Outsourced Development is Not for Us

Monday 13 June 2011

My development team would nod their heads in agreement to this statement.

No one wants to find that they are being replaced with a cheaper and more flexible option that operates out of a different continent and time zone.

From a commercial perspective the benefits can be huge but for our business it’s a big no no.

I’ve been around in IT for the best part of 25 years and have worked for small and huge corporate beasts that go about their businesses in completely different ways.

Specs written on the back of a fag packet to not being able to lift a finger until the PID has been written and signed off in triplicate before the army moves in.

Swarms of business analysts, project and test planners, change management teams etc… start chewing the fat for 6 months and documenting anything that moves and often things that haven’t moved and won’t move for the duration of the project.

At one point in my career, I ended up working on R&D type projects at AT&T where I got into rapid proto-typing.

Iterative small developments where we would try different ways of achieving a target goal and measure where we were at regular intervals but always with a mind to take a step back when required before moving forward again.

We use rapid proto-typing in my current company and it works for us.

We are constantly looking to innovate and find new ways of doing things. We make extensive use of white boards to explain and share ideas and try to test the limits of different technologies and how they may be used to create something new or do something in a better way.

Spontaneously coming up with and sharing new ideas and turning them into code is a liberating and rewarding experience that would not be possible were we to use more traditional design methods.

Allowing developers to express themselves (in every aspect other than the UI…) and try new things out works really well for us and allows us to be sure that a development is going to work and helps us understand the cost and time implications of turning a proto-type into a polished, integrated product.

Most recently we have been extending our BI software to integrate with unstructured Twitter feeds.

As we use search technology we can handle the deciphering of unstructured chatter fairly easily however the skill is then to make sense of it.

We have been proto-typing over Twitter.

Doing interesting things with the Twitter APIs. Doing interesting things with our VENN technology over Tweets. We have been looking at extracting sentiment out of Tweets and looking at how you can categorise Tweets based on things that are of interest to you.

Most interesting of all is that we have been using semantic vectors over the data to see what additional value can be made out of 140 character Tweet’s.

It is amazing what you can get back if you search for “Giggs”…

A byproduct of this work is what you can get out of semantic vectors over structured data, which is something we thought we’d try following the results over Tweets.

The first time someone tried to explain semantic vectors to me, my ears started to bleed and my brain attempted to shut down so I’m not going to go into any detail here. Needless to say you’ll find it on Wikipedia and Google!

We have created a load of building blocks, which are now in the process of being arranged into a uniform and easy to use application.

It’s been rapid in its conception, rapid in it’s production and will be available for general use at the end of the month.

Not a spec or project plan in sight for this one so here’s a pat on the back for in-house development!

Get in touch to discuss your requirements further