O'Reilly & Open Source

Ask Someone Who Knows^TM

 

By Dennis Kenney

(Tim) O'Reilly and Associates, Inc.

 

Tim O'Reilly is the guest speaker at a presentation, sponsored by MIT's ACM/IEEE chapter, on the third floor of the Julius Adams Stratton Building, MIT's Student Union. It's hard to separate Tim O'Reilly from the company that he is the founder and president of. Excuse the familiarity, Tim, but I'll refer to Tim when I mean the man, and O'Reilly when I refer to the corporation, unless the meaning is obvious from the context.

 

O'Reilly started out as a Unix documentation and technical writing consulting firm, which wrote Unix manuals to fill in its down time. The realization that there was an urgent need for the pamphlets on Unix utilities that O'Reilly was licensing to Unix vendors, nudged O'Reilly into the technical publishing business.

 

Demand for online versions of O'Reilly Unix books resulted in O'Reilly's participation in the Davenport Group's project on online publishing. The main product of this collaboration was the Standard Generalized Markup Language (SGML) DocBook Document Type Definition (DTD). See O'Reilly's "DocBook: the Definitive Guide" by Norm Walsh and Lenny Muellner. The DocBook XML DTD will be of greater interest to an XML developer. The subject matter of Tim's presentation is as wide as the subject matter of O'Reilly books, emphasizing naturally, O'Reilly's latest products and projects. The Web runs on open source software and open standards, many of which are documented in O'Reilly products as well as in other sources. There will be a quiz at the end of this article.

 

Tim's essay "Hardware, Software, and Infoware" in "Open Sources: Voices from the Open Source Revolution", edited by Chris DiBona, Sam Ockman & Mark Stone, clarifies Tim's proposition that Amazon.com is the new killer app on the Net, something that people will buy a new computer or device to use. Tim argues that such an Internet information application is a new breed of application, infoware. An infoware application is usually on a site on the Net with lots of information and a small amount of software. This is the realm of Web Services and the other new genre of software, peer-to-peer (P2P) networking, exemplified by Napster.

 

Capturing the Knowledge of Innovators

 

O'Reilly books specialize in capturing an innovator's knowledge: using an O'Reilly book is like having the author sitting next to you by your computer. Tim wants to grow O'Reilly by growing the industry. O'Reilly will grow as the whole pie grows, not at the expense of competitors, a win/win scenario for everybody involved.

 

RMS, mundane name Richard Stallman, laments that the lack of good free manuals is the biggest deficiency in free software. O'Reilly has filled a lot of this need with his not-for-free books. Tim tells the story of when RMS called him a parasite because he didn't give away his books. O'Reilly has done just that unwittingly when Web sites have put O'Reilly book CDs up on the Net, most notably in Romania. The story of one of Tim's encounters with RMS shifts to editors with RMS choosing to develop Emacs rather than use vi or ed. RMS's retort to Tim's stating that O'Reilly sells 3 times as many vi books as Emacs books, is that Emacs's online documentation is so good that no manuals are needed. Und so weiter.

 

Bill Joy's vi editor is part of the Berkeley Software Distribution (BSD) Unix, which was well on its way to World Domination when UC Berkeley was sued by AT&T for copyright infringement. The battle was settled by Novell CEO, Ray Norda, after Novell bought the Unix Systems Laboratory (USL) from AT&T. I remember the events clearly because some Novell Support Engineers (SEs) had a problem adjusting to selling UnixWare, Novell's version of the system which had been a strong competitor of Novell's NetWare.

 

O'Reilly books are by technical people for technical people. No "For Executives" summaries here.

"Animal" Books

Lorises and Lorrie

 

Edie Freedman was hired to design the covers for O'Reilly's first books on various Unix utilities. She describes her delight upon discovering some old 18^th and 19^th century Dover Pictorial Archives engravings of a pair of lorises. Edie knew immediately "That's sed and awk!" Beautiful images of animals have become identified with O'Reilly ever since.

 

Only a few authors have been able to pick their own animals--Larry Wall's camel and Andrew Odewahn's boll weevil come to mind. Many authors and their family members have actually questioned the undisclosed animal selection process.

 

Lorrie LeJeune is one of O'Reilly's best animal artist. Lorrie was hired for a position in corporate sales but ended up becoming O'Reilly's first product manager without ever working the corporate sales job. Lorrie discovered that it was more fun making books than selling them and has become an editor. One of Lorrie's "customers" was the Boston Network Users Group, which was trying to get some O'Reilly books for door prizes at a seminar the group was doing on Web Programming. Lorrie sent BNUG twenty-one books. Read about Lorrie's animal drawing techniques at oreilly.com.

 

"But they're ugly." "They're scary." Scary. You want scary. Take a look at the photographs of the authors on the covers of the WROX Publishing, Ltd. books, if you want scary.

 

Tim's presentation at MIT gave me a chance to give Tim BNUG's thanks for the books O'Reilly gave us for the Web Programming seminar.

"So You Want to Write a Book?"

 

oreilly.com has a roadmap for the aspiring O'Reilly author. O'Reilly's default style sheet extends "The Chicago Manual of Style" for O'Reilly's technical writers and includes a technical word list. UNIX is out, Unix is in. Sorry Tim, but I like eCommerce better than e-commerce and is it Unixes or Unices? "A Guide for New Authors" describes the "traditional" subject matter of O'Reilly books and the hot new subjects such as bioinformatics and wireless networking that O'Reilly is interested in. Most of you may know this already. The specific information on proposals, contracts, editing and the production cycle is must reading for any technical (or non-technical) writer. The frosting on the cake is the description of the O'Reilly view of marketing and revision processes for any successful book.

