and it’s in lights too…

I’ve done the impossible as far as I saw it a few years ago and totally changed careers. Quite a few people who read this blog would come here because of my old posts when I worked as an IT Architect concentrating on Open Source, my time at SUSE and the Linux books I wrote.

But others will be here for my posts on BuddyPing and the mobile world, my new job. I moved from big iron, open systems, compute clusters, and DR strategy to the tiny, closed, walled gardens to start a mobile social network. Actually, before that I did try my hand at a social network about London and also becoming an online IT retailer (for God’s sake don’t try it, the margins are too small).

A few months ago, I stopped freelancing (gotta pay the mortgage!) and spent all of my time on Ninetyten (the company I setup). The company recently took some seed funding, and we have seen great growth. I have learnt a lot over the past 2 years from talking to numerous people over beer, building our own social network, and the dreaded conference circuit.

I will be posting over the next few weeks my thoughts on social networking, mobility and the startup space in the UK. And this is the first one…

Social Networking is not just Facebook, Bebo and MySpace. Throw that idea away now, as trying to be the next “big guy” is the wrong way to start looking at a startup in this domain. A rough list of the types of social network could look something like:

Open Social Networks
A perfect example of this is MySpace. An Open Social Network is based on friendships, but sways toward pushing people to browse profiles, and then add that person as a friend. You may never have met a friend in real life on MySpace, and can end up p*ssing people off from the plethora of friend requests they get.

These social networks promote this openness through the user experience, and are sometimes used for vanity purposes (Look how many friends I have!).

Social Graphs
Mark Zuckerberg came up with the name, and it describes this type of social network perfectly. A Social Graph describes your real social network. That’s your real friends, acquaintances etc., and it is the more valuable type of social network from Facebook’s point of view, and also the user.

Because social network sites predominantly focus your interaction with the site on your social network, those links have to be strong. A Social Graph is based on the strongest relationships you have with real people. I am only friends with people on Facebook if I see a value in them (e.g. I want to see how my sister is doing at University, I want to arrange a beer with Andy), and if I know them. Unless I stop talking to my sister, or Andy drive me insane, I will find value in using Facebook to organise my social stream.

Facebook, and Social Graph based networks are utilities, and this should be the goal of any start up. Become a utility that people use, not because they are bored, but because it becomes necessary to use your tool to do something with an action.

Context Networks
A Context Based Network is something that tries to provide your social life (and the objects within it) based on contextual information. I will use BuddyPing as an example here, as it is all about on type of context; location.

BuddyPing uses your location to provide some context to the information you are looking at. At a very basic level, it is a friend finder. By aggregating your social stream’s location, we provide you with a quick way to see where all of your friends are, with maps and all that good stuff. BuddyPing is also an Open Social Network, with the bond between people being their location and distance from each other. This allows me to browse for people nearby, viewing their profile to see if they should join my network. This also works for photos (a live stream of photos being taken nearby), places (create a meetup at a place, and send it to all my friends’ mobiles).

Context is important, especially when browsing and searching for information, but is not a requirement in Social Graphs as the bonds you create are forged from existing relationships.

We chose location as it was something I have always been interested in (I like old maps OK! I’m not proud of it, but I am), and as importantly was something that could be modeled in a system. If you think about it, how would you finitely model the idea of “my mood” in a computer system? Would you browse all the profiles of people who are “bored”? How valuable would the results be?

Using context in a social network is a very interesting subject, and will be where strides will be made in the next few years on using context to personalise content in your network. Context benefits both parties, the user (personal experience) and the service provider (targeted advertising). For example, in BuddyPing we could post an ad to a user whose age and location we know, as well as the time of day. We build up a model of the user to target an ad to them. For the user, the unobtrusive (very important!) ad is content that is relevant to them. An example one of our guys gives is “I don’t want to see an ad about Pampers nappies when I’m a 28 year old beer drinker watching Coronation Street”. Ok, you won’t get the targeting right every time, but it’s a damn sight higher hit rate when you can target on multiple variables.

Push Networks

A push network is something like Twitter or Jaiku. It is a simple idea, that is extremely effective, and it solves some of the problems that come along with Open Networks. I am the root of the social network, and publish status updates, or information into a social stream. That social stream can be subscribed to by another person. When I publish information into the stream, all the subscribers receive that update, usually in an instant medium like email or SMS.

This is very heavily based on RSS, but changing the publisher from mass distribution to personal distribution, and changing the transport medium of the information.

I use Twitter to get SMS messages from the BBC Tech News stream, and it’s useful. I know some people who are constantly receiving SMS messages from the many streams they subscribe to.

It is a great idea, but usually has limited scope for expansion of the service from the basic “push information”.

