Using Samba / Learning Debian GNU/Linux / The Cathedral and the Bazaar: Musings on Linux and Open Source by an Accidental Revolutionary / MySQL & mSQL / Under the Radar: ...

Abstract

"Using Samba" by Robert Eckstein, David Collier-Brown and Peter Kelly, "Learning Debian GNU/Linux" by Bill McCarthy, "The Cathedral and the Bazaar: Musings on Linux and Open Source by an Accidental Revolutionary" by Eric S. Raymond, "MySQL & mSQL" by Randy Jay Yarger, George Reese and Tim King, and "Under the Radar: How Red Hat Changed the Software Business--and Took Microsoft by Surprise" by Robert Young and Wendy Goldman Rohm are reviewed.

Titles Reviewed

Eckstein, Robert, David Collier-Brown, and Peter Kelly. 1999. Using Samba. Sebastopol, CA: O'Reilly & Associates. ISBN: 1-56592-449-5.

McCarthy, Bill. 1999. Learning Debian GNU/Linux. Sebastopol, CA: O'Reilly & Associates. ISBN: 1-56592-705-2.

Raymond, Eric S. 1999. The Cathedral and the Bazaar: Musings on Linux and Open Source by an Accidental Revolutionary. Sebastopol, CA: O'Reilly & Associates. ISBN: 1-56592-724-9.

Yarger, Randy Jay, George Reese, and Tim King. 1999. MySQL & mSQL. Sebastopol, CA: O'Reilly & Associates. ISBN: 1-56592-434-7.

Young, Robert, and Wendy Goldman Rohm. 1999. Under the Radar: How Red Hat Changed the Software Businessand Took Microsoft by Surprise. Scottsdale, AZ: Coriolis Press. ISBN: 1-57610-506-7.

In the previous installment of this column, a series of books and manuals concerned with the Linux operating system were reviewed. This time, another set of new Linux books will be reviewed. Three of them are manuals providing guidance on various aspects of the operating system and its application. The other books reviewed below are commentaries on the development of open source computing.

First, the manuals, each of which is published by O'Reilly & Associates, the California-based publisher of computer books, probably best known for the visually arresting portrayals of animals that grace the covers of their books.

GNU/Debian is the Linux distribution of the moment' (or was at the time this review was written). Long regarded as the Linux distribution for purists because of its ties to the Free Software Foundation, Debian has recently been in the computer news because it is serving as the basis for the new Corel Linux distribution and for a version of Debian GNU/ Linux that is being marketed jointly by Silicon Graphics and VaLinux.2

O'Reilly's involvement illustrates a key and perhaps obvious point about the software industry in general and the commercial side of Linux in particular: support services of all kinds, including the publication of commercial documentation, can be highly profitable. What is most interesting about O'Reilly's approach to the Linux market is its decision to make several titles available as online books. Learning Debian GNUlLinux, by Bill McCarthy, is one of those titles.3

Learning Debian GNU/Linux is clearly written, carefully edited, and adequately illustrated, but the usefulness of the book will depend significantly on which version of Debian GNU/Linux is at issue, and what sorts of problems users encounter. Corel's goal is to sell its version of Linux, enhanced by the KDE desktop environment and its own office applications, as an alternative to Windows NT. Because many new Linux users complain about problems encountered in installing the operating system, Corel has taken great pains to create an installation procedure that is highly automated and requires little user knowledge of the hardware installed under the hood of the computer. Review of the first version of Corel Linux suggests that this effort has been successful, so the first several chapters of Learning Debian GNU/Linux, which deal with preparation and installation, may not be essential for Corel users.

However, for users working with the versions from the Debian Project or the O'Reilly-Silicon Graphics-VaLinux alliance, those first few chapters are among the best features of Learning Debian GNU/Linux. Operating on the reasonable assumption that a new Linux user will be installing the operating system in place of a Microsoft operating system, McCarthy discusses how to use various features of the Windows Control Panel to acquire information about the computer's hardware that will make the Linux installation considerably easier. The advice is basic but potentially valuable, because when Linux installations encounter a problem, it typically revolves around basic hardware identification and compatibility issues; for example, "Is the network card in my computer NE2000-compatible? Who is the manufacturer of my video graphics adapter and how much video RAM is installed on the video card?"

The chapter discussing the administration of a Linux system is useful, whether the user is working with Corel, Debian GNU/Linux, or some variant, but other chapters are considerably less valuable because they are generally superficial. The section of the chapter on networking that deals with Samba, the utility that permits a Windows-based workstation to be configured as a client of a Linux file or print server, provides a clear, straightforward account of how to configure both client and server, but McCarthy offers nothing about what to do when problems arise.

