Nice – a writable world by Terry Jones

www.fluidinfo.com is a new database where anyone can add data – hence “a writable world”. Think of data being the underlying data and the meta data – in Terry’s world the two are dealt with the same way – an article and a rating, a film and a comment. Anyone can write to the database and if that data is public then anyone can search on it.

The big difference here is that the database does not have complex tables and schemas – any filds can be added and then searched upon. In a typical database the  developers have to design the schema to be able to anticipate what you will be adding or searching for.

Show me all the web articles that I have liked that my Facebook friends like? Facebook would need to update their database, expose this information through their custom API and then another developer would need to create a database of information ralating to your favourite articles  so that the search query could combine and search across the two sets of data. Data would sit in proprietary silos and our freedom to connect information would be limited to what was exposed via the applications various customer APIs.

Terry’s approach allows all applications to use the same API (Fluidinfo’s) and for us to search across this data in ways that don’t have to be anticipated in advance!

Sounds great – watch the video – http://www.kyte.tv/ch/6118/284808

Bringing the open source model to data

The battleground today is for users’ data. Google uses this to great effect and has now made a grab for users’ browsing data with Google Chrome.

The various rumours about social networking sites sharing data have proved unfounded. The data is core to their business model and not for sharing!

How about a product that is open source and captures  data on the community for the community. A product that is entirely open and transparent about what data is captured and how it is used? What if this product allowed users to determine how they shared their data? What if this product charged commercial vendors for use of this data with the money going back to the community? Watch this space!

The Innovators Dilemma – again!

Reading through a piece about Microsoft’s SaaS initiatives I couldn’t help thinking of the great book from the mid 90s written by Christensen called The Innovators Dilemma, When New Technologies Cause Great Firms To Fail.

The only way for Microsoft to counter the threat of a new technology like software a a service (“SaaS”) is to create a separate division with transparent licensing over the Office suite and a lot of autonomy to go make it work! I would prefer a model where Microsoft supports partners to provide MS under a SaaS model with MS providing these partners with integration for users across products and services and a certification program rather like Salesforce with its App Exchange. A business with Google Email would ask its MS partner to easily integrate this with the Corporate Wiki, MS Calender and Push Email product. This would hurt Google more than MS by ruining Google’s “suite of good enough products” strategy but more on this later. Best of breed will beat best of suite.

Unstructued data – how to benefit or what has “plunge and squish” got to do with it all anyway?

Back in 1992 David Gelernter wrote a wonderful book entitled “Mirrors Worlds”. Gelernter’s work influenced the Sun team and the creation of JavaSpaces (technology that I based my first start-up on).

In this book Gelernter talks about “plunge and squish”. Imagine you have a huge vat of data that from a distance appears homogeneous but upon close inspection is a massively heterogeneous, some structured, some unstructured. You wish to know what data in this vat has a strong relationship to something or other, a particular thought, question or subject that you have in mind for instance. How do you find this related data? Gelernter imagined a system that allowed you to plunge your “something” into the vat of data and see the related data attracted or squished towards your “something” allowing you then to extract that related data to find the answer to your question. This is all very theoretical but still important in that to find a solution to a problem it helps to imagine what your solution looks or feels like. Gelernter didn’t stop there but went on to create Scopeware which was the equivalent of Google Desktop however launched well before Google’s product. Scopeware then disappeared from view (purchased by some group and the technology integrated into their product set) being faced with the competition from Google’s and Microsoft’s products in this area.

Scopeware was a first step in the road-map to solving the problem of data mining in that data was pre-searched during computer downtime and the user then used keyword searches to search across all document types desired. This is a first generation solution with the pre-search facility clearly helping in speeding up the searches. A second generation solution to this problem has not yet, as far as I am aware, been launched but I believe it is now worth investigating. BTW if anyone feels that second gen solutions are here already then do let me know.

One area that I am interested in and have been brainstorming with a friend of mine is in the area of knowledge management. How sensible is it to spend time researching a topic when someone maybe only two-steps down the office has done this research and has done so from within a similar context to your own? It is only sensible if the person is not able to be found by you simply and rapidly. If an enterprise system could quickly and efficiently find the person within the enterprise with the some or all of the knowledge that you are seeking to find this would potentially be of significant value. The identification of the person is the key to finding the information. Sometimes simple answers to simple questions are the most tricky to solve.

It’s done! Google plug-in is off my browser! For how long?

Yes, at last, after 8 years I have left the world of Google. The plug-in is going from my browser and I have installed snap in its place. It has good features, is slick, returns quality results, allows you to junk results and gives you a preview of pages before visiting them. What’s more it is founded byt he guy that invented pay-per-click. Loving IT.

Let’s see how long Google does stay off my desktop and how long before snap gets polluted as Google searches have with advertising.

Do VCs fund software?

It is now possible and desirable to build small pieces of valuable software and distribute them widely over the Web. How the software works with other related software is crucailly important determinant of success. Traditional marketing, channel management and enterprise software development processes are irrelevant in the early-stage.

How do VCs fund this type of business. The answer is that they don’t. VCs stick to either the more traditional enterprise software businesses (for which there has been a great exit market in recent years) or only invest after the software has gained traction with users.

Advice to entrepreneurs – pare down your software such that it provides a minimum value to users then launch it to early-adopters. Work hard to interact with early-adopters, learn from their feedback and carry on developing incrementally until the user base is singing its praises. Once at this stage think about how to fund the business. Up until that point you need to fund it yourself – importantly though, doing so is no longer an expensive task unless you are trying to do too much.