9
EL.SEVIER Computers in Industry 29 (1996) 27-35 Engineering design notebook for sharing and reuse Jack Hong * , George Toye, Larry J. Leifer Center for Design Research, Stanford University, 560 Panama St., Stanford, CA 943052232, USA Abstract The concept of an electronic or digital engineering design notebook used by designers to capture information for reuse and sharing is becoming reality in many different flavors. Our development of PENS (Personal Electronic Notebook with Sharing) responds to observed designers’ needs for a lightweight tool that is facile enough to compete with paper notebooks in functionality. Each time informal design notes are entered in PENS, a project information web automatically grows. As it grows, selections can be incrementally shared with collaborators over the Internet’s World-Wide Web (WWW>. In an era where both network security concerns and distributed collaboration demands are growing together, PENS has the capability for information sharing that is independent of security firewalls. To evaluate the utility of the PENS notebook concept, a prototype was developed and used by 14 mechanical engineering design teams, many of which were composed of geographically distributed team members. Keywords: Electronic notebook; Informal knowledge capture; Distance collaboration; Virtual teams; Computer-supported collaborative work, Concurrent engineering: World-wide web; Internet 1. Introduction The SHARE 111 project at Stanford University is actively developing and assessing tools and services to enhance the design process for distributed collabo- rating engineering teams. The unifying SHARE vision is one in which distributed engineers individually collect design infonnation, and share it over the Internet via electronic notebooks. While paper may still be best for pers’onal notes, electronic notes are vastly superior for sharing and reuse. Electronically captured design data can be quickly modified and reused in simulatiom and analysis. By building on top of the Internet infrastructure, we gain world-wide access and an open system architecture. Unfortunately, existing notebook products have ’ Corresponding author. Email: [email protected] met with limited success in the engineering domain. Electronic design notebooks from the research com- munity require continuous network connectivity [2,3], require specialized (expensive) workstations [41, and are not usable in meetings or outside the office. They are rarely tested across an extended product develop- ment cycle and focus on proprietary systems with uncertain prospects for scaling up beyond the origi- nal testbed. This factor makes it difficult to transfer the innovations from academia to industry. As a development strategy, we believe that col- laboration environments must be iteratively used, redesigned and improved. Designers must be en- gaged with real time pressures that are encountered in an extended product design cycle. In this paper, we present the technology behind our PENS (Personal Electronic Notebook with Sharing) prototype, and how it has been deployed in a live engineering testbed. By building it on top of the Internet’s popu- 0166-3615/96/$15.00 0 1996 Elsevier Science B.V. All rights reserved SSDI 0166-3615(95)00052-6

Engineering design notebook for sharing and reuse

Embed Size (px)

Citation preview

Page 1: Engineering design notebook for sharing and reuse

EL.SEVIER Computers in Industry 29 (1996) 27-35

Engineering design notebook for sharing and reuse

Jack Hong * , George Toye, Larry J. Leifer Center for Design Research, Stanford University, 560 Panama St., Stanford, CA 943052232, USA

Abstract

The concept of an electronic or digital engineering design notebook used by designers to capture information for reuse and sharing is becoming reality in many different flavors. Our development of PENS (Personal Electronic Notebook with Sharing) responds to observed designers’ needs for a lightweight tool that is facile enough to compete with paper notebooks in functionality. Each time informal design notes are entered in PENS, a project information web automatically grows. As it grows, selections can be incrementally shared with collaborators over the Internet’s World-Wide Web (WWW>. In an era where both network security concerns and distributed collaboration demands are growing together, PENS has the capability for information sharing that is independent of security firewalls. To evaluate the utility of the PENS notebook concept, a prototype was developed and used by 14 mechanical engineering design teams, many of which were composed of geographically distributed team members.

Keywords: Electronic notebook; Informal knowledge capture; Distance collaboration; Virtual teams; Computer-supported collaborative

work, Concurrent engineering: World-wide web; Internet

1. Introduction

The SHARE 111 project at Stanford University is actively developing and assessing tools and services to enhance the design process for distributed collabo- rating engineering teams. The unifying SHARE vision is one in which distributed engineers individually collect design infonnation, and share it over the Internet via electronic notebooks. While paper may still be best for pers’onal notes, electronic notes are vastly superior for sharing and reuse. Electronically captured design data can be quickly modified and reused in simulatiom and analysis. By building on top of the Internet infrastructure, we gain world-wide access and an open system architecture.

