InterBase became an open source product on 25th July 2000. There should have been dancing in the streets and unbounded joy throughout heaven and earth. The release should have brought praise for enlightened and forward thinking management. Instead there was anger and recrimination towards Inprise. What went wrong?
A Recap
Just before Christmas 1999 Dale Fuller's management team had a meeting with the senior executives of the InterBase division and told them of their new business plan. It has been stated by informed personnel that they had decided that the revenue stream was insufficient for the size of the division. They also said they wanted to lay off at least 50% of the current InterBase staff of 39 (including part-timers and contractors ). In addition, there would be no budget to restore the struggling US marketing and sales staff, which had already been reduced to one 3/4 time person and an intern. This was not a workable scenario to grow a product business. It has never been clear to me just what Dale and his team had in mind, but I believe it was intended to cancel the release of 6.0 and sit on the diminishing revenue stream from 5.n for a couple of years.
Possibly unbeknown to the Inprise management team, the InterBase team had been quietly building a sound and profitable (i.e., not loss making) business. One of the original developers of InterBase (Ann Harrison) had continued to provide excellent technical support for the product via the mers list server/news group. The product manager, Bill Karwin had equally spent many hours there helping developers. On the business side Paul Beach, responsible for worldwide sales (non-US) was basically generating about 90% of the total turnover. With little resources he had concentrated on building a formidable network of VARS that typically integrated InterBase with other Borland tools.
Their response (Karwin and Beach) to the new business plan was probably not what was expected. They resigned. Their resignations hit developers and VARS hard. Contractually they couldn't explain why they resigned, but gradually the above picture became clear. Thus the InterBase community was born. Suddenly we discovered that some serious business and serious money was riding on InterBase. Enough people came forward to buy InterBase several times over. A sale was not to be, but the development community was now forming itself into an organised force to be recognised.
Initial Public Offerings
Autumn '99 saw the Linux Bubble on the stock market. Companies associated with Linux and Open Source were IPO'ing like frenzy and seeing huge (paper?) profits on stock value. Dale Fuller's primary duty is to his shareholders and he worked hard to raise the share price. The development of Kylix for the Linux platform was widely touted. The venerable Borland C command line compiler was made available for free. And suddenly he had his top InterBase executives suggesting they Open Source InterBase if he wasn't going to fund further development. I guess he thought about this idea over Christmas and, come New Year, announced that InterBase would indeed be made Open Source. It was also announced that a new company would be formed to manage this. Inprise would hold some 20% of the stock and would help find venture capital to fund a large part of the rest.
Even the bitterest opponents had to admit that this was a smart PR move, if nothing else. It helped the Inprise stock price remain steady, maintaining a half yearly average of well above 10 dollars. (Previously the stock had been languishing, to put it mildly.) The consolidation of the stock price helped cement a good deal for the Corel merger.
In February 2000 it was finally announced that Ann Harrison and Paul Beach would run a new InterBase Management company. The original architect of InterBase, Jim Starkey, was also on board as chief technical advisor. The source would be made available by end of June at the latest and the company spun off as soon as the legal stuff was sorted.
After the black days of last December it looked as though Inprise had finally got things right. An Open Source InterBase was warmly accepted by just about everyone, in and out of the InterBase community. Ann Harrison was unchallenged as the right person to act as the figurehead for the new company. Paul Beach knew the InterBase business and, if anyone could make the open source project work money-wise, he could. Having Starkey back on the team was the icing on the cake. The synergy of a highly respected open source database shipping with every copy of Delphi, CPPBuilder, Kylix and JBuilder looked a winner.
The Bubble Bursts
During the year, it seems as if some of the assumptions Inprise made have changed. The Linux/dot.com stock market bubble started to burst. Then, the Corel deal came apart, much to the relief of many on the Borland side at least. Subsequently it became clear that IPOs were no longer a ticket to instant riches.
With the prospect of a hugely profitable IPO receding far into the distance, it appears as if the Inprise management decided that they should find another way to leverage their InterBase asset. It seems as though the original intention had been to live off the licence and maintenance fee income for the next three years, so they offered the new company a chance to pay this money upfront in one hit by charging them for assets that had not previously been valued as part of the deal. Rumours went around that $10 million was the asking price for such things as the documentation and the test suites. This rather changed the deal, which by that point had been all but signed.
Although the deal collapsed Inprise had still made a public commitment to open source InterBase. The bad PR that would arise from backtracking on that promise would have been huge, so they had little choice but to follow through. Thus the source was released, albeit a month later than promised.
Whatever the reason for the deal failing, it does appear that Inprise has rather shot itself in the foot. By spurning the new InterBase Company in this way, it has engendered the unremitting enmity of the InterBase development community. About 99% of them use Borland tools. Inprise are also likely to alienate some of the Linux development community too. Linux is very close to its Open Source roots and Inprise appear intent on writing the how-not-to guide for commercial companies wishing to climb on the Open Source bandwagon.
Inprise have issued a public statement saying that they intend to continue supporting and developing InterBase. It is difficult to take this seriously, given the events of the last year. The strength of this commitment is evident from the release of the source itself. Several bits were missing and they were released in piecemeal fashion over subsequent days. It was several weeks before all the parts were available in a single download. Worse still, the source would not compile. No serious work had been done to ensure that InterBase could be built outside of Scotts Valley. Things have improved slightly, but not at the speed of the Internet.
What's Happening Now
Within days of InterBase going open source, the InterBase development community had created the Firebird project on sourceforge (http://www.sourceforge.net/projects/firebird). Developers quickly got to grips with the task of fixing all the broken bits in the install routines. Since then, work has started on understanding how InterBase works internally. This will take a while but is a necessary pre-requisite to actually making enhancements or adding new features to the product. Work has also started on creating new ports of InterBase 6.0. Meanwhile, successful builds have been done for FreeBSD and Mac OS X, for instance. Inprise have finally released the core of the InterBase test suite, so it is now possible to certify new builds.
The new company that was not to be, has now formed as IBPhoenix (http://www.ibphoenix.com). Their website evolves so quickly that it is hard to keep pace. It really is the best starting point for all resources concerning IB. They have most of the best InterBase experts available to them and are offering support contracts for companies that wish to purchase them. They are also working closely with the FireBird project to ensure the future development of InterBase.
Interest in and usage of InterBase seems as strong as ever. The events of recent months have done nothing to diminish the range of third party stuff available. There is too much to list here but all the links can be found at the IBPhoenix web site. In addition to a raft of management tools (all built in Delphi) there is support for running IB with PERL, Python and PHP. There are two new ODBC drivers available for InterBase 6.0 and development o fan ADO driver is underway.
The community around InterBase continues to grow. At the last count there were some fourteen sites in the web-ring, covering three languages. Perhaps the best starting point is the InterBase developer initiative at http://www.interbase2000.org. It has many additional resources provided by developers.
It is difficult to predict how things will develop from here. For now, we shall just have to be content with an industrial strength database that is open source and freely deployable. All else aside, this is a major step forward. Let's hope it is not the last.