatory summury of the notebook pages. Summary of each page is generated using the "visual langiage". The set of summarized pages is then used to automatically generate an overview of pages.
  • Design Rationale Inferencer - capturing the rationale for a design change by comparing how the design meet requirements before and after the change. Assumption is made that the rationale for a design change is either to satisfy previously violated constraints or to "better" statisfy these constraints.
• Information about design alternatives is captured in form of drawings. Designers create sketches using predefined features. Conclusion: To make sketches understandable to computer an extensive mechnism for labelling the components and their connections must be developed.

[Uejio92] Capturing the Corporate Memory of a Product.

Wayne H. Uejio et al, Capturing the Corporate Memory of a Product, ?Mar 27, 1992

• WHAT of design (ie. data) is recorded in EDN by entering or pasting data, or by providing live links to other applications
• WHY of design (ie. context and design decisions) is captured by logging user actions. Users can later add notes and comments to this log.
• Two implementations of EDN: 1. based on Aster*x, and 2. based on FrameMaker

[Uejio91] An Electronic Project Notebook from the Electronic Design Notebook (EDN).

Wayne H. Uejio et al, An Electronic Project Notebook from the Electronic Design Notebook (EDN), In Proceedings of the Third National Symposium on Concurrent Engineering. Concurrent Engineering Research Center, West Virinia University, February 1991

• The proposed EDN architecture is based on the assumption that no single application can be the basis for theEDN.
• The system has a hierarchical structure with the project notebook at the top and note pages in personal notebooks at the bottom. Project members maintain ownership of thier notebooks.
• Organization of notes in notebooks and navigation among them is based on a calendar approach. Each notebook has also directory.
• Notes are put under configuration control in the shared information model
• A note page is a dynamic document which can have hyperlinks to other documents. Data from other docs can either be put directly (translated screen capture, in which case, however, all the structure of an included document is lost) into an EDN note page, or an active link to another application can be created.
• User can "publish" a note page by creating its copy in the project notebook. This copy is indexed by keyword phrases generated automatically from the shared information model scheme of attributes. A hyperlink to the published note page is also created.
• The indexes and links provide another way of navigation/organization of information.
• EDN services include a parser of the files from other applications which automatically creates hyperlinks and checks for variables to be put into the shared information model (but it is not explained how it is acomplished)
• Implementation was based on FrameMaker

[Fowler94] Experience with the Virtual Notebook System: Abstraction in Hypertext.

Jerry Fowler et al, Experience with the Virtual Notebook System: Abstraction in Hypertext, Proceedings ACM 1994 Conference on Computer Supported Cooperative Work, October 1994 pp. 133-143

The authors descirbe their experiences with the VNS after it has beed used in diverse environments and after it became a commercial product. The VNS provides environment for collaborative authoring/viewing of hypermedia documents. VNS, as a hypertext system, shares the basic charactersistics of the Dexter hypertext reference model. At the same time VNS extends the Dexter model, as the authors does not find it sufficient for their task.

Hypertext features of the VNS include:

• Links: explicit one-way connections between pages
• Dangling-links: allow for construction of link networks (eg. TOCs, outlines) to which data could be added after development of the outline. (not permitted in the Dexter model)
• Global-links across multiple libraries of objects.

1. Libraries contain notebooks
2. Notebooks contain pages
3. Pages can contain four types of components:
   1. Text - can be entered from keyboard, by importing, cutting&pasting.
   2. Images - either imported or captured from screen.
   3. Navigational links - as described above. They are implemented in the form of buttons with a linked page name or with a text label.
   4. Action links - invoke user-defined system commands. Examples of use: modifying the page contents, exporting the contents of a text component to a word processor and importing it back after changes have been made.

• Each notebook has an owner.
• Owner can grant access rights to the notebook to other authors.
• Authors can "subscribe" and "unsubscribe" to notebooks what provides the change notification mechanism in the VNS. An author subscribing to a notebook is notified as soon as another author has published annotation about the change to the notebook.
• Pages can be locked while an author is working on them.
• to prevent conflicting modifications of components, selected(?) components are automatically locked. After component changes are commited they are automatically propagated to all active clients.

• Strenghts
  • successfully facilitates collaboration,
  • supports multiple media types,
  • authoring and viewing/browsing tools are the same,
  • client are ported to multiple platforms
• Weaknesses
  • limited to the notebook metaphor
  • inability to share data across notebook hierarchies
  • no version control

[Burger91] The Virtual Notebook System.

Andrew M. Burger, The Virtual Notebook System, in Proceedings Hypertext'91, pp. 395-401

• Goal: support for collaborative work of scientific groups.
• The following issues of collaboration are addressed: ownership, integrity, concurrency (locking), managing change, and multiple viewpoints.
• VNS architecture is based on a client/server model. Different layers of the VNS communicate using TCP/IP.
• User interfase is based on X-windows.
• The VNS is a hypertext system.
• Elements of notebook structure.
  • resizable, scrollable page
  • resizing objects on pages (scroll bars)
  • links and references to other pages as icons and text on the page
  • objects can be moved among pages
  • page links can be duplicated and dragged to other pages or relinked to other objects
• Navigation and search in notebook.
  • outline browser. Based on a tool for casting hierarchical structures in outline form. Notebook outline browser displays page and link maps as an outline.
  • attribute/value tagging of objects for specialized (parametric) search and retrieval (presentation filtering)
  • WAIS-VNS integration facilitates indexing and free text search within notebooks. The result of the search is displayed in the browser for further navigation.
• The VNS was used to create specialized applications that dynamically construct or modify notebooks or provide unique views of the notebooks space.