Unfortunately, existing notebook products have

’ Corresponding author. Email: [email protected]

met with limited success in the engineering domain. Electronic design notebooks from the research com- munity require continuous network connectivity [2,3], require specialized (expensive) workstations [41, and are not usable in meetings or outside the office. They are rarely tested across an extended product develop- ment cycle and focus on proprietary systems with uncertain prospects for scaling up beyond the origi- nal testbed. This factor makes it difficult to transfer the innovations from academia to industry.

As a development strategy, we believe that col- laboration environments must be iteratively used, redesigned and improved. Designers must be en- gaged with real time pressures that are encountered in an extended product design cycle. In this paper, we present the technology behind our PENS (Personal Electronic Notebook with Sharing) prototype, and how it has been deployed in a live engineering testbed. By building it on top of the Internet’s popu-

0166-3615/96/$15.00 0 1996 Elsevier Science B.V. All rights reserved SSDI 0166-3615(95)00052-6

Page 2: Engineering design notebook for sharing and reuse

2s J. Hong et al. /Computers in Industry 29 (19961 27-35

lar World-Wide Web, PENS was immediately accessi- ble to testbed participants across the country with minimal infrastructure investment. Using a combina- tion of quantitative and qualitative data collected from our distributed testbed, we concretely examine our assumptions on tools design, and outline plans for further development.

2. Knowledge capture in collaborative engineering design

It is the custom in industry to conserve engineer- ing knowledge through formal documentation. Though this benefits redesign across generations of engineering teams, individual engineers often con- sider immediate benefits of documentation too small to justify the required effort. Consequently, docu- mentation is treated as a necessary evil to be done after the fact, once finer points of decisions and rationale have eroded from memory. While some engineers have such details written down informally in meeting notes and notebook entries, these tend to be personal accounts which are physically difficult to share with colleagues during the course of design.

In a survey of existing computer-supported col- laboration tools, we observed that most information exchange mechanisms are built on a metaphor of active discourse. Designers in a virtual co-located meeting scenario may use shared white boards, IRC (Internet Relay Chat), e-mail or usenet newsgroups to actively enter into synchronous and asynchronous conversation threads. These tools allow one to ex- tend meetings over space, time and offer opportuni- ties to record ongoing discourse.

However, as engineers know, meetings are only one part of teamwork. In fact, a significant portion of teamwork occurs in parallel outside of conference rooms. Indeed, the strategy of concurrent engineer- ing is to gain lead time by minimizing sequential dependencies and maximizing parallel task flows. To be effective in an engineering team, it is equally important to be aware of parallel tasks without en- gaging in prolonged discussions. This mode of infor- mal sharing is exhibited when teammates peek into one another’s offices, look over another person’s shoulder while they work, or ask one another, “What are you working on?”

Table 1

Comparison of discourse-based vs. reflection-based tools

Collaborative knowledge consists of both information generated

during collective discourse and individual reflection. E-mail and

conversations are conventional vehicles for discourse, while paper

notebooks are used for reflection. PENS fills the niche for reflective

knowledge in electronic collaboration.

Attributes\media type Discourse-based Reflection-based

Primary audience others self

Native tendency shared but not recorded but not recorded shared

Analog examples conversations personal notes

Internet examples e-mail FENS

2.1. The importance of monologues in collaborative work

In contrast with the discourse-based nature of meetings and conversations, we introduce the notion of rejlection-based information to characterize knowledge generated on a individual basis, as con- trasted in Table 1.

Discourse-based information transpires in the

course of communicating with others, in conversa- tions. While it is possible to record conversations, the inherent purpose is to exchange information and negotiate rather than archive. Conversely, reflection-based information is primarily intended for the individual. While it is possible to share reflec- tion-based information, it takes extra effort over its intended purpose of recording and facilitating thought. For effective collaboration, we propose that it is essential to integrate both knowledge generated collectively, as well as knowledge generated on an individual basis, with ratios of these modes of cap- ture variable with respect to domain and project phase.

