Thursday, 18 March 2010

Microsoft Mix 10 – designing Modern Web Apps

My current project is based around an interactive application delivered over the Internet to both full screen browsers and mobile devices with small form factors. Of key importance to the business plan is that the application be easy to use, and be capable of use by citizens who are not necessarily computer literate. (It’s also going to be delivered via Microsoft’s Azure cloud computing platform – but that’s another story).

As the product is being built with Microsoft tools, I’ve been up early this week watching the videos from the sessions given at Microsoft’s Mix 10 event in Las Vegas (you need to have downloaded or streamed the videos before the east coast of the US wakes up – from about midday onwards the response is very, very slow).

There is a lot about technologies not strictly relevant to my current project, but out of the Azure and Web Apps presentations I’ve seen to date, the best has been from Luke Wroblewski (not an MS employee) on the topic of Modern Web Form Design. In summary, Luke describes how to use modern web technologies/tools to deliver better end user experiences, and illustrates his talk with results from research into the end user acceptance, and use of, tools/techniques such as in-line validation, AJAX accordions and other such tools aimed at providing a better end user experience.

If, like me, you have an interest in this area (and would like to learn more about the methods to adopt in building web apps for small form factor mobile devices), then I thoroughly recommend the video (although be advised that it is over an hour long, just).

P.S. For those who want to know more about the Azure cloud computing platform at a fairly high level, then I recommend a Lap around the Windows Azure platform (although, yet again, this is about an hour long). This demonstrates the ease with which apps can be deployed to the cloud – although I can’t believe it’s as easy as the demo……

Monday, 15 March 2010

How to split large Government IT projects

I’ve been intrigued by the debate on large Government projects and the use of the larger service suppliers that has been prompted by the Conservative Technology manifesto. Some rush to the defence of the larger suppliers, whilst others, typically coming from the SME sector like me, point to the way the current procurement process fails to include SME’s adequately.

My experience of working as potential subcontractors to the big service suppliers is that even though you may have a market-leading software solution, they will try to find a way to prove that the end customer will be better off with a customised solution built, typically from scratch, with lots of chargeable days from the main contractor, rather than making use of an SME solution. And how many software package selections do main contractors make on the basis of the amount of services required from the main contractor to implement the solution (rather than possibly a better/cheaper solution that doesn’t involve oodles of services from the main contractor)?

These large services companies are in these large projects to generate services revenue for themselves, maximise their margins, and to make money for themselves - not their subcontractors – who they will use only when they really need to – and typically then only with loads of chargeable time from the main contractor to oversee the subcontract procurement and subsequent management of the implementation project.

But why would we expect otherwise – it costs a great deal to bid for Government work, and once it’s won, who would expect the supplier to do anything else. No – the problem lies in the way Government structures, procures and manages these projects, not the way the big services companies work.

Even with the Conservatives’ proposed limit of £100M on IT projects, the projects are likely to fall outside the types of project that SMEs can bid for directly. Central government needs to change the way that it structures larger deals, and uses the larger services suppliers to oversee them. Yes, use a main contractor in a management role or responsible for integration, but making it clear how far that role goes, and in particular that whoever manages the procurement and oversees the project cannot fulfil any of the other roles. Why not set a minimum percentage of the project value that must be spent with SME subcontractors?

More importantly, split the application software development out into separate projects from the implementation and roll out (and have separate infrastructure supply and support projects). For major new developments, fund two or three SMEs to develop software in competition, keeping the best solution but being prepared to throw away one or more developed solutions before the cost of implementation and roll out – even though the developments have been paid for. Get experienced software developers involved sooner, and in touch with the end users to develop software that really meets their needs.

Let’s get a contractual framework where the main contractors are focused less on where their own services revenue will come from, and more on how to provide the best solution for the customer.

P.S. The Conservative plan for a small in-house ‘skunkworks’ team, to develop low cost applications and advise on the procurement of larger projects seems like step in the right direction. But will Government be able to recruit the appropriate resources – with all due respect to the IT civil servants I’ve met, in most cases, they are not the types of staff that will be the best for this new role. As noted above, why not make use of those staff in SMEs, calling on a much wider pool of experience, and in many cases with experience relevant to the specific project in mind…..