We found another post on Twine. This is an analysis of what has been posted on the web and blogs. It isn’t a review of Twine since the author hasn’t been able to test it yet. We found it a good digest on what has been published on Twine. It’s too bad that it isn’t an actual review of the service itself.
Peter Rip, one of the VCs behind Radar Networks, give his own impression of Twine. He doesn’t yet use it to its full potential, however he does mention ways that he sends information to Twine. By sending bcc copies of some emails and by using a Twine bookmarklet.
For Twine to be functional and relevant, it will have to be able to process information from other services like del.icio.us accounts, emails and contacts from Gmail and Facebook friends. Working through those things might be a challenge.
First, in answer to the question about any potential relationship between Radar and DOD. Actually Radar Networks was never involved with DoD. In fact, the actual history is that several years ago we were invited to be part of SRI’s CALO project http://www.ai.sri.com/project/CALO — which was funded by DARPA – to invent next generation tools for knowledge workers. We were one small group among approximately 400 researchers in around 30 institutions including many of the leading universities in the US. CALO is a “who’s who in AI” — involving many of the leading academics in the nation. Radar Network’s role was very small — we provided a desktop disk-level RDF triplestore and helped with the ontology and user interface work. All the work we did was open-sourced under LGPL and anyone can see it and use it. You can see all the code here: http://www.ai.sri.com/software/IRIS
IRIS is a smart personal information manager for the desktop. Today IRIS may use only very little of the original codebase and I’m not sure how much, if any of our work, survives in the codebase today. It has been many years since we looked at it. You are welcome to look at that codebase and decide for yourself. As it is LGPL it is freely available and open to scrutiny.
In any case, we were just a tiny part of that project. We later raised venture funding from Paul Allen, and we began development of a completely different codebase. Instead of being oriented towards the PC desktop, like IRIS, we focused on building an “Internet-scale” server architecture and platform for the Semantic Web. The codebase we use today is completely different and not related to CALO or IRIS. Whereas IRIS was about exploring a future semantic desktop, Twine is about the semantic web – it is a totally different approach with completely different goals and constraints.
We are in no way affiliated with CALO or SRI, although we remain very good friends with our colleagues from that project — including some of the best ontologists and AI thinkers in the world.
We were honored to be part of the CALO project – one of the most ambitious academic research initiatives in artificial intelligence in years – and it’s wonderful that there is support for big ambitious breakthroughs from our government. After all many of the technologies we now take for granted as the Internet of today were originally funded by DARPA. Similarly, much of the core technologies of the Semantic Web stem from DARPA research, such as DAML + OIL.
But it should be extremely clear that Radar Networks’ codebase and product is totally separate from that past research. Neither Radar Netorks nor Twine is in any way funded by or affiliated with any government agency or program, and our position on privacy is very clear: we believe in the individual’s right to privacy. We also believe in open standards, and open-source. All of our code has been implemented to be potentially open-sourced in the future and while we haven’t decided if that makes sense to do at this time, there is nothing in our code that we are ashamed of or that would compromise anyone’s security or privacy deliberately. In fact many of the coders here at Radar Networks are hard-core open-source folks, like Peter Royal of the Apache Foundation, and Jim Wissner, our chief architect who is extremely in favor of individual privacy. My own personal views on this subject are equally strong. I am a passionate believer in democracy and the rights and protections granted to individuals by the US Constitution. I am very concerned about the present trend towards eroding those rights and as long as I have any say in the matter, I can promise that my company will do its utmost to protect privacy and the Constitution of the United State. Hopefully this has alleviated any doubts about this issue. If not, feel free to contact me and we can discuss this further until you are completely satisfied.