 

This is all very true, but can someone enter the Shaolin temple without being invited, say through the kitchen? The author of Java and XSLT (XML Stylesheet Language Transformation), Eric M. Burke, clicked on the "So You Want to write a Book?" link and ended up sending a proposal that was rejected. What O'Reilly really wanted was a book on integrating Java and XSLT. Burke's "Java and XSLT" (ermines) is one of the three O'Reilly books I keep chipping away at. The other two are the 4^th edition of "HTML & XHTML, The Definitive Guide" by Chuck Musciano & Bill Kennedy (koala bear) (XHTML was added to the title for the fourth edition), and "Open Sources" (no image). O'Reilly really likes ampersands.

 

Boston Linux & Unix User Group (blu.org)

Real Linux users don't whine.

 

Jabba the Hutt, Jaber, Jabberwock, jabr, and Jabber

 

Saeed Salem Jaber of Qator lifted a total of 1012 pounds in the three Olympic lifts to win three gold medals at the world weightlifting championships last year. This Herculean feat gives Jaber boasting rights to being the world's strongest man.

 

jabr@blu.org, mundane name John Abreau (Portugese not French DNA), the executive director of the Boston Linux Users' Group, tells me that Jabber is the free software peer-to-peer (P2P) Internet instant messaging (IM) framework, based on XML routing. See O'Reilly's "Programming Jabber: Extending XML Messaging" (bare-throated ballbird) by D. J. Adams.

 

Douglas Aircraft Company (DAC), which was assimilated long ago by the Big B, Boeing, had standardized its PC workstations so as to provide some permanence to the configuration. After all, I had one friend who had designed aircraft nose wheels for all of his 40 years at DAC. The configuration consisted of an IBM AT, EGA monitor, DCA IRMA card (remember the IBM 370?), Epson 100 dot matrix printer, a Hayes 1200B modem, and a 20-Meg hard disk. DOS, WordPerfect, Lotus 1-2-3, dBASE III Plus, and Sidekick formed the software office suite. What more could an engineer want? After all, BNUG had kept its 1500-member mailing list and the dBASE III Plus program itself on one microdiskette. jabr tells me that Linux has this software capability and more for free. A few other Boston Linux Users Group members say that the Linux GUI shells, GNU Network Object Model Environment (GNOME) and the K Desktop Environment (KDE), are even better than Windows.

 

My supervisor/boss had one administrative assistant (we called the girls secretaries in the glory days of Apollo) so we did our own letters. I will admit that I snuck over to the Chinese liaison office to use their HP laser printer and would have liked more than 80 columns for editing proportional fonts. HP BASIC, Fortran (simulations), and HP assembler were our software tools. Our flight test data was recorded with an HP 9845 mini that had two tape cassettes drives. In the days before AutoCAD it would take me two days to redraw a red-lined avionics wiring diagram that consisted of just rectangles, lines and text. I was a member of the upstart PC/BASIC generation. I had already missed the ARPANET and Unix waves of hacker history.

 

Fast forward a little to the Linux Special Interest Group (SIG) of the Boston Computer Society (BCS). Simson Garfinkel ("Web Security, Privacy & Commerce" with Gene Spafford and "Database Nation") had cancelled his presentation at the group a week before. The now Boston Linux (and Unix) User Group (LUG) was dependent on the Boston Computer Society magazine for scheduling information at the time and couldn't reach all of its membership to let them know of the change in plans. Dan Demus volunteered to do a short presentation on using Linux as a boot server. Dan's short presentation still left jabr scrambling to fill the remaining time of the meeting.

 

I was networking and noticed RMS and a female companion going to (my memory is that it was) a folk dancing class. I remember vaguely mentioning to the group I was with that WordPerfect had been ported to SCO (Santa Cruz Operations) Unix and that made it capable of being used with Linux. Wrong comment in the Free Software dungeons of MIT! RMS abandoned all of his previous plans (and his companion) and lectured us for close to an hour about his not understanding why anyone (especially someone in a Linux SIG) would PURCHASE a commercial word processing package when she could just as easily write a few lines of code. I was still trying to master DOS regular expressions.

 

Other memorable LUG meetings were Eric Raymond's and Linus's presentations at the group. Linus joined the group after the meeting for the usual trip to the Cambridge Brewery (the site of BNUG's annual, until this year!, Pub Crawl). I missed that Bierwanderung.

 

The Last True Hacker

 

RMS developed the General Public License (GPL) to prevent commercial interests from homesteading public domain knowledge, in particular, the software he was developing for future hackers. One of the places that RMS has documented some of his formative thoughts and activities is his "The GNU Operating System and the Free Software Movement" which appears, among other places in O'Reilly's "Open Sources", edited by Chris Bona, Sam Ockman & Mark Stone. The gutting of the Artificial Intelligence Lab at MIT by the departure of most of its hacker community to build the LISP-based Symbolics machine was traumatic to RMS. He decided to resolve the moral dilemma he faced by designing a "free" Unix-like operating system that he called GNU. GNU is a recursive acronym for "GNU's not Unix".

 

RMS quit his job at MIT in January 1984 to free himself from any claim that the university might have on the GNU software and started writing code.

 

The copyleft idea of the GNU General Public License (GPL) was designed to continue the "free" state of the software, preventing it from becoming proprietary as MIT's X Window System had to end users. RMS intended for any code used with GPL software to become "free", something that detractors have called the viral effect of the GPL. "The central idea of copyleft is that we give everyone permission to run the program, copy the program, modify the program, and distribute modified versions--but not permission to add restrictions of their own." RMS was creating a new software sharing community.

 

GNU Emacs (Editing MACroS)

 

RMS developed Emacs for editing before he turned it into an environment. He started a free software distribution business to support himself and share the program with those people requesting copies. RMS later developed the GNU GCC compiler which Linus considered the most important GNU program used by Linux. Quantum Books in Cambridge has several Free Software Foundation (FSF) "free" books for sale including O'Reilly's "Learning GNU Emacs" by Debra Cameron, Bill Rosenblatt & Eric Raymond. O'Reilly has no rights to the bison on the cover.

