I originally wrote this entry on September 23, 2004, and published it on blogs.sun.com.
I guess since there have been many start-ups in the Silicon Valley (and elsewhere) brought to the rank of credible companies by ex-Sun employees, I need to provide my own creative contribution to this frenzy in some small way. Otherwise, I may not be worth even half a true-blooded Sun guy.
I'd given away one start-up idea earlier on this weblog, and here's another, and there may be more in the future.
Over the last four years, I have been lucky enough to see, first hand and behind the scenes, a great trend in the telecommunications market to move applications of all sorts to the J2EE platform. In fact, I have been involved in a couple of R&D projects with a focus on evolving some of our partners' service platforms to the J2EE environment. This has all been very, very exciting work and some of it has been published in JavaOne presentations. (So, I'm not giving away any trade secrets.) In fact, some of this work fed requirements that led to the advent of the J2EE Connector Architecture 1.5. (Ram Jeyaraman, the lead for that JSR is a superb engineer. We've worked together for some time, beginning with our implementation of GIOP 1.2 in RMI-IIOP, also involving Mr. Anderson.) Now, the concept of [ connectors | telecom protocol stacks ] is congruent (remember your Abstract Algebra and Category Theory?) to [ JDBC drivers | Databases ]. The point I'm trying to make is that just like some corporations started out of Sun as database connectivity companies and then realized they needed to grow in adjacent areas in order to become significant weights in the software world, other start-ups can grow as telecom protocol connectivity companies and grow into significant weights in the telecom services software world. This is particularly important because high-value services involving the web, enterprise, mobility, multi-media, identity and telecommunications are on the roll. I've been so gungho about this idea that I've even suggested (in a moment of madness, I wonder?!) to some people here that if Sun incubated a startup to do this for a live-or-die period of two years, I'll be willing to risk joining it ! ! ! However, this was several months ago . . . time is running out . . . but I don't think the idea is quite yet passe.
It'll also be cool to have some JSRs to standardize the connector APIs above the protocol adapters. (In fact, there is an engineer in Sun/IEC who may help me do this sometime soon.) Having a standardized API will create a true market in these connectors. Yes, I'm not a big risk-taker right now, and it's fun working with the people at Sun, but if you're one, you don't have to let people know where you heard the idea first. Just go ahead and do it! It'll be good for everyone else in the market. Finally, I think it'll be great to stage an open source project that actually accomplishes what I'm saying here. However, in that case, given the telecom market's penchant for standardized components, the JSR route becomes even more necessary for commercial success.