5
JOURNAL OF ORGANIZATIONAL EXCELLENCE / Autumn 2001 © 2001 John Wiley & Sons, Inc. CURRENTS: TECHNOLOGICAL DEVELOPMENTS Mike Uretsky Mike Uretsky is co-director of the NYU Center for Advanced Technology in New York. * * * 79 OPEN SOFTWARE: CAN YOU AFFORD IT? CAN YOU AVOID IT? One of the hotter topics currently under discus- sion is the open source debate. In brief, it focuses on whether source code, the building block that is later converted into executable code that actually drives computers, should be made readily avail- able and at no cost. This is important because ac- cess to source code lets users make modifications to fix bugs to customize programs to meet their individual needs. Craig Mundie, senior vice president and chief software strategist of Microsoft, defined the Microsoft position in a speech given at the Stern School of Business at New York University. He said that at a time when we are all moving toward ubiq- uitous computing, a global firm like Microsoft, with its fiduciary responsibility to stockholders, cannot afford to make its software freely available for any- one to modify. Instead, it would share source code with organizations that agree to adhere to standards that Microsoft feels are consistent with addressing the needs of a large market. Richard Stallman, founder of the Free Soft- ware Foundation (FSF), defined the other end of the spectrum in a speech on May 29, at the Cou- rant Institution of Mathematical Sciences of NYU —a speech heralded as the rejoinder to Microsoft. (Videotapes of this presentation are available from: NYU Center for Advanced Technology, 719 Broadway, New York, NY 10003.) Stallman ar- gued that source code always should be free and that any limitations to free access were funda- mental limits to freedom of speech. He argued for the use of the GPL (general purpose license) of the FSF , which modified copyrights into copylefts—protecting ownership, but making a requirement to allow free copying. The current debate covers the spectrum from strategies for meeting the needs of a global soft- ware business to the conviction that there should be no obstacles to personal freedom. There are, of course, positions and groups in the middle. This important debate touches on a number of sensitive issues. How should innovators be com- pensated for their efforts? Which strategies are best for building a corporate information infrastruc- ture? Which approach leads to the most effective

Open Software: Can You Afford It? Can You Avoid It?

Embed Size (px)

Citation preview

Page 1: Open Software: Can You Afford It? Can You Avoid It?

Currents: Technological Developments

JOURNAL OF ORGANIZATIONAL EXCELLENCE / Autumn 2001

79

JOURNAL OF ORGANIZATIONAL EXCELLENCE / Autumn 2001

© 2001 John Wiley & Sons, Inc.

CURRENTS: TECHNOLOGICAL DEVELOPMENTS

Mike Uretsky

Mike Uretsky is co-director of the NYU Center for Advanced Technology in New York.

* * *

79

OPEN SOFTWARE: CAN YOU AFFORD IT?CAN YOU AVOID IT?

One of the hotter topics currently under discus-sion is the open source debate. In brief, it focuseson whether source code, the building block that islater converted into executable code that actuallydrives computers, should be made readily avail-able and at no cost. This is important because ac-cess to source code lets users make modificationsto fix bugs to customize programs to meet theirindividual needs.

Craig Mundie, senior vice president and chiefsoftware strategist of Microsoft, defined theMicrosoft position in a speech given at the SternSchool of Business at New York University. He saidthat at a time when we are all moving toward ubiq-uitous computing, a global firm like Microsoft, withits fiduciary responsibility to stockholders, cannotafford to make its software freely available for any-one to modify. Instead, it would share source codewith organizations that agree to adhere to standardsthat Microsoft feels are consistent with addressingthe needs of a large market.

Richard Stallman, founder of the Free Soft-ware Foundation (FSF), defined the other end of

the spectrum in a speech on May 29, at the Cou-rant Institution of Mathematical Sciences of NYU

—a speech heralded as the rejoinder to Microsoft.(Videotapes of this presentation are availablefrom: NYU Center for Advanced Technology, 719Broadway, New York, NY 10003.) Stallman ar-gued that source code always should be free andthat any limitations to free access were funda-mental limits to freedom of speech. He arguedfor the use of the GPL (general purpose license)of the FSF, which modif ied copyrights intocopylefts—protecting ownership, but making arequirement to allow free copying.

