An Epitaph for Be
Thursday, August 16, 2001 by Dave Winer.
What a great network we have. At 6:30AM Pacific I got an email from a reader with a copy of a press release saying that Be had been acquired by Palm. I went to the PR Newswire website, found the press release and linked to it. I went to other news sites to see what they had -- there was nothing.
This column is a little slower than the website because it goes through email; so I assume by now that you've heard that the assets and key personnel of Be, Inc. have been acquired by Palm Computing, for $11 million in stock. I'm gathering pointers to story about this event, on Scripting News today.
Expect some passionate writing -- this company evoked our better nature. They gave us a little guy to root for, even though many believed they were doomed. Expect some interesting writing as the day goes by.
And of course, I'd like to do my part. ;->
Be almost got it right.
While they were struggling to gain traction; and then to be acquired, I was reluctant to say publicly what went wrong at Be, why their platform didn't gain enough traction with developers to make a difference. This topic, how to create a successful platform, has been much on my mind through this summer, and will continue to be, probably through the end of the year and beyond.
Be's failure to gather developer support wasn't caused by lack of interest on the part of developers. Be was a carefully watched and much admired platform. It wasn't open source, and it's possible that if it were, that might have saved them, but I don't think that was the root cause for its failure with developers.
Here's the reason. Jean-Louis Gassee learned too much from his experience at Apple. In the late 80s when he was Apple's chief product exec, Apple was paying huge license fees to Adobe for the Postscript interpreter that was baked into the ROM of each LaserWriter printer. I heard estimates of $1000 per printer, going straight to Adobe. From this, Gassee incorrectly concluded that it wasn't safe to let developers have control of a strategic part of their technology. This meant that Be had to go alone into the market. When Jean-Louis promoted his platform to developers, it was always with a suspicious eye, trying to avoid another experience like the one Apple had with Adobe. When he promoted it to the public, he had little to sell beyond what was developed internally at Be. The platform never got the boost that developers can provide.
We had quite a few discussions about UserLand developing for his platform, dating back to the founding of both our companies in 1988. He was on the Board of Directors of my company. Were confidants back in those days but sometimes I felt like I was his therapist, constantly reminding him that "I am not John Warnock" but somehow I don't think it sunk in.
Yesterday I wrote about what platform vendors must do to earn the trust of developers, and today the issue is flipped around -- how can we make it safe for platform vendors? Is there a way to create a platform in 2001 that returns value to its shareholders and at the same time creates a safe environment for developers? Both things must happen in order for a platform to gain and hold traction, imho.
To the extent that Be worked, I believe it is because JLG gives great developer conference.
In 1995, I had given up Be as a goner when I got news from the Agenda conference. Everyone was excited about Be. "Amazing," I said. "And I know why."
Be didn't have a credible strategy in 1995, but most of the people at Agenda had never seen a Jean-Louis stage show. He's the best there is -- he knows that he's a pimp, not a whore. (Those are his words, sexual metaphors are deeply and justifiably ingrained in the JLG developer philosophy.)
Jean-Louis talked to the bankers, entrepreneurs and reporters at Agenda as if they were developers, and they loved him, which comes as no surprise to me. He is the best harvester of love I've ever seen.
Someday, if I'm lucky, I will get to do a developer conference myself, and if that day comes, my role model will be Jean-Louis.
OK, maybe Be, under its new ownership will take a chance and release the source code for the operating system under the Apache license. This would be a good move. They only paid $11 million for the company, and the key asset of the company hasn't done very well without an open source license. Perhaps it's time to try the other method.
Regardless, the leading edge is not at the device-driver and GUI level in 2001. Maybe it will be again in 2005 if we're successful at creating diversity in the layer that's being developed now, network services for desktop apps.
PS: Imho stands for In My Humble Opinion.
PPS: JLG stands for Jean-Louis Gassee.
PPPS: GUI stands for Graphic User Interface.