PeopleRank: the new directory inquiries

dierc.JPG

Introduction

In this article, we discuss a new form of search engine: a search engine of (verified) tags. Here, tags have the same meaning as they do on flickr etc except that they are verified and they could connote either businesses or people. This search engine is driven by an (as yet) hypothetical internal algorithm which I call for the lack of a better word, ‘PeopleRank’ (analogous to Google ‘page rank’). Much as I hate the word ‘PeopleRank’, it portrays the meaning I want to convey due to it’s similarity to the Google page rank .

The basic idea is ‘tags’ could be used to make calls from the web to any number. A directory based on an algorithm(which I have called ‘PeopleRank’) would act as a user interface(and in effect a virtual operator) and place the calls.

Its more about convenience. When Amazon was first launched, many people thought Amazon was about cheaper books. Today, we don’t necessarily expect cheap books from Amazon – but we get ‘something more’ i.e. the choice. Similarly, with this concept – you get all the numbers in one place, you get presence information and you can call the numbers from the web.

In a nutshell, PeopleRank indicates a complex algorithm for classifying people/businesses(a directory) but it also incorporates their profile/presence information. Thus, the PeopleRank directory/algorithm knows who you are, the context(for instance; where you are ) and how you wish to be contacted. Such a mechanism is driven from the web and works at the application level of the stack(in other words, it is not ‘fixed to mobile convergence’). It takes away the locus of power from the Mobile network operators and back to the user through a profile.

Central to the discussion below is the concept : I am not a number, I am a tag

The idea of ‘I am not a number, I am a tag’ is highly disruptive to the telecoms industry as we shall see below. It was first put forward by my co-author Tony Fish . This article expands on the original ideas and discusses potential mechanisms that would make this vision possible

The idea of ‘I am not a number – I am a tag’ was first discussed HERE

I am not a number – implementation

The idea of ‘I am not a number – I am a tag’ sounds enticing enough. But how would it be implemented?

Let us break down the functions

a) Profile: We need a mechanism which stores user information including context(location, presence etc) i.e. a profile

b) Identity: We need a trusted identity i.e. the ability to reliably identify the person and their information we want to call

c) Addressing: Finally, we need the ability to discover and address the mobile device independently of a telecoms operator(else it defeats the purpose and the idea no longer becomes disruptive). I call this ‘bridging the wireless gap’ i.e. the last wireless mile

Profile

The profile requirement is the simplest. Already, many social business networks maintain a profile. That profile can be used to store a range of contact information which may be made public or released to a selected audience. Presence information could be manually updated or could be set via rules.

Identity

Identity is a more difficult concept and needs a lot of thought. Recently, I flew across to the USA to speak at the Ajax seminar in New York. At the US immigration, homeland security has the new fingerprint system. At our end(UK), we have a new voluntary scheme called IRIS (pdf) which scans the retina of the eye. I joined that as well.

Thus, in essence, two entities(the UK govt and the US govt) can identify me from biometric information. This kind of identification arises from primary sources(biometric) and is validated by organizations which are accepted and recognised globally (the US and UK governments). Hence, my primary identity is defined by these primary facts(biometric information, birth certificate etc). This leads to conferred identity(secondary identity) such as driving license or passport. Providing the issuing authority of these documents such as a driving licence or a passport is recognised and accepted, the secondary identity is also as good as my primary(biometric) identity. Ofcourse, there is a risk that the secondary identity may be forged(such as fake passports). That risk, coupled with increased global threats, have caused many governments to revert back to primary identity(retina scanning/finger prints). These of course, cannot be forged.

Identity in the digital world

On the web, identity is still relatively primitive. Identity is what a system knows about me at the time of registration. This may not be true. Every time I register with a new system/site, I create a new identity. Each of these identities are not tied to any biometric information(currently). They are silos because each site has its own membership system. The obvious way to solve this was the ‘single signon’ method. Microsoft attempted this through their passport mechanism. But that did not receive much response. Initiatives like the liberty alliance are a consortia of companies. But .. Consortia often have an important member missing, in this case, it’s Microsoft.

