"Yeah, sure it's good, but can you make a living on Open Source?" Industry observers ask that every day, particularly in the wake of recent LinuxCare and Corel turmoil. Ajuba Solutions, formerly Scriptics, is out to model a happier answer.
A week ago, on May 22, a press release from Scriptics Corporation announced its reconstruction of itself as Ajuba Solutions, Inc. Reaction from the engineering crowd that favors the Tcl scripting language at the heart of Ajuba's technology was swift and nearly unanimous: the "suits" had taken over, and the consequences for Tcl would be dire. This, in turn, surprised Ajuba employees, who expected applause for a move they had designed to improve community relations. How did such a divergence arise, and where will it lead?
This isn't the first move for Tcl. John Ousterhout originally created it in the late '80s while a professor of electrical engineering at the University of California at Berkeley. Then, it was often called an "extension language," to emphasize its role within other existing programs.
Tcl has expanded and mutated in many regards. Through all the changes, though, Ousterhout has nurtured its character as the best "glue" available, that is, as a language unsurpassed for connecting together outside resources.
One example: his first substantial project with Tcl, called Tk, has become an indispensable and highly-regarded graphical user interface (GUI) toolkit outside Tcl that's uniquely portable across MacOS, Unix, and Windows, among other operating systems (OSs). Tk is so successful at tieing into "native" GUI resources that several other languages, including Perl and Python, use it as their standard GUI toolkit.
During the second half of the '90s, Sun Microsystems created a new division that employed the Tcl development team. Early in '98, Ousterhout spun the group back out of Sun as a standalone company designed to sell Tcl products and services to the business community. With him as chief executive officer (CEO), it appeared to enjoy operational success, with a well-received training section, several development contracts, and a couple of product lines.
The first released product, TclPro, is a professional quality development and support framework for Tcl that has received warm reviews. The second, whose name changed last week from "Scriptics Connect" to "Ajuba2," is a considerably more specialized business-to-business (B2B) application server with much higher margins and a longer sales cycle.
A continuing challenge through Tcl's history has been to distinguish clearly the differing interests of Ousterhout as the language inventor, the successive organizations which have employed him and the core Tcl development team, and the user community. There's nothing unique about this; similar questions arise for Java, Perl, BeOS, and CORBA, for example. What seems clearest is that Tcl has generally been growing through it all. More books appear on the shelves, download traffic from principal sites increases, the Usenet newsgroups which discuss Tcl grow in volume every year, and so on.
|
Ousterhout's challenge, of course, as CEO of a small venture-capitalized software company, was to deliver returns to his investors. He appears to have found the right niche: B2B is one of the few segments where there seem to be enough customers willing to pay for well-engineered software products. And he has the right technology to sell them, as Tcl's superiority in gluing contributes considerable value to businesses looking to automate disparate processes involving heterogeneous resources. Tcl's adaptability is particularly handy in accessing "legacy" data other development languages don't try to support.
Moreover, Ajuba2 (formerly Scriptics Connect) has the right "hook": XML. The Extensible Markup Language remains as hot a buzzword as the industry has, and Tcl is perfect to exploit that interest. In fact, Tcl is currently the single best language for XML development: far more productive than C or C++, more portable and easier to learn than Java, and more mature in its XML capabilities than any other language.
What should Scriptics/Ajuba do with these advantages? The company brought on board respected industry veteran Tom Thomas as President and CEO, as Ousterhout became Chief Technology Officer (CTO). The company has beefed up its marketing of "value proposition" messages that resonate with its prospective business customers. As Thomas was quoted, "Ajuba is devoting its expertise and resources to provide solutions that transform the way industries operate."
The name change is just another aspect of this. "Scriptics" sounded like a place where programmers were in charge, and that's an uncomfortable location for the clients Ajuba is pursuing now. "Ajuba," on the other hand, is supposed to be only a "simple, memorable name," although it has positive glosses in Arabic, where it means "miracle." The point of "Ajuba" is to solidify the company's position as what Thomas calls "a marketplace player."
There's nothing shocking in this. Only a few open-source business strategies have established serious records for profitability, and technology rarely takes the lead in them. Vignette's products critically rely on Tcl, and it might employ more Tcl specialists than any other company, but it's de-emphasizing Tcl in its latest press releases. Tcl stays strictly in the "back office" of AOL Digital City, Travelocity, and other well-known companies that make crucial use of the language.
|
Ousterhout summarized the transition: "For the Tcl community there isn't really any news in the name change. It's simply the culmination of our transition from a pure technology provider to a solutions provider. We will continue to work closely with the Tcl community to create great Tcl technology, while building high-value B2B integration solutions on top of that." If it's that simple, why did so many programmers respond in the tone of someone whose pet dog had been shot?