 

While drinking at the Muddy Charles at MIT with 6 doctoral candidates at the Sloan School of Management, the puzzle of what the M in RMS stood for came up again. The spokesman for the group said it meant "'rite a macro, stupid". A few other equally authoritative opinions were offered. I went to the Boston Linux Group's January 16th meeting to see if Bradley M. Kuhn of the Free Software Foundation could put together "Software Freedom, the GNU Generation, and the GPL" for me.

 

RMS and Linus received the Electronic Frontier Foundation's Pioneer award in 1998. The pair shared the Takeda 100 million yen (over $800,000) award with University of Tokyo professor Ken Sakamura this year for open development models for system software at the awards ceremony last December. Professor Sakamura started the TRON project (The Real-time Operating system Nucleus).

 

Dr. Ken Sakamura describes the TRON project in "TRON Project 1990", the edition of the TRON Project that I looked at. He hopes the TRON project will help enable a fully computerized society in the future. The embedded applications described are the TRON House, the TRON Intelligent Building, TRON Intelligent City Development, the TRON Intelligent Auto, BTRON Multimedia Communications Research, and the TRON Electronics HMI (human/machine interface).

 

Deja vi, all over again

 

I love to say "vee" instead of "vee eye" around my Linux bigot friends. They cringe beautifully even knowing full well that I'm putting them on. The vi tarsier, the "vi guy" is probably O'Reilly's best known brand. It appears on the "Learning the vi Editor" cover and peeks out here and there on the O'Reilly site. A tersier was on the back of a t-shirt that a local (Boston) O'Reilly editor gave me at the Visual Studio.NET roll-out. A hacker's editor is as personal as a saddle is to an equestrian and the source of much hacker folklore.

 

If it smells like a camel...

 

Larry Wall is a linguist by training. That's a lot like saying some of my friends are ex-marines. Once a marine, ... Larry's self-created language of choice is Perl, the Practical Extraction and Report Language or the Pathologically Eclectic Rubbish Lister. Take your pick. Larry and Perl occupy a unique position in the hacker tribal history. Perl originally was the scripting language of choice for system administration and has found a second home creating dynamic web pages. Larry has speculated that he was the irritant that caused Perl to form. He continues to cause scratching in response to his being an itch. I call it Perl ivy. The Camel Book, "Programming Perl" (camel) written by Larry Wall, Tom Christiansen & Randall I. Schwartz, is the "standard" Perl book. An easier place to start learning Perl is O'Reilly's "Learning Perl" by Camel Book co-author Randall Schwartz. Now an elder in the hacker tribe, Larry is perceived to be the hacker diplomat.

 

"Diligence, Patience, and Humility" by Larry Wall

 

"A lot of my thinking this year has been influenced by working with Unicode and with XML." Larry is still zinging from the latest update to Perl, caused by the need to make Perl Unicode-compliant. Newly discovered insights from XML and refamiliarization with the ideograms of the mysterious east have him developing a new yin and yang.

 

This essay is a sort of a state of the union for Perl and the open software movement. Oversimplifying Larry's thoughts, Larry describes the free software and commercial communities as binary planets (Pluto and Charon?), not close enough to have their tidal forces cause mutual destruction. Tidal distortion merges the two planetary atmospheres so that a person can fly but not walk between the planets. (See Rocheworld.) The planets don't collide because of copyright law.

 

Next comes the yin and yang or the light and dark sides of The Force to you new guys. On one side are the programmer virtues of laziness, impatience, and hubris. On the other side are the community values of diligence, patience, and humility. Larry, always the diplomat, says these are not opposites but different points of view, virtues that if followed, will carry the software community forward.

 

Many volumes of Perl later, Perl stars in O'Reilly's "Beginning Perl for Bioinformatics" by James Tisdall. Of interest to BNUG members is "Perl for Web Site Management" by John Callender (Probocus monkey).

 

Professor David Miller, the director of the MIT Space Systems Laboratory in the Daniel Guggenheim Aeronautical Laboratory has just finished a lecture on the Spheres project. His undergraduate students have designed and developed spheres that can control their attitude and relative positions in zero or micro gravity. Testing is difficult in one g and on "frictionless" flat surfaces. The moment of truth comes when the students test the spheres in "simulated" zero gravity in a modified KC-135, the infamous "Vomit Comet", a triumphant finale to the 3 semester course. A new group will start the 3 semester course in February, if the powers that be approve.

 

The Cous Cous and the Yang

 

Heading south towards the Charles River on Massachusetts Avenue, I end up by the crosswalk to the student union, at the main entrance of MIT, just outside of the lobby in Building 7. In a moment of indecision I'm stuck between the cous cous and the yang. You know that cockroaches have lived on the earth for over 400 million years and will probably be here 400 million years after man is extinct. Chef Boutros Youssef owns the Cous Cous Kitchen catering truck towards the south and Mr. Yang owns the Chinese "Delicious Lunch Box" truck to the north on Mass Ave.

 

Total World Domination Now

 

"The Linux Edge" by Linus Torvalds

 

"The importance of compilers was one reason that I chose to license Linux under the GNU Public License, since the license for the GCC compiler was GPL." The availability of compilers is what makes an operating system and programs portable; Unix and C had already proven this.

 

Linus decided that system calls would not be considered covered by the GPL. This decision meant that commercial, usually binary-only, modules or programs could be run on Linux without violating Linux's GPL. There's nothing earth-shattering about this--proprietary software can be written for any operating system. Linus's decision marks a commonly accepted data point on the fuzzy DMZ between the free and commercial software communities.