3. PENS: Personal Electronic Notebook with Shar- ing

On its own, a paper notebook is little more than a bound collection of paper. Used as an informal engineering support tool, however, it becomes an extended memory of design process and rationale. PENS has been designed with the paper notebook in mind, and is first and foremost a recording tool for

Page 3: Engineering design notebook for sharing and reuse

J. Hong et al. / Computers in Industry 29 (1996127-35 29

the individual. To support ubiquitous note taking with the understandmg that one takes notes almost anywhere, our first PENS prototype has been designed for the mobile designer who uses a low-cost, portable Macintosh computer. As with any paper notebook, a designer with a notebook computer can take PENS to site visits, vendors, meetings or working sessions, and take notes on pricing, planning, action items, or any other relevant activities.

To compete with a paper notebook, it is important to support agile browsing of note pages. The PENS

note interface, shown in Fig. 1, has been designed to support quick data entry and browsing between notes.On a single 640 X 480 pixel screen, the user is able to see not only the note, but titles of adjacent notes, and the top-level category under which that note is classified. Shifting between categories is accomplished via mouse click or the left/right arrow keys. The up/down arrow keys cycle through notes within a single category. Automatic time stamping

provided for a historical perspective on the design notes. This has been revealed to be a significant need in prior research in electronic notebooks [4].

PENS steps beyond the paper notebook when the user decides to share its contents. Upon returning to the office, home, or any other place with a connec- tion to the Internet (via AppleTalk or modem), a single button-push allows both contents and structure of this individual notebook to be woven onto a group notebook located on the World-Wide Web-the shared equivalent of Fig. 1 is shown in Fig. 2. This allows all team members, managers or trainees who have been given access, to view the complete set of all project notes via any Web-browser from any- where on the Internet, whether or not they are PENS

users.

The concept of structure-mirroring is important. PENS goes beyond the simple sharing of a piece meal document to support sharing of a full network of documents and their semantic relationships. Each

Outline

About SHAREing Notebook:

WWW as a Collaborative [ PENS Design Consideration

ME2 10 Testbed

Usage Data

Future Work

Acknowlegementa

SECOND CRLL FOR PAPERS

WET ICE ‘95 Fourth IEEE Uorkshop on Enabling Technologies:

Infrastructure for Collaborative Enterprises

Apri I 20-22, 1995

Berkeley Springs, West Uirginia

The Concurrent Engineering Research Center (CERC) at West Virginia

University, with sponsorship from the IEEE Computer Society and

support from RIRRI, will conduct the Fourth Workshop on Enabling Technologies: Infrastructure for Collaborative Enterprises on Rpril

20-22, 1995, olt the Cool font Resort, Conference Center, and Health Spa

in Berkeley Springs, West Uirginia. We also have applied for the

cooperation of RCIl (RCtl SIGOIS lent their cooperation to last year’s workshop).

The workshop cloal is to focus on infrastructural issues related to collaboration in diverse application domains, ranging from engineering

to healthcare. Papers reporting survey, original research, design and development, clnd applications of enabling technologies for collaboration are sought in the following areas:

Fig. 1. The FENS note interface. The first FENS prototype has four top-level categories for note classification, automatic time-stamping, and a

listing of all files in the current category, displayed concurrently with the current note. Once the ‘POST’ button is pressed, changes in the

current notebook are translated into HTh4L and then transmitted to a pre-designated Web server.

Page 4: Engineering design notebook for sharing and reuse

30

PENS notebook contains notes organized in sequence and categories specified by the author. When sharing these notes with a larger audience, this structure is mirrored on the Web, but also systematically woven into the group network of notes.

3.1. PENS client-server architecture

Technically speaking, PENS is a custom off-line WWW authoring application that consists of both client and server software components.

