Today, after what has been a bit of a frantic build up, we've gone live with a project that Osmosoft have been involved in for some 18+ months. Over the course of that time the scope of the project has shifted, twisted, turned and materialised into a number of things. Since the beginning of the project we've lost some good people from the team -- not specifically as a consequence of the stresses of the project I hasten to add. But with those of us remaining we managed to get through it!
There was no fanfare or applause when the go live button was pushed, and there certainly wasn't any champagne corks popping or pyrotechnics. But never the less it feels like a pretty good achievement and one that I personally have learnt a lot from. The work was for an internal BT project, and at the moment the stakeholders are a bit cagey talking about the specifics in public -- that'll change and I plan to elaborate a bit more about what was built. But in essence we started with building a prototype that helped business folk explore different ways of sharing information amongst themselves using positioning and sizing of snippets of information as a mechanism for conveying importance and relevance. An early prototype of this can be seen over here -- use 0 to zoom all, left/right cursor keys to navigate.
That early prototype led to a number of discussions around how further abstract models of sharing information could be used. There were some interesting conversations and ultimately what got proposed was a toned down, slightly more enterprisey solution. The key win however was getting senior buy in to move from a world where information was being shared in Word, Powerpoint and Excel documents via email to one where information lived and evolved on the web within a wiki.
What was built was an application whereby content was authored within a wiki, on top of that there's a layer that pulls content together and presents it within a structured manner with a separate interface. This allows for viewers of the content to be able to follow a step-by-step walk through in a logical manner, or take a more explorative path using all the goodness that wiki links provide.
Putting technology to one side the thing I've found most interesting about this piece of work is the reaction from management teams when they realised that individuals who are using the content -- most of it being processes, procedures, and best practise -- are also within the same pool of people creating and maintaining the content. For most that's been a massive paradigm shift from a culture of auditing and ownership coming from a single 'trusted' source.
Another interesting aspect was that the solution has been built on an open source stack. Now, this certainly isn't the first open source project deployed within BT but we still came up against some of the same misconceptions around what open source meant. Questions such as who do we go to (blame) should something go wrong, who 'owns' what has been built, can it scale, does it have Enterprise level robustness etc etc. Thankfully misconceptions around open source weren't shared by everyone and with some careful educating and demonstrating through building rather than pushing around spec sheets we were able to get the application deployed within a production environment...eventually!
(On the point of open source I'm not sure how relevant what we've built will be outside of BT, but where it makes sense we'll be putting the code up onto Github).
Whilst all the politics was playing out around strategic vs non-strategic and whether or not an 'open source' application was fit for purpose a related piece of work came our way.
The same group of people who were using the wiki to view and maintain process and procedural information were also backed up by a telephone support team who they were able to call should they come across a particular problem that hadn't been documented. These kinds of issues may be very specific and time bound, they might be relating to a particular customer or geographic location, or some may be quite generic and ought to be added to the wiki.
The details of the calls weren't being captured anywhere and the business were quite rightly concerned about valuable information exchange being lost.
The thing that we ended up building was pretty nice looking web form that as you entered information about the call it displayed details of previous call logs and related information from the wiki. Details of a call could be linked to related information to help promote reuse of content and highlight common themes. There is also some social voting where inspired by stackoverflow.com where answers to questions can be voted up or down. We also introduced some features that used the browser geolocation to pin point an individual on a map and layer on top additional information.
Our remit for this piece of work was to build a first iteration of the application that could be deployed for use in a live environment but also seeded ideas for its evolution. It feels as though we've done a good job on that and one of our small (but heroic) victories was around getting agreement to roll out a webkit browser to sit alongside the standard IE7 internal build. This has allowed us to demonstrate how recent developments in the HTML spec can quickly and (relatively) easily help tackle some key problems. Hopefully this will pave the way for some more interesting work over the coming months.
The applications that we've built have been hosted on a dev server for a while now and been in trial with a number of teams across the business. Today marked the day that the application moved onto it's own production environment with the trial badge removed and support taken over by a full blown 24/7 team.
It's been a fascinating project to work on providing a whole load of stimulating conversations around technology choices, business process, enterprise culture and change management. In a way it's sad handing it onto another team, but on the positive side it means we're now ready to move on and pick up the next piece of work!