The chapter that deals with configuring the X-Window system is one of the book's weakest parts. Setting up X-Window can be difficult, particularly if the user is installing Linux on a relatively new computer. Linux software developers tend to lag behind in developing drivers for newer video cards, and determining how to configure support of a video card not directly supported by the operating system can be tricky business. Unfortunately, Learning Debian GNU/Linux offers little advice that would be useful in a troubleshooting situation.

That is the real issue with Learning Debian GNU/Linux. It is a sound introduction to the installation and configuration of this version of the Linux operating system. But it is impossible to endorse commercial documentation that does not deal with common problems of installation and configuration.

More to the point, the Linux Documentation Project, available online and free of charge, offers far more value. In summary, the author and the publisher of Learning Debian GNU/Linux could have and should have done much better than this.

Another of the titles that O'Reilly is making available both in print and online is Using Samba. As noted above, Samba is one of the key technologies in the open source environment. The creation of Andrew Tridgell, an Australian computer scientist, Samba's purpose is to provide a means by which resources, such as files and printers, may be shared in a heterogeneous local area network. Simply put, Samba is the software technology that enables computers running Microsoft Windows and Linux (or any of the other UNIX variants, for that matter) to interoperate.4

Clearly, this kind of interoperability is a necessary condition for significant market penetration by UNIX or any of the open source operating systems. Increasing numbers of organizations are considering an alternative to Microsoft operating systems, but few of these organizations have shown a willingness to abandon Microsoft products altogether. Even in places where the complaints about Windows have reached a crescendo, neither users nor information technology managers are willing to scrap Microsoft at the desktop. (For example, a college library might elect to use FreeBSD UNIX and Apache as the basis for its web server while retaining at least some of its existing, Microsoft-based infrastructure to accommodate users and preserve the value of previous investments. Better yet, there is the example of Microsoft's own HotMail service, which runs not on Windows NT but on FreeBSD UNIX.) Consequently, there is widespread interest in the development and implementation of Samba; in fact, articles and postings have appeared recently on the Internet suggesting that Microsoft may adopt Samba to strengthen its hand in heterogeneous network environments.

Using Samba is a highly useful guide that deals in detail with every important aspect of Samba, from the perspectives of both Linux/UNIX and Windows. It is particularly useful in dealing with issues such as authentication and security; reconciling the differing properties of UNIX and Unix-like file systems with those of MS-DOS, Windows 95/98, and Windows NT; and the management of printing. In addition, Using Samba offers valuable guidance about tuning Samba servers to increase performance and using the Secure Socket Layer protocol to harden security. (O'Reilly has also provided a CD-ROM that mirrors the primary Samba Web site, including the binaries and source code for Samba 2.05a and supplemental tools.)

Unlike Learning Debian GNU/Linux, Using Samba is a model for commercially produced documentation. It will be regarded, properly, as an important addition to the systems administrator's working library.

The third manual under review is MySQL & mSQL. The aim of this book is to take readers through the whole process of writing a program to interact with a database, in the process dealing with issues ranging from installation and configuration to programming interfaces and basic administration. It succeeds, providing a thorough review of these two open source database systems, as well as basic guidance for programmers intending to use C/C++, Perl, Python, and/or Java to extend the capabilities of these packages.

MySQL & mSQL begins with a clear and thorough introduction to relational databases, covering key topics such as tables, fields, indexes, and normalization, and an equally pertinent review of how relational databases in general, MySQL and mSQL in particular, implement the structured query language. The book also includes information about where to obtain the latest versions of MySQL and mSQL, and how to install them.

One of the most useful parts of MySQL & mSQL is its introduction to the use of the Common Gateway Interface (CGI and Perl. This section offers extensive information on how to use mSQL Perl (and the emerging Database Independent, or DBI, standard) to develop CGI database scripts in Perl, including well-described examples. The book also describes how other programming languages may be used to support CGI interaction, with reference material to all of the relevant APIs included or noted.

The significance of MySQL and mSQL in the general computing environment is based on a desirable combination of impressive power and comparatively very low cost. In the more specific context of library automation, because relational databases are the core technology of the current generation of integrated online library systems, the availability of cost-effective alternatives without significant restrictions on their adaptability at the source code level presents interesting and potentially rewarding opportunities. That is certainly the view of the automation specialists represented in OSS4LIB (Open Source Systems for Libraries) and other confederations of similar interest and intent.5

Now, to review the commentaries.

The open source community is divided into two parts. In the first part, the unifying principle is the belief that there is civic, economic, or moral value in the free development and exchange of software. In the second part, the union is defined mainly by what is widely described as a new business model, in which open source development reduces the sunken costs of research and development, accelerates the rate at which systems and applications are developed and debugged, and creates open distribution channels that engender and then feed broader interest in various software products. Two recently published books, Eric S. Raymond's The Cathedral and the Bazaar and Robert Young and Wendy Goldman Rohm's Under the Radar, represent those two factions of the open source movement and provide readers with useful insights into the philosophy that unites them and the interpretations that divide them.