 

Linus's greatest accomplishment was the Linux development model. Many have said what made Linux develop so fast was the lack of barriers to contributing, the "bazaar" approach. The fact that Linus concentrated on the kernel and had most of a Unix type system available from already existing free software has been emphasized by the hackers in Cambridge, Massachusetts who would like to call the system GNU/Linux. Also mentioned is the fact that the hackers were just cloning propriety (left coast?) Unix.

 

Dare I suggest the possibility that the American legal system drove the American-dominated free software movement offshore (first to Finland and Germany), just as it did the light aircraft industry?

Open Source Licenses

 

Donald K. Rosenberg's "Open Source" published by IDG Books Worldwide, Inc. may be needed to make sense of open source licensing. Don't confuse this book with "Open Sources" from O'Reilly.

 

A truce of sorts is required now. Let's just say that the term "Open Source" is considered by some hackers as a rewriting of hacker history, a battle that I'm not going to get into any more than I have to.

 

Consider this a header file or an overview that hopefully will have the details filled in as we go. Free software is the software protected by the GPL license, such as Linux and GNU, while open source also includes other free source licenses such as the BSD and Netscape licenses.

 

Let's look at Tim and O'Reilly's history first. Tim was/is an active participant in the technologies formerly called free software that run the Internet as well as a founding member of the Open Source Initiative (OSI), an attempt to make the free software movement more amenable to business interests and unite the various open source groups.

 

Copyrights, Patents, Royalties, and the Common Good

 

Open Source and O'Reilly are concerned with intellectual property and who owns it. Oreilly.com has many articles about the continuing IP debate and flame fests.

 

Copyright is a government-issued right for temporary monopolistic control of certain intellectual property. Even material in the public domain can be copyrighted. Corporations have invested in academic research and have copyrighted or patented research done in the public domain. The problems of such collaboration and/or adoption of academic or public domain software is well documented in the AT&T/BSD Unix legal cases. Other researchers have been hired in industry for their expertise in a technology or have started their own companies using the academic research they were involved with. The Netscape Corporation is a good example here. With the research universities becoming more interested in patents and the resulting revenue from the licensing of patents, the boundary is even more blurred. Money, intellectual property (overloaded IP), and the public good separate and merge. The real turmoil occurs when there is a real or perceived inequity in the distribution of money or recognition associated with the privatized or just use of open source software. The pot came to a boil when it was perceived that business interests had assumed control of the open-source movement, through the Open Source Initiative.

 

The main thing that both the Open Source Initiative geeks and suits agree with is that the source must be available for modification.

 

I'm writing a novel on Little League baseball so I registered a "vanity" site to sell the book and hopefully to get some free help editing. The title plays on the various meanings of the words sweet and low in sports and everyday life. As soon as I registered the site sweetandlow.com I received a letter from an IP legal firm based on the New York Avenue of the Americas saying my site was diluting the trademark Sweet`N Low. They'd call it even if I'd sign over my rights to the site. My daughter-in-law, an Intellectual Property lawyer told me the battle wasn't worth fighting. An IP lawyer in Phoenix, in a preliminary free meeting, told me he'd write a letter, laying out my position for a $1000.

 

My old friend, Freddie Gallant (pronounced gallant), once told me you get as much justice as you can afford.

 

Berkeley Software Distribution (BSD) License

 

The anchor in all this is RMS and the Free Software Foundation. Stallman's "copyleft" license allowed the use and modification of the computer source code but required that the code remained "copylefted". Other licenses allow the modifier the use and modification of the source code without the requirement that the new code be published and the "freedom" to sell the code in binary-only form. Comparing these two extremes, the BSD license allows you to make modified free source code private, while the GPL doesn't.

 

Proprietary software is software where the source code is considered a "trade secret" and not made public even if the code formerly had been free software. Examples of companies that have taken open software private are Microsoft, Apple, Sun, and SAP.

Open Source Initiative

"Software is not software without source code." -NASA

 

Most of the technologies that make the Web possible are based on free software. Face it. You're not going to see the average windows user at a USENIX conference. Not that Windows is that easy to use. I've spent hundreds of hours using and learning Windows programs. Helping a "seldom" user to check her email reassures me that Windows is not "intuitive". Nonetheless, world domination is only possible by winning the "minds and souls" of the click-and-pointers. That is, if we want them in our community and dating our sisters.

 

When Bruce Parens and Eric Raymond tried to register the term "Open Source" the U.S. Patent and Trademark Office rejected the application. They considered the term to be too general.

 

"The Open Source Definition (OSD)" by Bruce Parens

 

Bruce Parens did much of the basic work on the Open Source Definition in a document called the "Debian Social Contract" which is used with the Debian distribution of BSD Unix. The work he did on the OSD and his hopes for the Open Source movement are described in his essay in "Open Sources", written before his flame fest with Eric Raymond caused him to resign from the Open Source Initiative. The subject of the disagreements was criticism of OSI supporters--first Apple Computer and then O'Reilly. Parens objected to O'Reilly's use of the term Open Source for his Open Source Summit conference. Parens had also felt that the Apple Open Source license, which would allow Apple to unilaterally pull their license, wasn't Open Source. Apple soon modified their license in response to Parens's criticism.

 

The Rifts--Quotes from "Open Sources"

 

"Teaching new users about freedom became more difficult in 1998, when a part of the community decided to stop using the term "free software" and say "open-source software" instead.

 

Some who favored the term aimed to avoid the confusion of "free" with "gratis"--a valid goal. Others, however, aimed to set aside the spirit of principle that had motivated the free software movement and the GNU project, and to appeal instead to executives and business users, many of whom hold an ideology that places profit above freedom, above community, above principle. Thus, the rhetoric of "Open Source" focuses on the potential to make high quality, powerful software, but shuns the ideas of freedom, community, and principle."