3.1.1. Client sofhvare

The PENS client is a custom HyperCard stack with embedded TCP/IP code for handling file transmis- sion. A PENS author takes notes in predefined note templates, shown previously in Fig. 1. When a ‘POST’ command is invoked, PENS looks only at notes which have been updated since the last posting, carefully translates each new note into an HTML document, and packages these notes into a single MIME message. This packaged message includes master index files as well as directory depositing

PENS Client \: PENSIWWW Server

Fig. 3. Flow diagram of PENS transport mechanism. When the

‘POST’ button is pushed, the PENS client generates a new index

file and integrates with other new note files to create a composite

MIME (Multi-media Mail Extension) tile for transmission. Upon

receipt, the server-side PENS code separates the original MIME

message into multiple parts and deposits them into appropriate

directories.

information. A TCP/IP connection is then opened to an SMTP (Simple Mail Transport Protocol) server, and this MIME message is transferred to a predesig- nated mail account for processing by the PENS/WWW server, as depicted in Fig. 3. By using an e-mail transport mechanism, the encapsulated PENS

notes package easily passes through network security firewalls and other e-mail gateways. This facilitates cross-industry information sharing crucial to the dy- namic partnership requirements of multinational cor-

Fig. 2. WWW version of the PENS note shown in Fig. 1. Once the ‘POST’ operation is completed, the server-side PENS code updates

masl ter group index file and deposits all new notes in a pre-defined notebook area on the Web server.

the

Page 5: Engineering design notebook for sharing and reuse

J. Hong et al./ Computers in Industry 29 (1996) 27-35 31

porations. Collaboration servers can easily be set up for the project duration, with no need for significant changes in tbe legacy information infrastructures of partner companies.

3.1.2. PENS server The PENS server code co-resides on a machine

with a functioning World-Wide Web server. The PENS notes package is received by an automated mail-handling account and serviced by the PENS

server, implemented1 using ServiceMailTM 111 with custom TCL scripts. This PENS ServiceMail applica- tion looks at the subject header of the mail message, and determines (upon acceptance) where to deposit the set of PENS elem’ents as files. It then decomposes the package into individual HTML files for access on the World-Wide Web. After separating and de- positing these files.. it launches a post-processing program which ‘weaves’ the new index into the existing group index. A mail message is optionally sent back to the PENS author, notifying acceptance and completion of the posting.

3.1.3. Administration The PENS server code determines where PENS files

are to be deposited on the World-Wide Web server, so that distinct personal and group spaces are main- tained. User authentication is implemented via cus- tom group names, login and password scheme which does not require establishment of individual UNIX accounts. For the PENS user, these settings are speci- fied in the preferences card on the PENS Client, shown in Fig. 4. Because the shared contents are located on the Web server, all Web access security and privacy mechanisms are inherited.

In our testbed environment, e-mail and HyperMail [5] are used for discourse-based information, while PENS has been introduced to provide a forum for reflection-based infomrrnation. The inclusion of reflec-

tion-based information allows us to recreate circum- stances of incidental knowledge transfer in a virtual team. What would have been learned in a co-located office by ‘dropping by’ a colleague’s desk can now be glimpsed from the master PENS index for the group. With PENS, we attempt to recreate the infor- mal knowledge sharing environment of an open stu- dio, or a ‘virtual design loft’. How effective this mechanism works can only be determined through

21OWebSettings

mstrr tit* for yaw notn: Jrk’r WWW Gwalopmdnt Notea

W*Lop*,nm Pwrrwd

jhpnO/ ( .-. I T*rrn Nrn L SlIRRE v

Fig. 4. Preferences card for PENS client. Specific user login,

password and group information is specified on the PENS client.

This determines server transmission privileges and destination.

empirical testing. The following sections give some background on our testbed and describes how PENS

has been deployed.

4. A distributed engineering design testbed for PENS

ME210: Mechatronic Systems Design, is a gradu- ate-level engineering design class offered by Stan- ford University’s Department of Mechanical Engi- neering. Each year, ME210 provides an environment for 14 + three-person teams to tackle a wide range of industry sponsored projects. The educational goal of the course is to create a shared learning experi- ence in collaborative mechatronic product design and development. Each student team benefits directly from working on a unique industry-sponsored project with a $10,000 design/fabrication budget, and indi- rectly from the design experiences of peer project teams.

