17-812 Seminar on Open Source Software Development
Instructor: Jim Herbsleb
University Units: 6
Wednesdays, noon to 1:20 in Newell-Simon Hall 3002.

We now have a course blackboard that will be used for communication and for submissoin of assignments.  This web page will continue to be the place to look for schedule, readings, assignments, and all other course information.  The blackboard provides us with a  class mailing list, as well as discussion forums. 

Description:
Open source software development is a recent phenomenon of great practical interest, as its practitioners challenge the basic practices and institutions that shape software development.  While discussions of open source software development are often bound up in ideology, it is a phenomena that has clearly become economically important, and may indeed shake up not only how we view software markets and intellectual property, but also how we think about software engineering tools, practices, and the coordination of technical work.  Open source software presents a fascinating sociotechnical phenomena which can only be understood and appreciated by bring multiple disciplines to bear. 

Some of the interesting questions include:
We will read and discuss a collection of recent research papers that address these and related issues in open source software. 

Prerequisites:
The course is a seminar appropriate for PhD students from a variety of disciplines.  Some familiarity with basic software development practices and tools is helpful but not necessary.

Schedule:
Date
 Readings
 Topics
 Questions
August 27
introduction, go over requirements, plan
September 3
 Herbsleb out of town -- no class

Make-up date
September 8
Room :
Newell-Simon Hall 2507
noon-1:30
 		Herbsleb, James D. & Mockus Audris.  "An Empirical Study of Speed and Communication in Globally-Distributed Software Development."  IEEE Transactions on Software Engineering, 29 (6), June 2003, pp. 1-14.

Grinter, R.E., Herbsleb, J.D., Perry, D.E.  "The Geography of Coordination: Dealing with Distance in R&D Work"  In proceedings, ACM Conference on Supporting Group Work (GROUP 99), Phoenix, AZ, November 14-17, 1999, pp. 306-315.

Raymond, Eric S.  "The Cathedral and the Bazaar."  Available at
http://www.catb.org/~esr/writings/cathedral-bazaar/, downloaded 7/10/2003.
 Background: geographic distribution in software engineering; insider view of OSS.

Discussion leader: Herbsleb
 1. The Grinter et al paper presents several "organizational models" used by Lucent for distributed software development.  As you are considering the problems identified in the Herbsleb & Mockus paper, which problems would you expect to be the most severe and the least severe for each model?  Explain.
2) Assuming for the moment that Raymond's views are accurate, which organizational model most closely resembles the development process he describes (or at least hints at)?  Are there aspects that fit none of the models?
September 10
 Halloran, Timothy J. & Scherlis, William L.  "High Quality and Open Source Software Practices."  Meeting Challenges and Surviving Success: 2nd Workshop on Open Source Software Engineering, International Conference on Software Engineering, Orlando, Florida, USA, May 19-25, 2002. 

Asklund, U. & Bendix, L.  "A Study of Configuration Management in Open Source Software Projects."  IEE Proceedings on Software," 149 (1), February, 2002, pp. 40-46.  		Open source development environments

Discussion Leader:
Vamshi
 In class on Monday, we discussed various approaches to coordination, including coordination by communication and coordination using various "coordination mechanisms," such as interface specifications.  Discuss the "walled server" concept and the configuration management tools and practices as described in the Asklund et al paper in terms of this distinction.  In other words, to what extent is communication fostered by the tools and practices discussed?  What other coordination mechanisms are introduced by the tools and practices?
September 17
 Mockus, Audris, Fielding, Roy T., & Herbsleb, James D.  "Two Case Studies of Open Source Software Development: Apache and Mozilla."  ACM Transactions on Software Engineering and Methodology, Vol. 11, No. 3, July 2002, Pages 309–346.