[Spillers93] Engineering Design, Conceptual Design

Spillers. W.R., et al., Engineering Design, Conceptual Design, and Design Theory: A Report. in de Vries., M.J., et al. (eds.) Design Methodology and Relationships with Science, Kluwer Academic Publishers 1993pp. 103-120

The authors present review of engineering design, elements of design theory and results from their case studies of conceptual design stage. Below only the latter is described since it is of our main interest.
Seeking to answer the question: how to support designers from various domains during the conceptual design phase, the authors conducted a study to determine characteristics of and representations used in this phase of design. Four expert designers from different domains including an artist (a sculptor) were interviewed. The results in the form of answers to questions are summarized below:

• The role of sketches: part of the thinking process, allow to overcome preconceptions about the design, used for communication (collaboration) purposes, useful to select among alternatives.
• The form of sketches: initial sketches are very rough, sometimes difficult to use because of lacking third dimension.
• One vs. many initial design ideas: it seems to depend on designer. Some designers start with one idea work on it, while other first generate (often in a group) several alternatives.
• Design details: details emerge only after rough sketching, often after considering several alternatives, and sometimes after the design is modeled.
• Features of computer tool for conceptual design: allow to rough sketch, also in three dimensions, manipulation of 3D rough sketches (note: there is a research conducted in this field at Alias), visualisation, cleaning up rough sketches, simultaneous comparision of alternatives, as details emerge the representation (rough sketches) should be transformed to illustrate their effect.

[Culverhouse95] Product Books.

P.F.Culverhouse, Product Books: archiving for information use and reuse, a replacement for CAD databases, Journal of Design and Manufacturing 5:13-24,1995.

The paper addresses the issue of human access to design information in engineering companies. A proposed solution is through the use of a structured constraint-based hypermedia system called "product-books".
The author proposes that design documentation be extended to embody engineering logbooks, design schematics, bills of materials and engineering constraints into one formally structured document that constitutes the design information for a product. Suggested structure for these documents has form of electronic books. The books are called "product books", and are based on a developed by the author object-oriented text and graphics database. Being object-oriented all text and graphics are accessible by automatic enquiry methods as well as manual ones. A set of product books is generated for each product during its life. This set is called a product encyclopedia. The encyclopedia consists of 3 product books each holding different aspects of product design information. Author makes a claim that by introducing a predefined, fixed structure of product book chapters, information enquires are made easier.
Critique. The described system stores many aspects of design information facilitating in this way design reuse. Thus the design WHAT is captured. It does not store, however, design decisions which led to the final design, nor it records design rationale. In particular, Product Book 2 describes a product's conceptual design. This book contains the early stages of a design, but it records them only in their final form. Furthermore, even though this book contains initial product concepts and solutions, it does not capture the design decisions lying behind them, it does not capture the design WHYs.

[Gessler94] PDAs as Mobile WWW Browsers.

Stefan Gessler, Andreas Kotulla, PDAs as Mobile WWW Browsers, Proceedings of the "WWW Fall'94" conference in Chicago.

In this paper authors present a WWW client for Newton. First, future applications of PDAs as global information browsers are pictured. Second, the possible architectures of a WWW client implementation on a PDA are outlined. Authors then describe current state of their development of WWW browser for Newton. The current WWW client is not generic in a sense, that it still requires some preprocessing done by the specific host (DECalpha is the only host supported). The main issues which had to be addressed in t his implementation were due to the limitations of Newton's hardware. These resulted in the following drawbacks in the browser's functionality:

1. Slow response time due to the limited memory and limited bandwith of serial communication.

2. Longer documents have to be split-up into smaller parts on the UNIX host due to the limited size of data blocks (32k) which Newton is able to store.

3. User has to scroll a lot to see the whole text due to the small display size.

Authors continue work on their project and plan to incorporate improvements in the Newton-WWW as soon as faster connections, larger storage capabilities of Newton, and the new HTTP specs become available.

The above research is of our interest because it presents similar design issues as in the NewtEEN project, and because a WWW client can be used to interface IKB server.

Installed alpha version has been tested with DECalpha over serial connection. Two main features which are missing: forms and security are not supported.

[Moran95]

Implicit Structures for Pen-Based Systems Within a Freeform Interaction Paradigm.CHI'95 Proceedings ACM 1995, pp.487-494

[Wolf92]

Communication and Information Retrieval with a Pen-Based Meeting Support Tool. CSCW'92 Proceedings ACM November 1992, pp.322-329

[Shipman95]

Finding and Using Implicit Structure in Human-Organized Spatial Layouts of Information.CHI'95 Proceedings ACM 1995, pp.346-353

[Moll95]

Articulating a Metaphor Through User-Centered Design. Design Briefing. CHI'95 Proceedings ACM 1995, pp.566-572

[Ostwald95]

Supporting Collaborative Design With Representations for Mutual Understanding. Doctoral Consortium. CHI'95 Proceedings ACM 1995, pp.69-70

[Rogers95]

The Use of an Automatic "To-Do" List to Guide Structured Interaction. Short Papers. CHI'95 Proceedings ACM 1995, pp.232-233

[Kramer95]

Supporting Design Activities in the Written Medium. Doctoral Consortium. CHI'95 Proceedings ACM 1995, pp.61-61

[Reeves92]

[Whittaker95]

Back to the Future: Pen and Paper Technology Supports Complex Group Coordination. CHI'95 Proceedings ACM 1995, pp.

[Gruber96]

Generative Design Rationale: Beyond the Record and Replay Paradigm. in Moran, Thomas P., Carroll, John M., eds. Design Rationale. Concepts, Techniques, and Use. Lawrence Erlbaum Associates Inc., 1996, pp.323-349

The authors argue that record and replay systems are not sufficient for design rationale recording, because they cannot answer all the designer's questions. They argue for generative methods of design rationale construction or inference...