The Chronicle: Colloquy Live Transcript

Open-Source Software: The Risks and Rewards

Friday, August 1, at 2 p.m. U.S. Eastern time

Are colleges on the right track by using noncommercial software such as Linux instead of more-expensive commercial alternatives? And should colleges themselves get involved in developing and licensing noncommercial, open-source code for specific higher-education needs? Or are such choices too risky for colleges, especially those that have relatively few programmers on their staffs?

A growing number of colleges see noncommercial software as a viable alternative to commercial software, in part because colleges have felt burned by companies that raise software prices without warning or stop supporting certain software altogether. In several key areas, colleges are now leading the way in developing software that fulfills specific needs of higher education. The most successful effort so far has been uPortal, which has produced open-source code for campuswide Web portals at more than 100 American colleges.

» Sharing the Code (8/1/2003)

Bernard W. Gleason is an Internet strategist at Boston College, where for 12 years he was associate vice president for information technology. Mr. Gleason has been an active participant in many higher-education organizations, including Educause and the Java Architectures Special Internet Group. Mr. Gleason serves on the steering committee of the Java group, which promotes the creation of open-source software for higher education. Mr. Gleason will respond to questions and commments on Friday, August 1, at 2 p.m. U.S. Eastern time. Advance questions are encouraged and may be posted now.

A transcript of the chat follows.

Florence Olsen (Moderator):
Hello, and welcome to Colloquy Live, The Chronicle's online discussion forum. I'm Florence Olsen, a senior writer for The Chronicle, and I'll be the moderator for the discussion. Today, we'll be talking about the open-source model for developing and licensing software and whether it fits the needs of colleges and universities.

Our guest today is Bernard W. Gleason, an Internet strategist at Boston College and its former vice president for information technology. He is also a technology consultant to the Boston Consortium for Higher Education. Thanks for joining us, Mr. Gleason, and for your article on open-source software, which appears in The Chronicle Review this week.

Bernard W. Gleason:
Good afternoon. I'm very pleased to be participating in the Live Colloquy on open-source business applications. Open source is a very important topic that is gathering a lot of interest within the higher-education information-technology community, but it is also a concept that is often misunderstood within information technology and the academic and business units.

Question from Charles Lowe, Kairosnews Editor:
In the recent Chronicle article, "Sharing the Code," the following doubts about open source are attributed to many college officials:

"What reasonable college, they ask, would give responsibility for critical software to volunteer programmers who get their paychecks from another college? The officials question how the work on open-source projects can be coordinated and accomplished by dozens of far-flung programmers working at different institutions."

How, then, is it possible that academics within the disciplines, spread across a multitude of institutions, find it possible to make knowledge within their fields? And that the open-source development model, Linux being a prime example, has successfully done the same by coordinating development internationally?

What is your response to those concerns and what open-source development practices would be best for academic-sponsored projects?

Bernard W. Gleason:
The skepticism that was voiced in the article was not in regard to Linux but rather major business applications, such as an accounting system, being developed as open source. Higher-ed business and information-technology professionals seem to be wedded to the idea that a product is only viable if an established software vendor supports it.

These skeptics are not necessarily shortsighted -- just unenlightened. They have not experienced the human dynamics, peer-review process, community spirit, and other characteristics of open-source initiatives. They also don't have a collection of references, such as a major business application built on the open-source model. Essentially the skeptics have not been cultivated, and there are not enough working examples to establish a comfort level. That is going to take time.

I think the first stage of open-source business applications will be a mix of proprietary and open source. For example, SCT is lowering the total cost of ownership for its enterprise-application suite by replacing a commercial portal engine with open-source uPortal. In this example, the customers remain comfortable that the software vendor is committed to support. You see this mix happening in other areas such as SAP offering support for MySQL, an open-source database.

Question from Kym Buchanan, NETC (nonprofit ed tech lab):
What are the leading arguments that dissuade higher-ed decision makers from choosing open-source software?

Bernard W. Gleason:
If we are talking about Linux, Apache Web servers, or other operating system or infrastructure components, I haven't experienced any strong dissent -- mostly positive backing. However, Chronicle articles focused primarily on open-source business applications in higher education. Most decision makers don't understand the open-source operating model. Their view is often that open source is free software that is not well supported. In the case of business applications, this view leads to concerns about the quality and reliability of the software and sustainability of the product and support.

In addition, there is concern that the soft costs -- such as installation, training, and service -- for open-source software may negate the cost advantages derived from free software.

Lastly, broad change is difficult to encourage in an academic institution. A case could be made that OpenOffice might be a better institutional choice than Microsoft Office, but attempting to orchestrate that change in an academic environment may prove to be an exercise in frustration.

Question from Richard Kirchmeyer, Western Kentucky University:
What do you think will be the consequences of the recent SCO Group suit against IBM and of SCO's demand for users to buy a Linux license? Does all of this make the long-term use of Linux more risky?

Bernard W. Gleason:
I am not a lawyer so I can't comment on the validity of the claim, but I hope that IBM will stand tall and not give in by making some form of settlement. That will just lead to more, similar tactics.