Moon, Jae Yun & Sproull, Lee.  "Essence of Distributed Work: The Case of the Linux Kernel."  First Monday, volume 5, number 11 (November 2000), URL: http://firstmonday.org/issues/issue5_11/moon/index.html.  		Collaboration in open source, how the work gets done

Discussion Leader: Alexandros
 Both of these papers briefly discuss how various aspects of open source processes and tools might be applied to commercial software development.  CITE EMPIRICAL FINDINGS from these papers in support of ONE of these three propositions: 
1) open source has little to offer commercial development, since OSS manages to avoid the hardest coordination problems,
2) open source development isn't really fundamentally different from commercial development, or
3) open source techniques (specify precisely which ones) could be of enormous benefit to many commercial developments.
September 24
 von Krogh, Georg, Spaeth, Sebastian, & Lakhani, Karim R.  "Community, Joining, and Specialization in Open Source Software Innovation: A Case Study."  Working paper, MIT Sloan School of Management, June, 2003. 

Cubranic, Davor.  "The ramp-up challenge in open-source software projects."  Meeting Challenges and Surviving Success: 2nd Workshop on Open Source Software Engineering, International Conference on Software Engineering, Orlando, Florida, USA, May 19-25, 2002. 		New members joining an open source community

Discussion Leader: Herbsleb 		Cubranic seems to conceive of the ramp-up problem as primarily one of gaining technical understanding of the software.  If this is accurate, to what extent would von Krogh et al agree?  To what extent could "joining" be facilitated by new or improved tools, if one adopts the von Krogh vew?  Be specific.
October 1
 Cubranic, Davor & Murphy, Gail C. "Hipikat: Recommending Pertinent Software Development Artifacts."  In proceedings, International Conference on Software Engineering, Portland, Oregon, May 3-10, 2003, pp. 408-418.

Orlikowski, Wanda J. "Knowing in Practice: Enacting a Collective Capability in Distributed Organizing."  Organization Science, 13 (3), May–June 2002, pp. 249–273. 		Tool-mediating mentoring; understanding capabilities of distributed organizatoins

Discussion Leader: Anupriya
 1. Applying the "knowing in practice" perspective to open source software communities, find corresponding activities within the open source context for the five practices she lists. For the 'face-to-face' interaction practice, you are free to interpret this literally as face-to-face or simply as several multifarious online interactions.
2. Based on what you have read in the Orlikowski paper, how successful do you think Hipikat is likely to be in practice? Does Cubranic propose a different perspective on managing knowledge or can Hipikat support some of the practices and activities you listed in 1?
October 8
 Herbsleb out of town -- no class

October 15 American Bar Association, "An Overview of "Open Source" Software Licenses: A Report of the Software Licensing Committee of the American Bar Association's Intellectual Property Section"  (undated).

Strasser, Mathias.  "A New Paradigm in Intellectual Property Law? The Case Against Open Sources."  Stanford Technology Law Review, 2001, http://stlr.stanford.edu/STLR/Articles/01_STLR_4.

Anonymous, "Halloween I."  Available at http://www.opensource.org/halloween/halloween1.php, downloaded 7/10/2003, reputed to be written by a Microsoft employee, outlining the threat posed by open source software and how to address it.  		Handling intellectual property concerns in OSS; one example of a possible industry response to the challenge

Discussion Leader: Patrick
 Which of Strasser's arguments support the position espoused in Halloween I?  Are Microsoft's interests (as represented in Halloween I) identical with the utilitarian interests that Strasser is trying to promote?  Explain. 
Make-up date
October 20
Room:
Newell-Simon Hall 2507
noon-1:30
 		Ghosh, Rishab A.  "Understanding Free Software Developers: Findings from the FLOSS Study."  Working paper, MERIT/Institute of Infonomics, University of Maastricht, June, 2003. 

Hertel, Guido, Niedner, Sven, & Herrmann, Stefanie.  "Motivation of Software Developers in Open Source Projects:  An Internet-based Survey of Contributors to the Linux Kernel."  Working paper, University of Kiel, Kiel, Germany, June, 2003.		 Surveys of OSS developers -- who are they, why do they participate.