-RMS's essay "The GNU Operating System and the Free Software Movement"

"On the technical front, we knew that the configurability of the system was key to delivering a "one size fits all" architecture. On the business front, we knew that a "one size fits all" was key to creating a unifying and beneficial standard for embedded system development. But we still could not figure out who was going to pay for this benefit. The two sides worked on their problem independently for a year and a half. R&D costs mounted. Unable to reconcile the Open Source paradox, many managers didn't make it." -Michael Tiemann's essay on the "Future of Cygnus Solutions"

 

"Open Source proponents found KDE (graphical desktop) objectionable because they perceived that their programmers were trying to blur the definition of what free software was to include partially-free items like Qt (proprietary graphical library). The KDE developers contended that their programs were open source, even though there were no runnable versions of the programs that did not require a non-Open-Source library. I, and others, asserted that KDE applications were only Open Source fragments of non-Open-Source programs, and that an Open Source version of Qt would be necessary before KDE could be referred to as Open Source."

-Bruce Parens's essay "The Open Source Definition"

 

It continues to this day with W3C Recommendations and various mixtures of licenses: the suits vs the geeks. See "Patterns, Royalties, and the Future of the Web" by Kendall Grant Clark, "Will You See Open Source J2EE Implementations?" by Mike Loukides, and "Patents and Web Standards Town Hall Meeting" by Michael Champion on xml.com.

 

"The Cathedral and the Bazaar" by Eric S. Raymond

 

Eric S. Raymond (ESR) is a modern participant/anthropologist comparable to Sir Richard Burton of yesteryear. O'Reilly has now published Raymond's "The Cathedral & the Bazaar" in a hardcover edition. "Cathedral" taught the Open Source participants how to talk and think about themselves much as Le Carr taught the "Cold War" spies about blind drops, etc. It is hard to discuss anything worth arguing about in the Open Source space without remembering something Raymond has written.

 

The success of free software in the mass consumer market required an explanation of the free software development model in terms that the average computer user could understand. Raymond did this in his initial manifesto, which was published and distributed freely over the Internet. The hardcover edition expands Raymond's essays, which he describes as being continually refactored and updated, usually in response to peer review, just like open-source code. "Cathedral" develops the idea that freedom encourages innovation in hardware and software while monopolies stifle innovation. But enough, already. Read the book or essay and arrive at your own interpretation and conclusions about free software. Don't even finish this article if you haven't at least read the title essay. That's what I did.

 

We are all familiar with the block-buster movie business model--lose money on the movie and make it up selling toys, t-shirts, and the sound track. Raymond characterizes this in O'Reilly's case as accessorizing--O'Reilly's hosting of conferencing, on-line white papers and tutorials, and the patron-like sponsorship of the high priests of software development as a means of selling O'Reilly's real product of high end technical books. Raymond's musings in "The Magic Cauldron" gives us insight into the lukewarm reception of the Linux crowd to Sun's Java and JINI marketing plans and Netscape's release of Mozilla to Open Source. And finally, embedded in the "Cauldron" essay is the conclusion that "open source peer review is the only scaleable method for achieving high reliability and quality."

 

A major advantage in trouble-shooting free software is that there is better documentation. The free software world spends a lot of time keeping the source code clean and accessible. Propriety code doesn't need to look pretty; it isn't open to public inspection. I've done a lot of documentation (including requirements) after the development phase was finished-- because the documentation was a "deliverable", required by contract.

 

Something is wrong with my mother's PC. I know that the performance of its Microsoft Windows Office 2000 could be improved by a little more RAM. The symptom is that I end up with corrupted Word 2000 files. It occurs when I save my files to microdiskette for printing on another machine. Somehow the machine doesn't make the jump of changing the target from the disk drive to the hard drive. This "problem" caused me to lose or have to retype about half of my work on a previous article. Is this problem caused by a bug, a Word virus, or a bad disk drive?

 

 

"The Revenge of the Hackers"

 

This essay was written after Raymond's and GNU's fortunes had improved. "Twenty years of living in a ghetto--a fairly comfortable ghetto full of interesting friends, but still one walled in by a vast and intangible barrier of mainstream prejudice inscribed `ONLY FLAKES LIVE HERE.' " Netscape's opening up the source code of their browser was the catalyst for a marketing campaign, involving elements of the Linux community, Eric, and Tim, later named the Open Source Initiative (OSI).

 

Pithecanthropus (Java man)

 

Two full days of classes at Software Development 2001 East, Web Services, with Elliotte Rusty Harold (Intro to XML & Java and XML courses) and a few shorter sessions with Scott Means and I thought I knew what processing XML was all about. Harold and Means co-authored "XML in a Nutshell". Harold also authored "The XML Bible", "Java Network Programming" and "Java I/O". Harold taught from the original handout text in his laptop, which I understood was written in XML using the XML DocBook DTD. He was editing his notes in real-time as he taught the class. Means is new to the conference circuit and I look forward to future books from him.

 

Java, "write once, run anywhere", is a product of Sun (Stanford University Networks), as is the Java 2 Enterprise Edition (J2EE) platform. Comparing Sun and BSD licensing with MIT/FSF licenses can become as cantankerous as the Route 128 vs Silicon debates (DEC vs Sun and other workstations). I found the free tutorials and white papers on oreilly.com to be very helpful, especially the tutorials on XML, Java, and LDAP.

 

RMS's "GNU Manifesto" as a Business Plan

 

Michael Tiemann, one of the three founders of Cygnus Solutions read RMS's GNU Manifesto at the end of the GNU Emacs Manual and saw a business plan. Cygnus Solution was recently acquired by Red Hat and is the best example of a successful Open Source company in embedded systems.

