Information Wants to be Valuable:

A Report from the First O'Reilly Perl Conference

by Keith W. Porterfield

The Twilight Zone

Presented for your consideration: a grim scenario of a world without free software...

I recently traveled to San Jose, California to attend the First O'Reilly Perl Conference. Tom Christiansen, Perl guru and co-author of the book that has come to be known as the "Perl Bible" (Programming Perl, 2nd Edition), gave a pre-conference tutorial on Perl programming for the World-Wide Web. During the course of the day, he repeated (with considerable sarcasm) a quip heard from one of the new generation of Internet hypesters: "Without advertising, there wouldn't be an Internet." I'd like to modify that statement slightly:

"Without freeware, there wouldn't be an Internet."

Why such dependence on freeware?

In today's Internet environment, there are commercial software alternatives for most of the functionality that has been historically provided by freeware. But despite a prevalent attitude at management levels that freeware is somehow substandard, software engineers still prefer gcc over expensive commercial products, and webmasters overwhelmingly prefer Apache. In fact, some of them have gone so far as to modify the Apache server to masquerade as a commercial product, fooling their management into thinking that they are actually using the commercial server the company had shelled out big bucks for.

Even if most corporate managers don't recognize the importance of freeware, Bill Gates does. At last year's annual Microsoft meeting with financial analysts he noted that, in the Web server market, "Apache is our biggest competitor. It's gaining market share faster than Netscape." Since January of 1997, the UNIX-only Apache server has consistently held over 40% of the web server market (as measured by The Netcraft Web Server Survey). Though Microsoft's Internet Information Server product currently dominates the growing Windows-NT web server market, Apache will be released for the NT platform by the end of this year. Apache's market penetration here will be telling: both servers are "free", but they are children of radically different parentage.

The preference for freeware over commercial software is driven by two very simple desires: higher quality and more responsive support.

Much freeware is fundamentally better software than its commercial competition.

How could this be true? Essentially, it boils down to this: commercial software developers simply cannot afford the resources that a well-coordinated freeware project brings to bear on the task at hand. Literally hundreds of competent and highly motivated (they're obviously not in it for the money) programmers may participate, contributing architecture and design insights, core program code, bug fixes, and documentation. The freeware philosophy of "release early, release often" to a massive pool of users (instead of the relatively small base of beta testers used in a typical commercial project) enables the finding and subsequent squashing of far more bugs during the development cycle. For a fascinating analysis of this style of software development and why it works, I refer the reader to Eric S. Raymond's article The Cathedral and the Bazaar.

Support for freeware is far superior to that available for most commercial software.

In a recent survey by the market research firm Service Intelligence, questions taken from vendor-published FAQs were posed to the telephone support personnel of several major software companies. About 25% of the time, callers were either given wrong answers or told that the problem was "unsolvable". Most users of commercial software have spent what seems like (and in unfortunate cases, may actually be) hours on hold waiting for answers that don't work, or are never forthcoming.

Most freeware is supported via mailing lists, Usenet newsgroups, or a combination of the two. It's typical to receive correct answers to complex technical questions in less than an hour, often from the very programmer who wrote the code. But the myth that freeware is not well supported still persists, and is often cited as the primary reason for the corporate prejudice against free software.

Culture Clash

The rift between freeware and commercial software seems to be mostly cultural. In his opening remarks at the Conference, O'Reilly & Associates President and founder Tim O'Reilly drew some parallels based in cultural anthropology:

Unfortunately, it is extremely difficult for those inculcated in the mores of either of these two cultural categories to understand, let alone appreciate, the value system of the other. In the case of software development, especially unfortunate because both the commercial and freeware communities have a lot to learn from each other.

Tim O'Reilly has undertaken the mission of finding that middle ground:

Almost always when freeware is pulled within a commercial entity, barriers are built between the commercial entity and the freeware community that cause development to wither. Not only do we recognize this as a mistake from the commercial point of view, it runs completely counter to our historical roots and belief system....We believe we can create a model for how the business side of O'Reilly can live with the freeware side, to the benefit of both. Our past indicates we will.
Perl Q & A, O'Reilly & Associates, http://perl.oreilly.com/Q_&_A.html (NOTE: This URL is no longer valid as of 05/23/2001.)

Larry Wall, the mastermind behind Perl, began working for O'Reilly & Associates in mid 1996 in order to devote his full-time attention to the language. In addition to putting on the Perl Conference, O'Reilly's recent efforts include providing a high-bandwidth home for Tom Christiansen's Perl Language Home Page, a ramped-up publishing schedule for new Perl-related books, and public relations efforts to gain recognition of Perl's importance in the mainstream computer press and 'legitimize' the language from the corporate point of view. Commercial software development tools for Perl are forthcoming from O'Reilly, accompanied by a commitment to release freeware versions, as well.

The Future of Software

"Information wants to be free," is a popular quip among certain factions of the freeware community, but I prefer the modulation that Larry Wall offered in his keynote address at the conference: "Information wants to be valuable." Free software is deprecated by most of the world, and efforts to increase its value in exchange culture terms (make money with it) usually end up cutting off the creativity and support of the freeware culture that created it. Yes, freeware authors would like to be paid for their efforts, but straight-line commercialization of a free software project just doesn't work.

Though I don't believe that freeware is in danger of completely disappearing, the spectre of a world devoid of these creative gifts is frightening. Without the support of commercial entities and professional public relations, freeware products will continue to be under-reported in the media and under-valued in the corporate world. And without an understanding of the intrinsic worth of the freeware development process (and why it engenders high-quality software), most efforts to commercialize free software will be doomed to a dismal mediocrity.

If O'Reilly is right (and I think he is), the future of software lies somewhere in a yet to be explored synergy between the clashing cultures of the freeware and commercial worlds. This pioneering experiment with Perl creates a whole new model, and will go far toward creating that future.


Links and References


Keith Porterfield, formerly the 'Technicalities' columnist for Internet World magazine, is an independent computer consultant. He previously worked as a Systems Architect at Mecklermedia and ManyMedia.