The Cathedral and the Bazaar is an important, if deeply flawed, book. Perhaps more to the point, The Cathedral and the Bazaar is likely to be viewed as an important book. One of the principal reasons is timeliness. The software industry is big news. The recent growth of Linux in the marketplace has been remarkable. Microsoft's problems in and out of court have garnered even more attention. If Y2K has taught us anything, it is that software and how it is made are issues of critical importance.

Moreover, the essay from which The Cathedral and the Bazaar's title is taken is arguably one of the most important Internet documents of the last few years. It has given specific voice and vision to a movement that many believe is revolutionary. It has been identified as a major influence in attracting a number of major software firms, including IBM, Oracle, and Corel, to the open source model and as a principal basis for Netscape's decision to make its browser source code open. The publication of Raymond's essays in book form is, in effect, an assertion that his ideas, and the larger milieu in which those ideas developed, are important.

Viewed from another perspective, The Cathedral and the Bazaar is a collection of essays that form a personal memoir of computing over the past two decades and a meditation on the organization of software development.6 The collection is engaging as a memoir, provocative in its contemplation of work, and highly problematic as a blueprint for software development.

Raymond describes how the open and decentralized approach to the development of the software tools and utilities that surround Linux works, using his own fetchmail project as a principal source of evidence. He likens the centralized models favored by proprietary developers to the cathedral building and the process out of which Linux has emerged as an environment like that of a bazaar. The essence of the book is Raymond's belief that at the heart of the remarkable success of Linux, there is the putatively anarchistic bazaar and a lesson of transcendent importance about the nature and organization of work. The problem is that the polemic he advances by way of explanation is deeply flawed.

It is essentially a moral argument that Raymond makes, albeit one that he has obscured through an idiosyncratic blend of Fabian economics, technology, and Utopianism. The rhetoric varies from essay to essay, but Raymond's basic stance throughout the book is that what he calls "hacker culture" is by nature altruistic, and that altruism makes better technology. According to this argument, Linux has succeeded because the hackers who form the core of the open source movement are more enlightened and less self ish than the programmers who direct and work under proprietary development models.

A much more plausible explanation is that the success of Linux is the result of the coalescence of a series of factors in the early to mid 1990s, including the availability of open source programming tools (created mainly by Richard Stallman and the Free Software Foundation), access to the electronic mail and file transfer capabilities via the Internet, and the presence of a cadre of programmers talented enough to contribute in significant ways to the development of what is, in reality, a derivative operating system. The idea of this "critical mass" is implicit in several of Raymond's essays, but he clearly prefers the more dramatic interpretation of events.

There is also the problem of Microsoft. Microsoft is the constant counterpoint in all of the arguments for open source software. Raymond's central argument for the superiority of open software development is built in significant measure upon the notion that the rapid evolution of Linux (and Microsoft's problems in delivering various products on time and in stable form) is an indication that centralized and proprietary software development models are fundamentally defective.

It is an argument that is easy to embrace. Microsoft's problems are well documented. Its products are bloated and buggy, with Windows 2000 promising to set new standards for both bloat and bugs. Also, Microsoft recently was found guilty of unfair business practices in a federal investigation. Hence, Microsoft is the perfect straw man for polemics such as those offered by Raymond.

But the arguments that play off of Microsoft are wrong because they ignore a variety of attendant realities. Proprietary models do work; in fact, there is at least forty years of evidence to indicate that the models Microsoft uses-what Raymond refers to as the software engineering equivalent of cathedral building-are successful and sustainable. Centralized development models work, too. The FreeBSD Project, which is less publicized than Linux but no less important in the context of the open source movement, employs a centralized development, suggesting that centralization and the open source movement are not necessarily, and perhaps not at all, incompatible. (It is worth noting that many of Microsoft's software development practices bear a striking similarity to practices that Raymond lauds when he finds them in the open source setting [Bezroukov 1999]. And when Linux development is divested of its sociopolitical embellishments, what is left is a process that is loosely organized but not altogether decentralized, as witnessed by the role played by Linus Torvalds.) Again, there is the moral dimension: Raymond's arguments against proprietary development suggest that the desire to control products for the purposes of profit taking is immoral, and therefore that the resulting technology is bad.

These lines of argument lead to a tangle of ideas, many of them interesting and thought provoking, some of them overwrought and wrongheaded. But all of its shortcomings notwithstanding, The Cathedral and the Bazaar is well worth reading. The essay titled "A Brief History of Hackerdom" is a concise, straightforward look at how and where the hacker culture grew.' Raymond establishes the lineage of open source computing-how we got from Multics to Linux in fewer than thirty years-and provides insight into the logic, political as well as technological, that shaped important initiatives and trends. "Revenge of the Hackers" is an update, focusing largely on the burgeoning interest in Linux and various reactions to Raymond's own work. Of particular interest is Raymond's account of his involvement with Netscape and the 1998 decision to make the source code for Netscape Communicator open. The other essays in the book, "Homesteading the Noosphere" and "The Magic Cauldron," deal with a wide array of issues engendered by the title essay, ranging from notions of ownership and use rights under the open source scheme to an assessment of what Raymond calls the "business ecology" of software development.

