How was BingoBo's technical solution worked out? (Part I)
With all the requirements being defined in place, it could be overwhelming to design such an architecture in order to implement all the features to fulfill them. It's not a simple prototype created in a few months, it's based on 16 years of work through deep thinking and massaging. Plus, there were overtures that I grabbed during my previous work experience, which led me toward a better solution to put the puzzle together.
I'd like to thank all the people who worked with me before, especially those who directly or indirectly had an impact on the idea and worked toward a complete solution. Looking back, it's hard to imagine how things have come along so harmoniously that brought the idea to light . Maybe I did filter out the noise around it. I also feel grateful to the people who forced me to do non-development work, such as Release Manager to manage the build process and Quality Assurance (QA) work of a software product, including the integration work involved with three independent ERP systems. All of these gave me the competitive edge to do what seemed impossible to accomplish.
It was such a sweet memory when I was working in Singapore. I first started working on various Internet protocols, such as HTTP, FTP, SMTP, POP3 and TCP/IP etc., and implemented them into a common interface in C. It helped me to build a solid foundation for the state-of-the-art web technologies.
Soon after that, I joined a team to build a multilingual search engine called iAgent led by Dr. KF Lai from the University of Wisconsin-Madison. I was assigned with a component called SpiderTalk that extended JAT 0.2, a Java Agent Framework from Stanford Ph.D student Rob Frost. The SpiderTalk enhanced the JAT multi-agent communication system with broadcasting functionality and applied to iAgent for exchanging search results among distributed local search engines from Singapore, Malaysia, Indonesia, etc.
The job was well done and I exchanged emails with Rob Frost about the extended work while he was about to leave the research lab from Stanford. Then I moved onto the KQML and JAT Light online forum to discuss the further development of the Agent Technology in the field. I talked with Charles Petrie many times and he invited me to visit Stanford whenever I got the chance. Although it's no more than a friendly gesture to show his acknowledgement of what I was doing, it's a great incentive for me to continue on the design and development of Agent systems in the later years.
Why do I mention these points? I have to say, without them, there would be no BingoBo. In those years, among researchers and professionals in the Agent Technology group, we couldn't see the light to find the way out. At that time, we hadn't had XML yet. KQML is the best for handling message communication between peer to peer agents and brokers. New interpreters must be developed and downloaded for each new type of KQML message being added. Later on, when someone proposed to use XML to implement the message exchange protocol, we all laughed at him. Now we are proved wrong.
Technology evolves in a hard way with so many people involved and so many failing endeavors. We are standing on top of other people's shoulders to reach higher and higher. It's our moral duty to remove the frictions along the startup's way. I will make every effort to bring the BingoBo solution to the audience to help in their daily life.