Discussion Leader: Jeff
 1. Ghosh and Hertel approach the question of motivation from somewhat different theoretical perspectives -- explain. 
2. In what ways are their results consistent?  In what ways are they inconsistent? 
3. How might you explain the inconsistencies you identified?
October 22
 Schiff, Aaron.  "The Economics of Open Source Software: A Survey of the Early Literature."  Review of Network Economics, 1 (1), March 2002, pp. 66-74.

Lerner, Josh, & Tirole, Jean.  "Some Simple Economics of Open Source."  The Journal of Industrial Economics, 50 (2), June, 2002, pp. 197-234.

Leppämäki, Mikko &  Mustonen, Mikko.  "Spence Revisited - Signalling with Externality: The Case of Open Source Programming."  Working paper, University of Helsinki, May 28, 2003. 

Han, Roberts, Slaughter, & Fielding. " Why Do Developers Contribute to Open Source Projects? First Evidence of Economic Incentives" Meeting Challenges and Surviving Success: 2nd Workshop on Open Source Software Engineering, International Conference on Software Engineering, Orlando, Florida, USA, May 19-25, 2002.  		Motivation from an economic point of view.

Discussion Leader: Tayo
 1.  Productivity paradox
It is generally assumed that OS products are superior, performance wise, compared to proprietary alternatives (let us believe this to be true). It sounds plausible then to assume that OS developers are more productive because they are more motivated. This sounds contradictory given the fact that most developers are unpaid and no formal structure exists to militate against "shirking" or that ensures an "appreciable degree of alignment between a developer's interest and the organization's goals. Explain this paradox
2. Signaling effects
Is there reason to think that software companies, the OS community and OS developers will be more conscious of these effects in the near future? Explain.
October 29
 Baldwin, Carliss Y. & Clark, Kim B.  "Does Code Architecture Mitigate Free Riding in the Open Source Development Model?"  Working paper; Harvard Business School, June 1, 2003.

Shah, Solani.  "Understanding the Nature of Participation and Coordination in Open and Gated Source Software Development Communities."  Working paper, MIT Sloan School of Management, June, 2003.  		Motivation, contribution, under varying technical and IP conditions.

Discussion Leader: George
 1.The properties of game theory and financial options theory models are well known but the mapping between the OSS domain and the models is a new area.  Discuss the accuracy of their mapping from a variety of perspectives (e.g., motivations, participation style, knowledge, resources, opportunity cost).
2.Shah creates a detailed profile of open and gated source developers.  Is his data consistent with Baldwin and Clark's theories and assumptions? Would the options and game theory models predict different behavior from  open vs. gated source software developers?
November 5
 Kogut, Bruce & Metiu, Anca.  "Open Source Software Development and Distributed Innovation".  Oxford Review of Economic Policy, 17 (2), pp. 248-264.

von Hippel, Eric & von Krogh, Georg.  "Open Source Software and the "Private-Collective" Innovation Model:  Issues for Organization Science."  Organization Science, 14 (2), March–April, 2003, pp. 209-223.

Bessen, James.  "Open Source Software: Free Provision of Complex Public Goods."  MIT Center for Research on Innovation, July, 2002.
 Open source as user innovation

Discussion Leader:  Cheng
 1. Both von Hipple and Baldwin (last week) are attempting to address the "paradox" of apparent altruism. How do thier explanations differ?
2. One might argue that Besson's model of open source development is a special case of the von Hipple and von Krogh private-collective model.  Is this argument persuasive?  Explain the particular way in which this is true, or give specific reasons why it is false. 