Ideally, users should control their own identity. The current thinking around identity is to work towards ‘Open Identity’: with users controlling their own profiles and storing their data at an independent place on the web. Open identity is still at an early stage. However, from the perspective of mobile web 2.0, a simpler approach could suffice. Let us not forget that we are making a phone call here! A simpler interpretation of identity may suffice for our purposes. That identity could be ‘reputation’ and this leads us to the idea of ‘verified tags’.

Verified tags

As we have seen before, conferred identity (for example driving licence), can be used instead of the primary identity. For example, conferred identity can be used to buy a range of services in the physical world(for instance you can show your driving license to buy services). In the digital world, tags / avatars etc are all a form of identity. However, they are not verified.

To achieve the concept of ‘I am not a number ..’ you need a verified tag.

How can we verify the tag? You could use something like the Liberty alliance or similar mechanism. But that’s too ‘top down’, complex and expensive. But let us put this in perspective first .. A phone call is not a transaction!. The stakes are a lot less lower. In this case, rather than a full fledged approach(such as liberty alliance), which is expensive, a simpler more organic approach could suffice. This approach is based on the concept of ‘Identity = reputation’.

Reputation is what others say about me on the web. That’s an organic way to ‘verify’ me i.e. my tag. A verified tag allows me to be identified and verified, but not because who I say I am, but because others can validate what I say. This makes the mechanism a ‘closed loop’ system. In contrast, the Internet allows me to set up an ID called ‘John Smith’, without proof. I can then take someone’s verified ID can communicate to another channel. As this is open; it can be abused.

The next logical question is: how to capture that ‘reputation’? And here, lets introduce another word ‘Pingerati (or technorati 2.0) as I explained in a previous blog.

An algorithm (PeopleRank) could then process all that information to create a ranking/probability indicating who I am. In my case, the algorithm would process a number of tags such as ‘Mobile web 2.0’, ‘London’, ‘Futuretext’, ‘Ayn Rand’, ‘Tom and Jerry’(the last two because I follow Ayn Rand’s thinking and love ‘Tom and Jerry’ cartoons!’) and ‘identify’ me if you wished to call me. But then again, with a name like ‘Ajit Jaokar’ – I don’t need too much verification – there can’t be too many people by that name :)

Addressing

So, thus far we have seen how we could identify a tag based reputation. Our goal is to make a ‘phone call’ that tag now that we have reliably identified the tag. The question arises, how do ‘place’ this call from the web? If this is a mobile phone, how can we bridge the ‘fixed to mobile’ gap?

From a telecoms perspective, to place an IP call to someone on a mobile IP network, you need their IP address. That address may be mapped to any identified(tag, avatar etc), but an IP address is required. We also need to know presence information: Is the ‘tag’ available to accept phone calls. On first impressions, the requirement of knowing the IP address in advance sounds very daunting. In fact, it may not be so as we show below.

I can think of two ways you can ‘call’ a phone.

a) Skypeout and

b) Naked SIP

Skypeout

The first and the simplest is to use an enhanced version of a mechanism like Skype out http://www.skype.com/products/skypeout/ . Skype out has a profile and can already call any phone including a mobile phone. So, we need the Identity and presence features which are missing at the moment. Presence can also be provided by the user manually or through rules(Between 7 to 9 in the evening, call me on my home number).

Naked SIP“Naked SIP” is SIP without IMS. I have explained the concept and significance of Naked SIP before

Conclusions and Observations

From the above discussion, we see that :

a) I am not a number, I am a tag : can be implemented at multiple levels – either as a simple skypeout call or through naked SIP in future

b) Identity can be implemented organically through ‘Pingerati’

c) To some extent, we already use the web to check identity manually. How many times have to ‘googled’ someone to ‘check out’ who they are? That’s an ID check!

d) Its more about convenience. When Amazon was first launched, many people thought it was about cheaper books. Today, we don’t expect cheap books from Amazon – but we get ‘something more’ i.e. the choice. Similarly, with this concept – you get all the numbers in one place, you get presence information and you can call the numbers from the web.

Acknowledgents:

Dean Bubley:

Marc Canter : Breaking the web wide open