The current debate covers the spectrum fromstrategies for meeting the needs of a global soft-ware business to the conviction that there shouldbe no obstacles to personal freedom. There are, ofcourse, positions and groups in the middle.

This important debate touches on a number ofsensitive issues. How should innovators be com-pensated for their efforts? Which strategies are bestfor building a corporate information infrastruc-ture? Which approach leads to the most effective

Page 2: Open Software: Can You Afford It? Can You Avoid It?

JOURNAL OF ORGANIZATIONAL EXCELLENCE / Autumn 2001

80 Mike Uretsky

economic development? But this debate may wellbe a precursor of a larger and, in a sense, moreinteresting one. Growing use of the Internet is notjust a technological phenomenon. It is being ac-companied by a great number of changes andstrains upon legal, social, and economic systems.

SELECTED ARGUMENTS IN BRIEF

Traditional software distribution, like other pub-lishing and intellectual property pursuits, is basedon an assumption that someone owns the con-tents and the developer has a right to both pro-tect the quality of the creation and recover thefruits of his or her labor. In the commercial soft-ware business, this generally means that innova-tions are protected by various legal means andthat buyers get only compiled code along with auser license. The product has presumably beentested before release, and supplier-owned sup-port teams resolve problems. As a practical mat-ter, some corrections are made immediately,while others await the issuance of patches or newproduct releases. If the supplier is a large com-pany, it has the potential for issuing a de factostandard that others can use as a platform in or-der to reach mass markets.

The open source movement argues that mak-ing source code readily available (but not neces-sarily free) has significant advantages. It frees youup from a false sense of security, from dependenceon suppliers (who may or may not be willing orable to supply timely fixes), and makes it easier todistribute corrections. Here are a few of the morespecific alleged advantages:

• Open source code is higher in quality be-cause it is freely available and, hence, sub-ject to a much wider level (and perhapsbetter quality) of peer review than wouldotherwise be the case.

• Because open source is of better quality, itis more reliable and, therefore, a superiordevelopment platform.

• An open source development environmentis inherently collaborative, thus openingthe potential for contributions and assis-tance from widely dispersed experts.

• Open source development makes it pos-sible for programmers (and by extrapola-tion, the company) to use the besttechnologies at each point, rather than acomposite of proprietary tools that mightbe available within a given developmentshop. This assumes that, since commercialdecisions are affected by a wide range ofconsiderations, the more readily availablecommercial platforms are really a portfo-lio of good and fair components.

• Corrections are more readily available andtimelier, because support is sought fromthe entire development community.

• Development cycles can be much shorter,because one has available both a libraryof source code and a widespread develop-ment community.

• Building systems from pieces of opensource code (which can be entire modules)is really no different than assembling anyother type of product—for example, acomputer or a car.

There is an extensive library of open sourcecode and programs available. Although the list islong, it is useful to note that it includes ApacheWeb Server, various versions of Linux, HTML, HTTP,Perl, Python, Sendmail, and Enhydra ApplicationServer—all major development tools. Many ofthese are basic building blocks used in the Internetand in internal software development.

One should not fall into the trap of thinkingthat the open source movement is just a collectionof techies. Large companies like IBM and Oraclesupport it, to varying degrees. It is used, in part,by such groups as Fermilabs and NASA.

This leaves us with the questions raised byBrian Behlendorf in the book Open Sources(O’Reilly and Associates, 1999).

• Is this really a new way of building soft-ware?

• Is each of the successes in open sourcesoftware a fluke of circumstance, or isthere a repeatable methodology to all this?

• Why on earth would I allocate scarce fi-nancial resources to a project where my

Page 3: Open Software: Can You Afford It? Can You Avoid It?

Currents: Technological Developments

JOURNAL OF ORGANIZATIONAL EXCELLENCE / Autumn 2001

81

competitor would get to use the same code,for free?

• How reliant is this whole developmentmodel upon the hobbyist hacker or com-puter science student who just happens toput the right bits together to make some-thing work well?