In Autumn 1994, ME210 took on tbe additional challenge of being offered concurrently to both tradi- tional on-campus students and off-campus Honors Co-op students through Stanford’s Instructional Television Network (SITN). This resulted in a hybrid class of local and remote students. Based on different

Page 6: Engineering design notebook for sharing and reuse

32 J. Hong et al. / Computers in Industry 29 (1996) 27-35

communication needs and life-styles, four classes of students are identified in Table 2 below: Traditional students who live in the dormitories; traditional stu- dents who commute; SITN students in the San Fran- cisco Bay Area; and out-of-state SITN students. In this distributed environment, paper-based communi- cations no longer reached the whole class in a timely fashion. Consequently, we moved to a completely Web-mediated classroom infrastructure.

While precedents exist for the support of Web-

mediated classrooms in distance education, they tend to be didactic lecture courses [6], and the deliver- ables are non-physical pieces of information. Resem- bling an engineering consultancy more than a lecture hall, ME210 defies the convention of typical class- rooms. In the ME210 environment, students teams, typically not facile with Internet technology, manage a real budget while working on real industry prob- lems, specifying, fabricating and delivering tangible artifacts with both co-located and non-co-located partners. The one distinguishing difference between our student teams and industry engineering teams is that immediate market pressure is replaced with an even more rigid academic time line - there is no slippage allowed. It is exactly this close parallel to

industry settings that makes ME210 a superior testbed for PENS.

To exercise the vision of an ubiquitous electronic notebook, each ME210 student is asked to use a Macintosh Duo 2 10 as their design notebook. Among other tools, this computer has Eudora [7], Netscape Navigator [8], and PENS on board with MacTCP [9], and MacSLIP [lo] for off-campus students. Each project team has its own home page under the ME2 10 class Web. Each project home page has links to a group HyperMail archive, individual PENS notebook

spaces and a master PENS index page for the group notebook. For confidentiality purposes, access to these pages are restricted to the Stanford campus network, except for remote students, coaches and sponsors who also have access to the World-Wide Web.

5. Testbed results

In 20 weeks of deployment between November ‘94 and April ‘95, close to 700 PENS postings were logged by the PENS server. Plotting the distribution of

these postings by project team, we find that three

Table 2

Overview of ‘hybrid’ design teams in ME210, academic year 1994-95

Engineering teams were composed of both traditional and honors co-op students. The highlighted teams were remote partnerships and the

most heavy users of PENS.

Team name Traditional students Television students Remote sponsor?

dorm commuters SF Bay Area out-of-state (default is yes)

GM-Window

FMC GM-Door

Lockheed

Raychem

Redwood

Ford

Pfizer Quantum

Seiko

3M

HP

JPL

1

1

3

2

2

1

2

2 2

2

2

3

2 1

2

1 3 (CMU)

1

1

1 n

1 n

1

1

2 n

1

2

4 (Mass.) n

1 (Oregon)

Total 23 9 8 8 48 [students]

Page 7: Engineering design notebook for sharing and reuse

J. Hong et al. / Computers in Industry 29 (1996) 27-35

120

60

Teams Fig. 5. Total number of ENS postings for each ME210 project team. More than 700 PENS postings were made over a 20 week test period.

Leading adopters are plotted in dark gray. The HP team was completely remote and did not have access to PENS.

Number

of

Posts

Fig. 6. PENS posting history for a single ME210 project team. Internal PENS posting history for the Pfizer team shows most contributions being made by a single member during the first two weeks, with sporadic use by other team members.

Page 8: Engineering design notebook for sharing and reuse

34 J. Hong et al. / Computers in Industry 29 (1996) 27-35

teams have made far greater numbers of PENS post- ings than others, as shown in Fig. 5, contributing to 70% of the total postings. Two of these teams are widely distributed. Members of the JPL project team are all remote SITN students distributed in Oregon and California, while FMC project members are distributed between Stanford University in California and Carnegie Mellon University in Pennsylvania.

7. Future work

The first generation of PENS has been successfully deployed in a 48-person testbed. In the interest of timely deployment, many planned features did not leave the drawing board for first stage implementa- tion, these include:

The internal PENS posting history of the Pfizer team is plotted in Fig. 6. This distribution reveals

that most of the postings for this team are con- tributed by a single team member (b) in the first few

weeks, while few postings were made by members (d) and (a). Internal plots for the two previous teams show a more even distribution.