 

"I reached a different conclusion (than RMS in the GNU Manifest), one which Stallman and I have often argued, which was that the freedom to use, distribute, and modify software will prevail against any model that attempts to limit that freedom. It will prevail not for ethical reasons, but for competitive, market-driven reasons." Michael also explains how being an Open Source maintainer can give a company a competitive advantage. The essay describes a fascinating business plan even if Michael seems somewhat politically insensitive. Michael's quote in the Rifts section pertains to an original development project at Cygnus, which would abstract the hardware interface of the roughly 100+ processors in the world from a common development library.

 

John Gilmore, the GNU debugger (GDB) maintainer at Cygnus comes across as a hacker's hacker. John is presently working on GPL cryptographic tools.

 

"Cygnus--We make Free Software affordable. -John Gilmore

 

I bought "Inside AutoCAD" in volume from New Rider's Publishing when the company worked out of a garage/office cell in an industrial park in Thousand Oaks, California. "Embedded Linux" by John Lomgardo is a recent release from New Rider's.

 

"Free for All" (or The Geeks versus the Suits)

 

Yes, I stole the title from Peter Wagner's great book about the free software movement and Linux (great programmers steal). Tim and RMS are credited with being reviewing readers in the same paragraph. I think "Free for All" is more objective than "Open Sources" because Wagner isn't a stakeholder in the open-source movement and has no agenda to promote. The book puts the Open Source Initiative in a different perspective, while "Open Sources" was written by the "winners" who (re)write history. "Free for All" is very well written describing the conflicts, politics, and who got the money and glory. The pendulum swung towards the suits' side.

 

The "save the world" hackers (and that really used to mean the hackers' world) as personified by RMS and some followers still believe that those who have gone over to proprietary software and money, as opposed to supporting wealth for everyone, have succumbed to the Dark Side. This argument is a little hard to understand for somebody that went to a land-grant university just to get out of the stinky paper mill town that he grew up in. At least "Free for All" dares to say that there seems to be some inequity in who got the money in the companies benefiting from Open Source software. The question of whether Red Hat's value-add to GNU/Linux is worth a market cap of (at one time) $9B is also absent in "Open Sources".

 

Peter Wagner refers to Linus Torvalds by his last name explaining that writing etiquette requires this convention. I like "Free for All' well enough that I'm going to extend the same courtesy shown Linus in Open Source literature and call him Peter, even though his name isn't as distinct as Elvis or Linus. Peter nominates Buffalo Springfield's "For What It's Worth" as his candidate for the free software movement's anthem: "There's something happening here/ What it is ain't exactly clear." Peter's last paragraph in "Free for All" is a sort of rallying cry for Fremen everywhere. Some of the expressions may sound strange if you haven't read the whole book.

 

"For all these reasons, this grand free-for-all, this great swapfest of software, this wonderful nonstop slumber party of cooperative knowledge creation, this incredible science project on steroids will grow in strange leaps and unexpected bounds until it swallows the world. There will be battles, there will be laws, there will be martyrs, there will be heroes, and there will be traitors. But in the end, information just wants to be free. That's what we love about it."

 

.NYET

An Internet Operating System

 

.NET, dot net, is Microsoft's Web Services platform, with integrated Simple Object Access Protocol (SOAP) and XML support. Tim thinks "Microsoft has laid out the clearest and most ambitious strategy" to control the OS of the Net, the network-centric operating system layer. See "An Interview with Tim O'Reilly" at oreilly.com. O'Reilly has long covered Java, Open Source, and the Unix federation. Actually O'Reilly is just a state of mind and Microsoft (and Mac) programming gurus and wannabees could use some "animal" books. Now O'Reilly feels that the O'Reilly approach can and should cover .NET and has published many Microsoft books with more in the pipeline. See dotnet.oreilly.com. May I suggest the use of an image of a 900-pound gorilla on the book covers? I'll rewrite this section of the article if somebody threatens my pizza-eating privileges at the Microsoft .NET User Group.

 

Microsoft's Waltham office has moved to 201 Jones Road, on the 6^th floor. Here we are at Microsoft's new office in Waltham to celebrate remotely Bill Gate's rollout of Visual Studio.NET in San Francisco. There's a lot more room in the new offices but I still like the old office with its view of the reservoir and Route 128 traffic. After a few dozen O'Reilly C# books and many t-shirts with a tarsier on the back were given away for any reason that came to mind for the various presenters, we watched a partial showing of Bill's speech in `frisco. The author of O'Reilly's C# Essentials book joined Chris Pels, the Boston .NET user group director, and other luminaries to answer questions about the whole life cycle and evolution of .NET. When it came time for the BIG raffle the second prize was the Visual Studio.NET enterprise edition. It lists for around $1400 but if you add up the value of the individual packages it's value approaches $50K. First prize was the Microsoft Xbox game station. All in all a first-class roll out even if the Rolling Stones didn't provide the theme music.

802.11b

 

Tim mentioned a new book that O'Reilly wanted to call "802.11b". O'Reilly was negotiating with the IEEE which objected to the use of their trademark in the title. The IEEE has trademarked the ethernet standard and other standards that they maintain. I thought at first that "Building Wireless Community Networks" (Carolina parakeet) by Pat Finkenger, an O'Reilly network administrator, might be this book. The winter O'Reilly catalog lists "802.11b Wireless Networks: The Definitive Guide" (images of bats) by Matthew Gast as being released in February 2002.

 

The local Boston user community is showing a growing interest in Wi-Fi^TM, the wireless Internet infrastructure based on the 802.11b standard. In local academia, about 90 MIT researchers have joined with Hewlett Packard in a Wireless Initiative that should greatly expand what wireless technology does today. There is some spillover from the MIT graduate students' 802.11b (not MIT's) net in the Muddy Charles pub, allowing sporadic connection to the MITnet.