• Does this threaten or make obsolete mycompany’s current methods for buildingsoftware and doing business?

HISTORICAL UNDERPINNINGS

In some sense, there is nothing new about the defacto concept of open source. The developmentof computer technology has been built on a cycleof openness followed by closure, followed byopenness, and so on.

The early days of computing were built onhardware technology. Users, generally academics,developed systems software, tools, and applica-tion packages. They readily shared their innova-tions through a combination ofmanufacturer-supported user groups and academicsocieties. This approach accelerated advances insoftware development and created the tools neededto speed up commercial use of computer technolo-gies. Standard (stabilized) versions of the basictools and selected applications were made avail-able through hardware manufacturers.

The general feeling during this period was thatthe hardware was the asset of value. Companieslike IBM protected their hardware markets by de-veloping and protecting proprietary interfacesbetween computers and peripherals. With thegrowth of IBM’s marketing position, the courtseventually forced placing the interface informa-tion in the public domain.

De facto standardization of software may haveslowed down innovation, but it certainly provideda basis on which others could build. The empha-sis shifted to the development of applicationsneeded to run companies.

This same general pattern was present in theintroduction and development of the personalcomputer. The early personal computer was es-sentially an open device. Users could write pro-grams, develop new circuit boards, etc.

Manufacturers made selected software availableto registered users of their equipment. The initialopenness of the development process provided alow-cost selection of software and hardwareboards. Proponents of open source argue that thegrowth of the personal computer market stemmedfrom an open system that allowed each user todevelop his or her ideal operating platform. It ismore likely, however, that the growth was due tomany other more important factors—legitimiza-tion of the market resulting from IBM’s entry, themarketing power of IBM, etc. From the perspec-tive of our argument, de facto standards are setwhenever there is a need to reach broader mar-kets and provide stable application platforms.

Although the Internet was originally a gov-ernment-sponsored activity, many of the devel-opments came as the result of open sharingbetween scientists having shared interests orneeds. Then, however, there was a need for stan-dards that would lead to platforms. Once againwe had a move toward stability, this time througha combination of pure competitive forces andindustry consortia. The competitive forces in-cluded the widespread seeding of the browsermarket by AOL, as well as Microsoft’s launch ofInternet Explorer. The international industry con-sortia, it should be noted, grew out of the need toprovide relatively stable platforms, recognitionof impending convergence of technology andindustry lines, and a need to share large finan-cial risks in a rapidly growing market.

A FORESHADOWING OF THEMACRO FUTURE

We are already seeing a shift in business attentionto the Web as a paradigm for interrelationships.The focus on Napster is just an early indicator ofthe focus on peer-to-peer networks, where systemsadvertise their wares—products and services—andintelligent agents then access them. There is alsocurrent talk about the formation of electronic com-munities of various kinds and with widely vary-ing geographical membership.

In this sense, one can observe that current dis-cussions regarding the client-server-operating sys-tems environments and open source are a precursor

Page 4: Open Software: Can You Afford It? Can You Avoid It?

JOURNAL OF ORGANIZATIONAL EXCELLENCE / Autumn 2001

82 Mike Uretsky

of something much larger—something that isevolving at this very minute. Nobody would ar-gue that Microsoft Word is an application. What,then, is the conceptual difference betweenMicrosoft Word and a company website? Noth-ing! Microsoft Word is an applications packagemaking it possible for you to carry out a selectedset of word-processing functions. A companywebsite is an applications package making it pos-sible for you to carry out a selected set of searches,retrievals, etc. In other words, the website existson a more meta-version of a stand-alone applica-tions package.

The focus is thus shifting to the network as anew knowledge space. Windows, as a client-serverfor moving data back and forth, may be on the vergeof extinction. Instead, we may be evolving toward anetwork operating system with a focus on moremacro knowledge-related functions—a merger of thetraditional operating system, search and metasearchengines, analytic engines, and presentation spaces.