Different Network Models fulfill different problems, it really isn’t just the MySpace/Bebo/Facebook. I’ll be posting more random thoughts in the next few weeks…

Well, after TomTom’s purchase of TeleAtlas earlier in the year, I honestly thought Google would have done this deal as mapping data is so important to them. But left of field comes Nokia, purchasing Navteq for $8 billion.

One thing that amazes me is that 2 companies, who yes have interesting IP take licensed data from copyright owners and aggregate it to licencees are worth that much!

Well, that is one way of lowering the bottom line impact of offering mapping data for free! I knew I was right that location and mobile go hand in hand

Seeing as Aimee and I are living on the south coast at the moment because our purchase of a flat fell through, I have a 2 hour train journey into London every day, which on the face of it sucks quite badly.  The upside is that I get to do really mundane things while speeding through the countryside to the smoke.

One of those things was to clean up the BuddyPing J2ME app and make the 10,000 odd lines of code manageable and change the way it all fits together.  If there are any people out there like me who came from a non-OO world to J2ME and wrote in a monolithic manner, I cannot stress enough how important it is to start off on the right foot.

At the moment, I am doing a half/half approach, part OO, part monolithic to get the benefits of the language, but at the same time maintain the speed of the app itself.  I have been re-designing the code to act on a screen by screen basis with a new class for each big chunk.

For example, in the Meetups code ( a new feature for the 1.0 release ), all the the UI and action code is in one Meetup.java file, which all of a sudden mean I know exactly what is where, and allows me to change things with out the dreaded s/r cycle happening.

Of course, the Network, XML, blah blah blah is in separate classes, but is all abstracted away into their own larger classes for the rest of the system, leaving the Meetup Controller to fire off what it needs to do.  I can see the code moving closer to an MVC model, but at the same time remaining true to J2ME principals and this is something I am so happy with.  I am getting the benefits of both sides, and as long as it is kept under control, a solid framework will come out of the end of the process and allow rapid development of new projects Ninetyten hs in the pipeline.  All good stuff!

Last week, I had my Blackberry stolen from my car (yes, I know it shouldn’t have been in there etc etc.), and I got on to the phone to T-Mobile to blacklist my handset and de-activate my SIM/BB service.  I was put through to a very nice chap in the US who was on the Blackberry team, who took me through the steps and it was all over in 10 minutes.  They even checked the last numbers dialed to see if anyone had tried to run up my bill.

So that’s the end of the story as far as that goes.

Part II is getting a new phone, of which I decided to go for the Nokia E61i as I wanted a camera, 3G and Symbian as an OS.  Alas, my phone was not insured (again, I know, I’m stupid), so I purchased the E61i from the Nokia store.

That arrived this morning, and I will be putting my newly activated SIM in it tonight.  But, I am jumping ahead of myself.

I rang T-Mobile the day after to see what options I had for phones that supported the BB Connect service, as it is a God send for me to have access to mail, and use the BB Messenger to keep in touch in realtime with people in Ninetyten.  And this is why T-Mobile rocks, twice.

When I rang them, I was put through to the “dial 0 to speak to an operator”, not the BB team. not the business team, but a normal operator.  I asked the question “Does T-Mobile support the BB Connect service on the E61i” ?  There was no oooh, please hold while I talk to my manager, or any click click click from the computer says no.  The answer was yes, straight away.  They didn’t treat me like an idiot, they didn’t try and sell me another BlackBerry, it was just “yup, that’ll work”, along with telling me the process I had to go through.  It was a real person, from the UK, who understood my issue and helped me resolve it very quickly.

Round 2 was about five minutes ago when I rang to ask if they needed me to do anything before I got home to activate the service, and I was told that they would be able to re-activate my BB service, and will walk me through the entire process over the phone (different from my mobile, as the whole receive SMS/turn off/turn on process happens).  The guy knew exactly what I was talking about, he knew what an IMEI number was, and he was really helpful.

How long did I have to wait to get through to him?  about 5 seconds after pressing 0.

Now, yes I may have been put through to a BB team member because it was on my account, but even this is brilliant that their system did this.  I didn’t have to bounce around different departments, and I had all the information I needed.  The only downside is that I always feel like I shouldn’t have rang up and wasted their time because they solve any problems I have so quickly.

I had heard horror stories about T-mobile and their coverage/data support, but I have never had a problem, and would genuinely recommend anyone looking for a decent operator to give them a ring.  They are a joy to work with.

This contrasts Orange, Voda and O2 who I have used before dramatically.  As an example, I had originally ordered my BB from O2 that was supposed to arrive the next day.  10 days later, it still hadn’t arrived, and after calling them every day to complain that it still hadn’t arrived, and that I was waiting for it to arrive at home, I cancelled the order.  T-Mobile on the other hand, I had the phone the next day, and BB activated within a few hours.

T-Mobile, thankyou for restoring my faith in MNO’s!!!