Glenda 1
I & D Smiths both -100
The UK government's controversial plans to cut and shut benefit payments into one system was defended today by Work and Pensions Secretary of State Iain Duncan Smith. Just as Duncan Smith took his seat for a relatively easy grilling from the Work and Pensions Select Committee, his department put out a statement about the …
I know it's yet another bad government IT story but 'too small scale', and 'start with the easy stuff'. Isn't that how you are supposed to do Agile? Get the easy stuff working, then tackle the corner cases and expand the scope. Far better than banging down a huge wad of cash up front.
It's just a shame that it hasn't been planned or phased as an Agile development process. It's just turned into an "Oh F... this is all that we have that even sort of works at the moment" rollout. For something this complex, with so much impact for everyone involved, it's not something that can be dropped in place and then upgraded as and when new features are added. It's definitely something that needs to be designed and piloted in parallel with existing working* systems, especially to see that complex cases can be handled.
It actually appears that this is the way it is being delivered - but from the way it's being presented, only because it got so screwed up it could not be launched in any other way.
*Ok, maybe calling the existing system "working" is pushing it a bit, but at least it is there
That is indeed how one is supposed to do Agile. It's one of the reasons that Agile is rarely a fit for serious large-scale development projects. For example- to get a system that works for a few people in Ashton Under Lyme would be relatively easy, but if you haven't designed how it will handle the traffic levels involved when the 2.5 million unemployed are all required to use it from the very start then all you've built is a prototype. There are situations where Agile works great, but when you are going to need massive scale and excellent security you need way more than a couple of iterations worth of architecture ahead of you to be able to build what is required.
See also: http://cartesianproduct.wordpress.com/2013/04/28/universal-credit-death-march-goes-on/
When I test my code I use extreme data, not the easiest scenario, as I want to prove its competence from the start.
Finding a very expensive system cannot cope with the real world years into its development is a total waste of time and money, but why am I surprised, this seems to be the way of government procurement, fire countless resources at a project while getting soothing results from tame testing, until the day it has to prove itself.
Incompetence and public funds are in a bed near you creating abominations!
A most suitable candidate for eugenics if ever there was one.
"I know it's yet another bad government IT story but 'too small scale', and 'start with the easy stuff'. Isn't that how you are supposed to do Agile? Get the easy stuff working, then tackle the corner cases and expand the scope. "
Err, isn't the idea of agile that you proceed quickly between levels that have shallow steps between them? Rather than
Weeks 1-25 Single people. Single parents with 1 child. Disabled with 1 condition.
Week 26 Cohabiting couples with 5 children between them by different spouses (with another on the way), Quadraplegic diabetics with liver cancer etc.
"
It's obvious that this is a textbook government IT project. And as such, baby steps are needed to limit the damage and give the firefighters the 3 years they need to get this thing fit for purpose, which sounds like what is proposed by IDS - Get the basics right FIRST, then start adding complexity after you've fixed the initial (inevitable) problems.
Yes, it's a proposal that is due a lot of flack for not getting it right earlier, proves that the project was poorly planned and executed and isn't going to deliver on time, but compared to the alternative, it's far better than just rolling it out regardless and bringing the entire system to a standstill, specially when it affects the lives of so many of societies vulnerable people.
Don't get me wrong - it's an omnishambles, but in terms of dealing with said shambles and damage control, (and I can hardly believe I'm saying this) I think IDS is probably doing the right thing.
But there are 2 things about this that really p**s me off about this plan.
1) There are already systems in place delivering these benefits. True they might be s**t but you can study them. What they do, how they do it, how staff handle the times it craps out etc.
2) Change in government systems is guaranteed Rates, allowances, structures all can be expected to change. Any architecture that does not embrace change (parameters in data files, table driven procedure calls, decision tables for logic capture and verification etc) is going to be a goldmine for conslutants forever.
Despite the regular bitchslapping HMG insists on giving the taxpayers HMG never seems to learn.
The problem with the existing system is that it doesn't interface very well with HMRC, so when someone starts earning a bit more or less money from part time work, they can't adjust the benefits to suit. The claimant can and presumably will tell DWP about a change in circumstances, but the DWP system works on the assumption that the claimant gets regular weekly income that doesn't change very often, when in reality, they get a few days here and there whenever the recruitment agency phones to say they have work available.
Well the system isn't just the stuff inside the computer. One of the problems is the high number of errors made by claimants in data input on the website, so maybe the questions aren't clear enough. Then you have to look at how staff interface with the system and how it fits into their workflow.
Neither work properly and neither are for the benefit of those that will suffer from their ideas. This is intended as a replacement for all other benefits to be paid once a month. The chance of this going so badly that it leads to suicides and homelessness when payments are screwed up is so high yet good old IDS keeps insisting everything is fine and going as planned. That is only true if the plan is to royally screw up the whole system.