November 12
 Franke, Nikolaus & von Hippel, Eric.  "Satisfying Heterogeneous User Needs via Innovation Toolkits:  The Case of Apache Security Software."  MIT Sloan School of Management Working Paper # 4341-02.

von Hippel, Eric.  "Horizontal Innovatoin Networks -- By and for Users."  MIT Sloan School of Management Working Paper No. 4366-02, June, 2002.

von Hippel, Eric & Katz, Ralph.  "Shifting Innovation to Users via Toolkits." Management Science 48 (7), July 2002,  pp. 821–833. 		Open source as an example of user toolkits, innovation networks

Discussion Leader: Anupriya
 von Hipple argues that there is a close connection between market segmentation and user innovation -- explain the connection.  What does the idea of innovation networks add to this?  Can one reasonably think about collections of open source tools as an end-user toolkit?  What strengths and  limitations do you see to this "user toolkit" view?
November 17
Make-up class, noon to  1:30, Wean Hall 4625
 		Benkler, Yochai.  "Coase's Penguin, or Linux and the Nature of the Firm." Yale Law Journal, 112, Winter 2002-2003.

O’Mahony, Siobhán.  "Non-Profit Foundations and their Role in Community-Firm Software Collaboration."  Forthcoming, Making Sense of the Bazaar: Perspectives on Open Source and Free Software O’Reilly & Associates Publications. 		organizational considerations -- neither firm nor market . .

Discussion Leader: Tayo
 1.What features of the software industry make it amenable to the commons - based peer production model? Do you believe this production process (the peer production model) occupies the 'middle ground' between markets and hierarchies? Explain.
2.In what way has the increased collaborration between firms and the community created the need for a more formal structure for the community? Discuss briefly how this is being achieved.
November 19
 Dedrick, Jason & West, Joel.  "Adoption of Open Source Platforms: An Exploratory Study."  Working paper, University of California at Irvine, June, 2003. 
"Open Source vs. Closed Source: Public Policies in the software market" by Stefano Comino† and Fabio M. Manenti
 Are commercial and governmental  organizations adopting  open source software?  When and why.

Discussion Leader: Srini		 
Competition occurs only in the segment of informed consumers" : Do you
agree that market forces are incapable of informing the consumer about OSS?
November 26
 Thanksgiving holiday -- no class

December 3
 no class, Herbsleb out of town (make-up Nov. 17)

December 10
 papers due


Written Assignments:
For each session, a written discussion of 2-3 pages will be due at the beginning of class.  Please prepare 2 copies, one to hand in, and one for you to refer to during class.  The written discussion will have 2 parts: part 1 will be a summary of each paper, no more than 1 or two paragraphs each.  In total, the summaries may be no more than 1 page.  The remainder of the discussion will be a response to discussion questions which will be provided each week.  Please submit your papers electronically (before class) as attachments (Word, pdf, ps, text) in the appropriate discussion forum on the cousre blackboard site. 

Class Discussion:
Each student will lead one or more class discussions.  It is expected that the discussion leader will prepare some brief  introductory remarks pointing out what is interesting and important, as well as wrong or just controversial in the readings, bolstering these views with background reading of additional materials where appropriate.  The leader will then facilitate a discussion of the issues.

Papers:
Papers for the course can be of two types -- literature reviews and proposals.  Literature reviews are exptected to be lucid, well-organized evaluative reviews of literature relevant to open source software development.  Depending on the student's interests and exertise, they may take any any of several disciplinary perspectives (e.g., software engineering, economics, sociology, organizational behavior, etc.), or focus on a particular problem or issue from multiple perspectives.  The other type of paper, the proposal, will frame an interesting research question, in the context of the existing literature, and propose a study or series of studies that will address the question.  The studies are expected to be realistic within reasonable time and resource constraints.

Initial ideas for papers should be discussed ASAP.  In any case, a preliminary proposal (1 page or less) is due September 19, and final topics must be chosen, in consultation with me, by Septermber 30. 

Method of Evaluation:
Students will be graded on the basis of written assignments (25%), class participation (25%), paper (35%), and leading discussion (15%).