 

Free Software, Paid Support

 

This is a no-brainer. Consultants have always written basic books and white papers on subjects that they were selling their expertise on. Once a project manager knows a little more about the problem, say a Java app rollout, he also knows an expert to call on, to give his group a jump-start. The "we're better than your engineers" approach didn't work that well and has been replaced by "we can give your staff a running start". Sell service, give away the software (or support free software).

 

BNUG's Virtual Private Networks (VPN) and Firewall Seminar

 

"Building Internet Firewalls" by Elizabeth D. Zwicky, Simon Cooper & D. Brent Chapman and "Virtual Private Networks" by Charlie Scott, Paul Wolfe & Mike Erwin are in their second editions.

 

The flyer given out at the BNUG December General Meeting described the BNUG seminar as being "Defending your Network: Network Security and VPNs". Denise Olliffe, O'Reilly Associate Publicist/User Group Manager, sent BNUG four O'Reilly books for door prizes for this BNUG seminar which was presented Saturday, the 2^nd of February, 2002. "Incident Response" by Kenneth R. van Wyk & Richard Forno wasn't in the stack. Snort and Alerting Dragon.

 

Cisco presented their version of the BNUG seminar at the Boston Sheraton. Each participant was given a copy of "Cisco Security and VPN Solutions Folio", a collection of excerpts from Cisco Press books and some CDs. I don't think Cisco knows that they're infringing on the Cisco Press (a newspaper in Cisco, Texas) name. Maybe I'll buy the newspaper and force a nuisance suite on Cisco.

Bioinspiration

 

Man builds with concrete building blocks and ttl circuits while nature uses cells. Many processes that challenge us today have already been solved in biological systems at "normal temperatures" with a minimal expenditure of energy and without producing toxic waste. Tim identified computer applications in bioinformatics (computational biology) as being the next rapid growth market. O'Reilly produced his Bioinformatic Technology Conference this year putting his money... This conference took place January 28-31, 2002 in Tucson, Arizona. Tim feels that genomics, proteomics, and other biotechnological projects will be the first revolution of the new millennium. The rumor mill has it that while the "redundant" part of the workforce in the early 90s got Novell certification (often paid for by the state) their counterparts in the new millennium have been seen madly studying O'Reilly's two bioinformatics books.

 

If you attended the conference, I hope that you allowed yourself enough time to visit the Biosphere II and the only public-viewable ICBM (in its silo) in the "free" world (on the road to Nogales).

 

Trade School on the Charles

 

Tim's presentation in Twenty Chimneys (room 306) is over, we've eaten all the cookies and crackers and drunk all of the soft drinks that we can. There's nothing left to do but raffle off the O'Reilly books and t-shirts. A student representative for MIT's ACM/IEEE chapter, today's host, comes up with the first question: "What book has the MIT mascot on it?" Hint 1: MIT students are very energetic, and keep themselves as busy as they can constructing dams, structures, virtual worlds, etc. Music's over. Hint 2: The Society of Women Engineers at MIT has an annual Beaver Dash in April. Right, a beaver. Now for the hard part. What O'Reilly book has the beaver woodcut on the cover? Don't feel bad--nobody got that one. It's "Windows NT Event Logging".

 

I received my Midwest handle of "tree hugger" (on the Great windy Plains "tree hugger" describes a spectrum from environmentalist to ecoterrorist) because of my defense of a beaver that was chomping all the landscaping at the Allied Signal facility near Kansas City. Being a card-carrying member of the Sierra Club is, of course, still safer in Dole/Limbaugh territory than being pegged as a Boston liberal. I'll be forever indebted to the blond cashier in the cafeteria who finally upgraded me to the more endearing handle of "big-time slacker". Okay, so I can't afford to buy a Segway.

The Invisible Computers

 

I have worked for three different corporations that produce Flight Management Systems, the minicomputers that sit at the top of the hierarchy of all the other computers in commercial jets. Obviously the term minicomputer is outdated with the explosive growth in microcomputer power. An FMS is less critical than an inertial navigation system (INS) or a global positioning system (GPS) that outputs the aircraft's position in 3-space. The flight control system (FCS), which stabilizes the aircraft and receives commands from higher systems is still more critical; the gyros that sense which direction is up and down end up being the most critical. INSs have been replaced by inertial reference systems (IRSs), a new digital ARINC standard and a new technology, laser gyros. The less critical a system is the easier it is to certify so the certification game involves pointing to more critical systems and proving that they will supply the critical function, not the system you're trying to certify.

 

There are 5 levels of criticality, a through e, with a being the most critical. The functions of a certain level are normally isolated in one box, and a federated system of boxes is integrated. It's a whole new game when functions of various levels are combined to share processing power and reduce space and weight. Now the real-time operating system (RTOS) has to guarantee that functions of less criticality can't compromise those of greater criticality.

 

Safety policies with FCSs won't allow any tampering once the system is certified and properly accepted into an aircraft type. Does this mean that the code will always remain proprietary and closed? Probably, since the knowledge encoded in the boxes is Honeywell's crown jewels, allowing Honeywell to pretty much control all the avionics in the majority of the free world's jumbo jets. We already saw that some Open System licenses allow the use of binary-only software derived from open-source software.

 

I've always considered embedded system software to be an Alchemist's spell that turns cheap sand into magical devices. The crime of proprietary software was that I couldn't see the magical chants and if I did or developed a spell on my own, my paymaster, by law and contract, considered it to be his property.