Regarding the questions about what we have at Radar Networks that is “novel” and “innovative.” There is a lot of new work and intellectual property here. Most of it is related to the Semantic Web platform, the way we do machine learning, the way we enable knowledge management, new ways to improve search results, new ways to do personalization, based on uses of the “semantic graph” we are building. We are not claiming to own the idea of natural language processing (NLP), nor have we patented that. Instead we apply a wide variety of methods including NLP, statistics, graph theory, machine learning, ontologies, in an innovative combination, along with human input and folksonomies. Ultimately we are about open standards, open API’s, and even open-sourcing. This will all be revealed over time. We have been working since 2001 on this. We have a very large amount of code and IP, and over time we will unfold this and release more and more of what we have been working on so that others can play with it. I hope that in the future, when the codebase has evolved further, we can even release an open-source stand-alone version. We’ve done some experiments around that and we have some interesting ideas. I can’t promise that yet because we are not sure it is key to our business model to do it yet — in fact it may just be simpler to focus on being a SaaS (software-as-a-service) business and open the platform via open API’s (which we have and are working on). That is a decision we will make in the future. Regardless of that, for now our mission is very clear: Make Twine an amazing and delightful service for individuals and groups, and then enable outside developers and services to integrate with Twine in numerous ways. For this first phase our focus is really on learning and entering into a conversation with users and developers. I am sure there will many wonderful discoveries and synchronicities along the way, and I guarantee that we cannot imagine them yet. This is a great adventure and we are very excited to be on it, and I look forward to feedback from all of you as we explore this new frontier together.
On future patents:
Regarding our potential future patents, there is a lot of novel IP in our work. It’s not a matter of merely combining existing ideas in a new way. We’ve invented a lot of new stuff especially in the last 2 years, and even quite recently. But who knows what the USPTO will do. I wouldn’t deign to predict that. I’ve seen a lot of patents granted that have absolutely no merit and don’t deserve to exist. I’ve also seen denials of very novel ideas on completely misguided grounds based on examiners not really understanding the domains well enough. There have also been many very legitimate, well-deserved patents. It’s a bit of a chaotic system it seems.
The patent office examiners are overloaded and the entire field of software patent law needs an overhaul. There are too many old patents lying around that nobody is using, making it hard to do anything new without risk of stepping on a landmine. At the same time, the standard for what is patentable is too fuzzy in some cases, and so people try to patent things that just don’t deserve it. There are too many patents being filed and granted.
One solution to the problem would be to make all (past, present and future) software patents have a shorter lifetime — say 4 years max. That would relieve the backlog on the PTO, and would also prevent barriers to progress, competition from innovation that results from dormant patent land-mines lying around. First of all only serious players would bother to apply for a patent if they could only get 4 years of protection. Secondly, if one got a patent, but failed to do anything with it within 4 years, then they would not be able to sit on it and just wait for someone else to sue in the future.
A four-year timeframe would be fair because if an invention is not commercialized and monetized by its inventor within 4 years from being granted a patent in the software space (a long time in our field), then it should be opened up to others. That would really reduce the number of lawsuits (spurious or otherwise), and enable the software industry to flourish.
Software is not like mechanical devices — it evolves much faster. Our present patent system wasn’t designed for software — it just can’t keep up with the pace of software innovation — it was designed for an earlier, more physical-mechanical world. The friction between the pace of the software industry and the intellectual-property process we currently have for it, is actually holding back our emerging new economy. It increases costs and risks for everyone, and reduces the spread and reuse of new ideas. Another area where I feel very concerned is genomics patents and patents around drugs — these are for all intents and purposes very similar to software patents — and it’s very worrisome to see what is going on in that space. When fundamental operations on building blocks of the human, plant and animal kingdom become patentable “inventions” it just seems wrong.
In any case, regardless of my opinions above, I do believe that companies and individuals who invest time and money in inventing things — software or otherwise — should be granted some protection for a limited period of time so that they can have the first shot at monetizing their investment. That’s only fair. Without this protection at least some amount of incentive and investment in early-stage ideas would dry up, and that would hurt innovation and progress just as much as having too much protection.
In any case, until the patent system for software changes (for everyone, equally), then I and my company have to work within the present system. I do believe that our intellectual property is novel, but time will tell. It will probably be many years before we’ll know for sure.