(1) Graphics. We are reminded frequently by users that engineers need to see pictures, yet the first generation of PENS only supported text. Support for

rudimentary graphics, as well as intemote hypertext linking has been developed and will be included in our next test cycle. Still-frame video capture has now been enabled and is compatible with the sub $100 Connectix QuickCam.

6. Lessons learned

Based on data for all the teams, using e-mail and Web-based class surveys over the 20-week period, the following lessons were learned:

(2) Reading activity. Due to limitations of our current server implementation, we were not able to track individual PENS reading activity for each team to complement the data we have gathered for indi- vidual team posting. An improved user-tracking sys- tem will make this possible next year.

(1) More buy-in to PENS was seen from teams with

remote partnerships. In addition to PENS posting data, class surveys indicate that local teams do not perceive a compelling need for electronic communi- cations in light of being co-located. To further inves- tigate collaboration dynamics of distributed teams, we will consider requiring all ME210 project teams to have at least one remote student in the coming

year.

8. Closure

(2) A rough metric for inter-team work load was

obtained. In evaluating group engineering projects, it is often difficult for management to assess individual contributions without a heavy time investment. PENS

provides a mechanism for better understanding com- munication activities within a team and catching opportunities for intervention as the project pro- gresses. While this data can not be used exclusively, it provides a quantitative dimension which can be used to augment qualitative performance assess- ments.

We have described the concept, architecture and deployment history of a World-Wide Web-based electronic notebook for distributed engineering de- sign teams. While we have focused on the utility of this concept for engineering design, the group note-

book concept can be generalized to other domains, such as marketing, field survey work and construc- tion industry coordination. As the notion of a glob- ally sharable notebook is a new one for most of us, our strategy is to continuously deploy PENS in a variety of cross-domain testbeds and iteratively im- prove its design based on user feedback.

Acknowledgements

(3) PENS clients are needed for non-Mac plat- We thank the students and teaching team of forms. We have deployed PENS on the Macintosh ME210 over the past two years for putting up with platform to take advantage of Stanford’s highly Mac- our development efforts. Special thanks to Brian intosh-centric campus environment. As we increase Luehrs for making sure the 210Web stays ‘live’. enrolment of remote students, PENS clients must be PENS was developed under the SHARE project, carried developed for other popular platforms in the work out under the Department of Navy Contract NOO014- environment of remote students. 92-J-1833.

Page 9: Engineering design notebook for sharing and reuse

References

J. Hong et al. /Computers in Industry 29 (1996) 27-35 35

and Cl. Toye, “Mapping design information”, AAAI Work-

ill

PI

[31

[41

G. Toye et al., “SHABE: A methodology and environment for

collaborative product development”, IEEE Proc. 2nd Work- shop on Enabling Technologies: Infrastructure for Collabo- ratiue Enterprises, Morgantown, April 1993, pp. 33-47.

A. Burger, B. Meyer, C. Jung and K. Long, “The virtual

notebook system”, juroc. Hypertext ‘91 Cof, November

1991.

M. Guzdial, N. Rappin and D. Carlson, “Collaborative and

multimedia interactive learning environment for engineering

education”, accepted for ACM Symp. on Applied Computing ‘95. F. Lakin, V. Baya, D.M. Cannon, M. Brereton, L.J. Leifer

shop on Design Rationale Capture and Use, AAAI 1992,

San Jose, CA, June 1992.

[5] K. Hughes, HyperMail Documentation, http://www.eit.

com/software/hypermail/hypermail.html.

[6] D. Perron, “Learning on the WWW: A case study”, Elec- tronic Proc. 2nd WWW Co@, http://www.ncsa.uiuc.edu/

SDG/IT94/Proceedings/Educ/perron/perron.html

Chicago, IL, October 1994.

[7] Eudora, Qualcomm Incorporated, Copyright 1992-93.

[8] Netscape Navigator, Netscape Communications Corporation,

Copyright 1994-1995.

[9] MacTCP, Apple Computer, Inc., 1987-93.

[lo] MacSLIP, Hyde Park Software, Copyright 1993.