Question from Fred Beshears, University of California at Berkeley:
How should open-source development consortia collaborate with open-standards and open-specification consortia such as IMS? Incidentally, James Dalziel has written on this topic, open standards vs. open source in e-learning, at http://www.melcoe.mq.edu.au/res.htm

Bernard W. Gleason:
While most of the concerns regarding open source are focused on sustainability and support issues, you have hit on two less talked about points: open source and open standards and consortium activity coordination.

While it is not required for open source to be based on open standards, it does make a lot of sense if you want to gain the interest of a broad base of developers and commercial software companies. There may be a good reason for putting proprietary source code in an open repository, but I tend to think of these type of applications as being "openly available."

I also believe that open standards are of more basic importance than open source. Some e-learning vendors have a "we are the center of the universe mentality" and getting them to open their applications and provide open application programming interfaces has been a struggle. Doing that may be a sufficient first step that would reduce consideration of developing an open source e-learning application. The e-learning software vendors should continue to exist and play a major role in standards efforts.

I don't have a good answer for the issue of coordinating consortia activities. This is a real problem that can lead to redundant or competing efforts.

My involvement with open standards has been through the Java Architectures Special Interest Group (JA-SIG). In addition to W3C, IETF, and OASIS specifications, JA-SIG anticipates using the IMS, HR-XML, and PESC transcript standards in future projects and for integration. (Those are a lot of acronyms.) JA-SIG has maintained active communication with the OASIS WSRP and internationalization technical committees. We have been very pleased with the advice and assistance that we have received from them.

We are also following Internet 2 Shibboleth, the federal e-Authentication, and Liberty Alliance efforts. JA-SIG UK is working with the Joint Information Systems Committee (JISC) and we expect to follow its guidance as well.

My experience, albeit limited, has been to become involved in networking with interested parties in various groups and in tracking industry directions. This is getting more and more difficult because applications, in many cases, are no longer narrowly focused but rather fit within an overarching application architecture.

Question from Florence Olsen:
To what extent do personalities of the project leaders determine the success or failure of the open-source model for developing software?

Bernard W. Gleason:
The personalities of project leaders always have an impact on the success of a project. With open source it is no different except for very special qualification. In the open source community model, the project leader is less of an "in-charge" person and more of a benevolent dictator. At some point, the project leader and/or architect needs to act as the arbitrator.

Question from Patty Gertz, Princeton University:
Some of the anxiety generated by an open-source solution is not having anyone to call for support. Do you see partnerships between commercial companies and open-source initiatives solving that problem?

Bernard W. Gleason:
When open-source efforts, such as uPortal, are coupled with commercial companies such as SCT and Unicon, the customers win. Development happens faster and there is a built-in, on-going support structure for institutions that may not have the resources to support the product in-house. Post a uPortal question on the community list server and you will have multiple responses from around the world in a matter of a few minutes. But having commercial partners, such as SCT, also ensures long-term viability and enhancement of the product.

Question from Mark Olson, NACUBO:
We have seen some exciting and successful open-source initiatives in higher education, like the uPortal project, and others. Do you think that we will see open-source applications like, say, financial systems, emerge as candidates for open-source solutions? What would the issues be to move in this direction?

Bernard W. Gleason:
I think that you will see a financial system emerge as the first enterprise open-source administrative system. Unlike student systems, which can vary greatly in scope and customization from one institution to the next, financial systems are pretty stable, and they scale well across campuses -- large and small, public and private. While different FASB/GASB standards apply, there are common financial reporting, accounting, and processing requirements. Associations, like NACUBO, in this instance, should be prepared to play a key role in open-source initiatives by facilitating dialogue, reviewing standards and building community between chief financial officers and chief information officers.

Question from Jon Allen, instructional media+magic:
How do you see the use of open source in colleges and universities evolving over the next year?

Bernard W. Gleason:
I am hesitant to take out my crystal ball and to make predictions, but it appears to me that the industry is showing signs of migrating to open-source database management systems. As licensing and annual support costs continue to increase, it is reasonable to predict that commercial database management systems are going to be replaced by open source systems such as MySQL and PostgreSQL. The Linux operating system, Apache Web server, and Tomcat application server are already widely used.

Question from Tim Mentzer, Holy Cross:
Open source is obviously attractive due to the lack of costs involved, but do institutions who run many open-source applications need to have higher staffing levels to support them? What factors help determine staffing levels?

Bernard W. Gleason:
No matter what situation you look at in applying information technology there are going to be trade-offs of hardware and software costs and required personnel resources for support. In the case of open source the risk/rewards need to be calculated. For many small institutions with limited technical staff in both skills and numbers, the choice is often to use vendor-supplied or outside resources. Open source applications are going to have commercial support providers.

Florence Olsen (Moderator):
Keep the questions coming. If you have a question for Mr. Gleason, now would be a good time to ask it.

Question from Tim Mentzer, Holy Cross:
Have you heard of any open-source content management systems that have been successful at colleges or universities?