Hidden in Plain View

 

O'Reilly's offering for embedded systems is weak so far. While the summer catalog had 3 Palm books, there was only the solitary "Programming Embedded Systems in C and C++" by Michael Barr. Michael Barr will be presenting classes on Java, Programming Embedded Systems, and J2ME at the Embedded Systems Conference (ESC) San Francisco which runs from March 12 to March 12, 2002. I assume he'll be at ESC-Chicago and ESC-Boston (Hynes) as well. Michael Tiemann of Cygnus Solution was a keynote speaker at last year's ESC-Boston.

 

First Microsoft and now Linux (among others) is targeting the embedded consumer products space. It's probably time for somebody (like John Gilmore) at Cygnus Solutions to step up to the plate with an "animal" book on Linux embedded systems.

 

Real Fast vs Real-time

 

Even the smallest Linux kernal is too big to be real-time. It becomes a question of what is fast enough; Linux may be fast enough for an FMS but too slow for the inner loops of IRS and GPS even using the tricks of time-stamping data. All of the past experience has to be "refactored" in the light of faster microprocessors, superior algorithms, and fast, small, and cheap RAM.

XML-RPC: XML over HTTP

 

Dave Winer, CEO of Userland Software, describes the development of RPC extensions at his company in the forword of "Programming Web Services with XML-RPC". I've kept an eye open for jelly fish ever since I had my arm temporarily paralyzed by a jelly fish while swimming at Waikiki in the "lumber yard". There's a cousin of that Honolulu jelly fish on the cover of "XML-RPC" by the XML regular Simon St. Laurent, Joe Johnston & Ed Dumbill. Dave deigns to talk about the simultaneous development of SOAP with the excuse that he's saving that for another O'Reilly book. This cop-out is softened by the explanation that XML-RPC is simpler and ready for immediate use while SOAP is more complex.

 

The authors of "XML-RPC" are the first to state that a community develops the content of their books that they have their names on. And this statement goes beyond "hacker humility".

 

Web Services: Enabling Computers to Communicate

 

"XML-RPC reuses infrastructure that was originally created for communications between humans to support communications between programs or computers." The implications and criticism on this reuse is covered in the book as well as the major programming environments--J2EE, JavaScript, Python, Perl, Java, and ASP that need to communicate with each other. XML-RPC is to the Internet what VBScript is to the Microsoft environment.

 

Web Services is just another of the bleeding edges that O'Reilly has caught the wave of. Another closely related technology is peer-to-peer (P2P) networking, which was the theme for an O'Reilly conference this fall.

P2P = Presence, Identity, and Edge (PIE)

 

Chapter 1, by Dale Dougherty, of the "2001 P2P Networking Overview" by Clay Shirky, Kelly Truelove, Roel Dornfest & Lucas Gorge is available for download on oreilly.com.

Some pundits considered the Web broken because of its centralization. Napster showed that a new mindset had arrived with the generation that grew up with the Web. Web Services tend to concentrate content and power in portals but P2P pushes content to the edge--the PCs, users, and content at every leaf of the Net. Napster wasn't completely decentralized; it had coordinating servers to get users and music files together. This architecture made Napster the most scaleable software application in the world by using the cumulative power of the edge resources.

 

Identity is the ability to identify resources even within the limitations of IP4 addressing, while presence is the capability of determining if a resource (or recipient) is available. For example, Instant Messaging (IM) allows a sender to locate a resource and know if the resource is available.

 

Microsoft's .NET framework, Microsoft Messenger (Pie), and Passport identity services (pIe) seem to be forward positioned for the P & I of the PIE battlefield even though they were designed primarily for Web Services.

 

Man with the Golden Peanut

 

The Twenty Chimneys room at the Student Union has cleared out. Nothing left but to go to the Middle East to watch the navel maneuvers or to the Cantab Lounge on Massachusetts Avenue near Central Square to see the legendary Little Jim Cook (Peanut, ah oh, ah oh, ah oh). Jim wears a golden peanut on a gold chain commemorating his hit record from the 50s.

 

Bookmark these sites!

 

bsd.org - Berkeley Software Distribution Unix resources

conferences.oreilly.com - O'Reilly's Bioinformatics Conference

dotnet.oreilly.com - .NET articles and book reviews

emlinux.com/specialized_linux_projects.html - Embedded Linux

fsf.org/manifesto.html - RMS's GNU manifesto

ibm.com/developer/ - IBM's Open Source/Linux zone

java.sun.com/learning/ - Java tutorials from Sun

linux-embedded.com/software.php3 - Embedded Linux

linuxrouter.org - The Linux Router Project

loaf.ecks.org - Linux on a floppy (LOAF)

MetaLab.unc.edu/LDP/intro.html - Linux Documentation Project (LDP) page at UNC's MetaLab

opensource.org/Halloween - Microsoft Internal Memo--aka the Holloween memo

oreillynet.com/mac - Mac articles and reviews of Mac books

oreillynet.com/wireless - Wireless data developments and LANs

perl.com - Perl articles and Larry Walls preview of Perl 6

python.org - Python programming--a good first language

stromian.com/copyleft.htm - D. K. Rosenbergs "Copyleft and the Religious Wars of the 21^st Century"

tuxedo.org/~esr - Eric Raymond's home site

xml.com - XML tutorials, articles and white papers

xmlrpc.com - Resources for XML-RPC

"Use the Source, Luke."

 

I was going to name my distribution "Marx nix" (pronounced macht's nichts) but I was afraid I'd be called a communist. The 3 CD package sells for $15 and includes a limited-edition black designer t-shirt compressed into a hexagon. The instruction book suggests that you remove the t-shirt before inserting the CDs into the CD-ROM drive.

 

Quiz: Name a cathedral and a bazaar.

Can a suit be free?

 

How am I driving? kenneydennis@hotmail.com Updated January 21, 2002.

 

Eoff (end of flame fest :-) --from the final email (Niels Skov Olsen) of the Tanenbaum-Torvalds Debate as reprinted in the appendix of the "Open Source" book.