Under the Radar was written by Robert Young, the chief executive officer of Red Hat Software, and Wendy Goldman Rohm, who is identified as a technology journalist. It is a less provocative, less ambitious book; in fact, Under the Radar is written so poorly as to suggest that it may have been rushed into print mainly to support the first public offering of Red Hat stock. But, in some ways, Under the Radar may be more soundly instructive than The Cathedral and the Bazaar.

Red Hat Software is far and away the most successful of the for-profit organizations that have been spawned by the development of Linux. Under the Radar traces the development of open source computing from the business perspective and the evolution of the business model that Young has played an important part in developing.

The book is interesting on two levels. Young's anecdotes about people like Richard Stallman and Linus Torvalds are interesting, and often telling. (My favorite is Young's account of how the Free Software Foundation came into being. According to Young, Stallman encountered problems driving a printer. He asked the manufacturer for a copy of the driver's source code so that he could fix the bug in it, but the manufacturer refused to give him a copy. The resulting frustration must have been monumental, because Stallman established the Free Software Foundation not long thereafter.) His accounts of business dealings with executives from established companies such as IBM, Intel, and Dell and Red Hat's interest in ventures such as Gnome Project and the Zenguin installer offer useful insights into the current and probable trajectories for commercial versions of Linux.

Unfortunately, Under the Radar compels the reader to endure some of the obligatory cant of high-technology industry. For example, Young and Rohm write:

If you move into an industry playing by existing rules, you'll just become another player. You won't revolutionize that industry unless you change the paradigm.

Does the open source movement really constitute a "whole new way of thinking" for the software industry? It is doubtful. New paradigm or not, the reality is no less arresting. Entrepreneurial firms such as Red Hat and established powers such as IBM and Oracle have discovered that open sources offer considerable leverage, inasmuch as they provide the means to establish a for-profit venture on the basis of assets that are available without significant restriction or cost, and that are under more or less constant development. As competition grows keener and the investment climate more demanding, using open sources will make more and more sense.

Footnote

Notes

Footnote

1. Unlike the other versions of Linux, Debian is actually kernel independent. It currently uses the Linux kernel, but work is in progress to provide Debian for other kernels using the GNU Hurd. The GNU Hurd is the GNU project's replacement for the UNIX kernel. The Hurd is a collection of servers that run on the Mach microkernel to implement file systems, network protocols, file access control, and other features that are implemented by the UNIX kernel or similar kernels, such as Linux. Currently, the Hurd

Footnote

runs on i386 machines, but the Free Software Foundation plans to port the Hurd to other hardware architectures or other microkernels in the future.

2. Profits from Debian GNU/Linux sales will be donated to Software in the Public Interest, a nonprofit organization for open source projects including Debian.

Footnote

3. Learning Debain GNU/Linux is available as a seperate printed title, as an outline book at http://www.oreilly.com/catalog/debian/.chapter/ index.html, and as part of the software package being distributed by O'Reilly, Silicon Graphics, and VaLinux.

4. The basis for Samba is the Server Message Block (SMB) protocol. Developed initially by IBM and later incorporated by Microsoft, SMB provides a methos for client applications in a computer to read and write to files on and to request services from server programs in a computer network. Using the SMB protocol, an application can assess files at a remote server and at other resources including printers, mailboxes, and named pipes. With the SMB protocol in place, a client application may read, create, and update files on the remote server. It can also communicate with any server program that is set up to receive an SMB client request. Microsoft has offered a public that is set up to receive an SMB client request. Microsoft has offered a public or open version of SMB for the Internet to the Internet Engineering Task Force. Called the Common Internet File System (CIFS), this new protocol provides more flexibility than existing Internet applications such as the File Transfer Protocol (FTP). CIFS is envisioned as a complement to the Internet's HTTP protocol for web browsing.

5. The URL for OSS4LIB is http://info.med.yale.edu/library/oss4lib/.

Footnote

6. Readers who have encountered Raymond's essays on the Internet should be advised that several of the chapters in the book are little different from the versions that are available on his personal web site. The URL is http://www.tuxedo.org/-esc

7. By Raymond's reckoning, a hacker is "an enthusiast, an artist, a tinker, a problem solver, an expert." A hacker is not a vandal or a criminal.

References

Reference

References

Bezroukov, Nikolai. 1999. A second look at The cathedral and the bazaar. First Monday 4 (12).