Bernard W. Gleason:
Content management along with other generic services such as workflow in my opinon is the next big opportunity area for open source. With regard to content management I would suggest that you take a look at what Columbia University is doing with their project CUCMS. I know that Columbia is in the process of obtaining grant funding to expand CUCMS into a very robust, open-source application.

Question from Justin Tilton, instructional media + magic, inc.:
How will open-source projects and products be supported?

Bernard W. Gleason:
We are going to see new types of companies and new business models emerging. Companies, such as Red Hat are providing support to open-source products based on packaging, technical support, and, in some cases, proprietary extensions. We have seen companies -- Unicon and SCT, for example -- providing support for uPortal, and I expect other open-source products will be similarly supported.

Question from Bob:
Bernie, with the open-source application movement, how are you addressing intellectual-property issues?

Bernard W. Gleason:
I am not sure that I understand the intent of the question. With proprietary software, the vendors are interested in protecting the investment that they have made in R&D and development. In the open-source model, the source code is freely available for all to use and modify. The advantage for the submitter of code is on the back end, where there is a broad community providing support and enhancements.

Comment from Aaron Hamid, Cornell University:
Let's also not forget that much of the infrastructure of the Internet and modern computing was developed as open source under academic programs, so not only is it feasible, but it already is a great proven success.

Question from Beth, small community college:
How easily does open-source software integrate with distance-learning course development software like Blackboard, Angel, WebCT, or SUNY Learning Network? I am guessing the information-technology department at the college would need to write code for integration.

Bernard W. Gleason:
It is important to understand the relationship between open source and open standards. Open standards are the key to achieving interoperability. Interoperability influences pricing of software and instructional content, and makes instruction and research more productive. The e-learning software products and companies that you mentioned can ease this problem of interoperability by opening their applications and providing application programming interfaces based on open standards. This is the purpose of the Open Knowledge Initiative (OKI), an open-source project at MIT.

Question from Allan MacDougall, large community college:
It may be heresy to ask this question, but is anyone taking an open-source applications approach based on a Microsoft infrastructure?

Bernard W. Gleason:
I don't think so. While most open-source projects are Java-based, there is nothing that I can see to prohibit Microsoft ASP code from being placed in an open-source library. I got an e-mail with the same question yesterday with regard to a faculty management system at the University of South Florida, which uses ASP and SQL Server and supposedly is being made available as open-source code.

Question from David Wright:
Doesn't the addition of corporate partners reduce the benefits of open-source software, namely the cost factor? The companies will need to generate revenue somehow. They won't support these efforts free.

Bernard W. Gleason:
Originally, I thought of open source as a direct competitor to vendor-supplied software. From my experience in adapting to the open-source community and getting to see an open-source project (i.e., uPortal) mature, I have come almost full cycle.

Commercial vendors are going to be key to the success of open-source applications. In the case of SCT, uPortal replaced a licensed commercial portal engine in its Luminis product, reducing SCT's costs. Hopefully, some of those savings are being passed on to customers.

Comment from Chris Coppola, r-smart:
Bernie, comment adding to your response to Beth: In addition, companies like ours are getting involved in open-source projects to help with those integration issues. The advantage of open-source projects in this regard is that as other institutions adopt the software and integration code gets written, it is shared so those institutions without the resources also have access to those improvements.

Question from Darren, Georgia Tech:
What can you share with us regarding your experience with campus portals in higher education and the use of open-source software vs. a product from a vendor such as Oracle?

Bernard W. Gleason:
In addition to being freely distributed, uPortal is open and is not dependent on a proprietary environment. Vendors such as Oracle make their portal available at an attractive price, but the catch is that you also need the rest of the Oracle product set. In addition commercial portal software is generic, that is, built all types of environments. uPortal is standards-based, but it was also designed for higher education. JA-SIG members like to say "by higher education for higher edcation."

Question from John, large private university:
Let's imagine a not-so-distant future in which an open-source consortium would have developed an integrated suite of the big "enterprise resource planning" (ERP) components, such as student records, financial aid, admissions, and would be going head-to-head with major vendor products. Those of us who've gone through proprietary ERP projects have heard the mantra of "vanilla" implementations, that is, don't customize particular applications too much or you won't be able to handle the next version upgrade.

How would this dependency be any different in an open-source environment, even with " access to the source code"? Wouldn't there be evolving "versions" that user institutions would have to implement and adapt to?

Bernard W. Gleason:
When we talk about replacing an ERP system with an open source equivalent the first question that is usually asked is, "How much will we save?" We all know that the cost of the software is only a small portion of the total cost of implementation. However, we also know that we are locked in to the vendor and have little control of future costs. Of course version control will still exist.

Florence Olsen (Moderator):
I see that it's time to wrap up this discussion of open source. I thank you all for participating, and Mr. Gleason, thank you for answering questions. Do you have any closing comments you would like to make?

Bernard W. Gleason:
The topic of open source is obviously very complicated. I don't think that any of us know all the "how and when" answers, but I think the open-source movement will be major factor in setting the direction for the design and development of application systems in the future. Thank you all for your participation.