Part of the reaction probably stemmed from the fact that Ajuba initially failed to even mention Tcl in their press release or on their web site. Those omissions were misinterpreted by some programmers and fueled inaccurate gripes about licenses and restrictions that simply weren't true. Some members of the Tcl community began to worry that development might stop and that the future of the language might be neglected.
Other fears, better-grounded in reality, also came out. Tcl's fans universally praise its productivity; programmers like wrapping up work quickly. Almost equally universal among Tcl users is the experience of "wasting time" explaining the language. Managers and clients demand to know why Tcl isn't like Visual Basic, or Perl, or C++, in some particular.
Scriptics' role in these often-frustrating conversations was to give comfort as "The Tcl Company." Though Ousterhout judges "there really isn't any news in the name change," Tcl practitioners in the field agree that the absence of a stable Tcl lightning rod makes their persuasive jobs harder. Steve Blinkhorn of Psychometric Research and Development Ltd., for example, tells how difficult it is to communicate the reliability of the Tcl language, when its parent company appears so fickle.
More ambiguous are Tcl's technical prospects. Ajuba depends on Tcl's health, that is, on its correct operation, on a self-sustaining mass of external recruitable users, on its continuing openness to industry trends, and so on. Ajuba appears to have as much incentive as Scriptics ever did to invest in Tcl's maintenance and growth, and at least as much financial ability. How will the community of users react, though? Will they continue to contribute new code and other key deliverables, or will Ajuba's reorganization so damage their motivation as to strip Tcl of its technical edge?
Ajuba employs long-time Tcl developer Jeffrey Hobbs as "Tcl Ambassador." Here's his intent:
For the greater good, it is no longer our (Ajuba's) goal to try and make a buck selling extensions to Tcl. Instead we will be interested in seeing more extensions created and thriving, with the hope that we'll be able to make use of them here. To this end, we will be putting more organizational effort into the Tcl Developer Xchange to spur community support and interest in the language.
The open source dance is a delicate one, with the same players alternating between voluntary action and occasional substantial material gains. Can the pattern stay healthy past an initial burst of enthusiasm?
Zope is a telling example of an apparent open-source success. Digital Creations, Inc., received venture capital backing at roughly the same time as Scriptics was founded. Digital Creations is a consultancy that gives away Zope, a Web application server and object publisher that is its principal intellectual property. Business is good enough, though, that despite receiving no income for the product, Digital Creations is solidly profitable and growing rapidly.
Digital Creations is quite clever about how it presents itself in public. Superficially, it differs from Ajuba in being a service rather than product company, and its language base is Python, generally regarded as a rival to Tcl. In operation, though, it seems to live several of the principles Ajuba aims to exploit. In a recent conversation with Paul Everitt, CEO of Digital Creations, he emphasized:
Zope has its own PR challenges within the Open Source community, of course. On Monday, May 29, Digital Creations issued a press release with ActiveState Corporation announcing its intention to bind Zope to the Perl programming language, as well as Python. This set off a predictable storm among Python advocates who see "language wars" as a zero-sum game they don't want to lose to Perl. In general, though, Zope's users appear to have a reservoir of goodwill for Digital Creations. They largely trust its ability to balance business and engineering needs.
Tcl has lost out on that sort of confidence and "buzz" recently. Python, for example, has grown more rapidly than Tcl over the past couple of years, as measured by conference attendance, download growth, book sales, and so on. Much of the outcry about Ajuba's name change really is concern that Ajuba will acquiesce to Tcl's perceived stagnation. A variant of this sees Ajuba corrupting Tcl by taking the language in a direction that serves only Ajuba. Larry Virden, who maintains the authoritative FAQ collection for the language, speaks for long-time Tcl contributors:
|
Pertinent Links |
As Ajuba moves away from being identified as being driven by technology, other companies who have strong foundational ties to Tcl are concerned that in some way the free nature of Tcl will become less free; that their investment of time and energy will be lost somehow. I get the impression that many will be never be satisfied as long as Tcl is 'held' in the hands of people involved in commercial interests."
Hobbs knows this, and already has plans for new releases of Tcl and TclPro. Ajuba recognizes it needs the external Tcl development community. Will Ajuba achieve the financial success it's after? Will it use that success to reinvest in the programmers who help build its technology? The year 2000 looks to be a true turning point for Tcl.
Cameron Laird is the vice president of Phaseit, Inc. and frequently writes for the O'Reilly Network and other publications.
Discuss this article in the O'Reilly Network General Forum.
Return to the O'Reilly Network Hub.
Copyright © 2007 O'Reilly Media, Inc.