The preceding discussions may be an earlyskirmish in a much more interesting war. A hintof this is found in comments from the Frenchwriter, Pierre Levy, author of Collective Intelli-gence (Plenum Press, 1997). He claims that tech-nology, like hypertext, increases the knowledgespace and opens up increased potential for collec-tive intelligence. In other words, it opens up arti-f icial boundaries, and makes intelligence apatchwork that is created from a wide variety ofsources—a shift from the classical “I think” to apluralistic “we think.” Marshall McLuhan’s glo-bal village may be closer than we all thought.

RELATED LICENSING CONCERNS

Let’s return to open source code concerns. Some-one working with open source code can presum-ably be assured of several rights:

• To make and distribute copies• To access source code• To make improvements to the code

As seen by the open source community, opensource is far more than simply providing accessto source code. Similarly, it is not a specific li-

censing agreement. There are many different li-censing agreements. Moreover, there are a num-ber of conditions that must all be satisfied in thelicense. In a loose sense, here are the conditions:

• The license may not provide restrictionsfrom selling or giving away the software,even if it is bundled with sourceware com-ing from different sources. The programmust contain both the source and compiledcode—thus making it possible for othersto maintain or modify the program.

• Modifications and derived works must bedistributed under the same terms.

• Source code can have limited distributionif there is provision for the distribution ofpatch files that modify it.

• There can be no discrimination againstanyone or any group.

• There can be no discrimination against anyfields.

• No additional licenses may be required.• Licenses cannot be linked to a specific

product—for example, a specific versionof Linux.

• License cannot place restrictions on othersoftware that is distributed with the li-censed software.

It is possible for an author to release multipleversions of works—one closed and one within thegeneral definition of open source.

Note that the authors of these programs arenot surrendering their rights and placing every-thing in public domain. Instead, they are keepingcontrol of their products, and saying that they maybe made available if certain conditions are satis-fied. There are variations on the licenses used.Here are a few of the major ones:

• The GNU General Public License—GPL .Thetext of the GNU general public license isnot open source and it cannot be modified.Modifications cannot be taken private, norcan GPL programs be incorporated into pro-prietary programs. Enhancements, distri-butions, and code incorporating GPL codeall fall within GPL. Thus, it is viral.

Page 5: Open Software: Can You Afford It? Can You Avoid It?

Currents: Technological Developments

JOURNAL OF ORGANIZATIONAL EXCELLENCE / Autumn 2001

83

• The BSD License. You can do anything youwant to the source, including taking themodifications private. You must give creditto the originators.

• The Mozilla Public License. Mozilla is theopen source version of Netscape Naviga-tor. Any changes must have the sameMozilla public license. Distribution can belimited to links to proprietary code.

• The Netscape Public License. A variant onMozilla public license, in which Netscaperetains the right to take modifications pri-vate, improve them, and then re-licensethem.

How do firms operating in this environmentmake money? The most frequent routes are chargesfor software consulting, training, and 24/7 techni-cal support. You cannot take the license terms andbenefits for granted. You must read them and de-termine the impact that they might have on anyproposed activities.

… AND THE BOTTOM LINE

There are some indications that at least some soft-ware provided under open source arrangementsmay be more stable and reliable than their morereadily available commercial counterparts. It mayalso have valuable features that are lacking inother versions. Still, you should ask the follow-ing types of questions when thinking of movingin this direction.

• Can your organization deal with the addedcomplexity of building platforms based on awidely varying number and type of sources?Do you know how to select and use “best ofclass?”

• To what extent are you willing to depend onwhat may be an informal network of externalcollaborators?

• Do your employees play well with others—that is, can they work well in a collaborativeenvironment?

• Will the long-term total cost of ownership beless with a standard commercial package thatmay lack some desired features, or with sup-posedly more reliable open software thatplaces a greater emphasis on fitting modulestogether and getting technical support fromfree will collaborators?

• Are you running the risk that the system de-signed and implemented may become so com-plex that it is hard to understand and maintain?

• Will the system be truly modifiable and scal-able as your needs change?

The answers to questions like these in any organi-zation cannot come exclusively from the IT depart-ment. They reach to the heart of the functioning of abusiness, and must be taken up by managers of lineand other staff functions to make sure that thecompany’s answers align with its core business model.Alternatively, the business model needs to be adaptedto the opportunities newly available from the func-tioning of the technology. �