Openistas question UK.gov's £300k crime-mapping website
Portal wobbles at launch
Open data advocates are questioning why the Home Office splurged £300,000 on its shiny new police.uk crime-mapping website, which launched this morning. Sort of.
The portal has been struggling to stay stable over the past few hours since the government hit the live button on the site.
A Home Office spokesman told The Register that its tech team was currently "looking into" what had gone wrong.
Reports on Twitter suggest the site is largely failing to cope with all the interest it is currently receiving from UK residents keen to see how much crime is taking place on their doorsteps.
It's a crime
Meanwhile, some of Blighty's developer community are perplexed by how much the government has spent on the portal, given that the David Cameron-led Coalition has made such a big play for championing open data online.
Rewired State's Emma Mulqueeny, who tweets under the handle hubmum, was exasperated by how much cash the Home Office had spent on police.uk.
"£300k paid by taxpayer funds c/o Home Office for data driven map in the name of open data? Exactly the opp of the whole flipping point," she said.
The Google Maps-powered API could have cost a lot less if the open-dev community had got their hands on the data, she argued.
We asked the Home Office if it could provide us with a breakdown of costs and also put Mulqueeny's concerns to the government department.
A spokesman said it was unlikely the figures could be released "for commercial reasons".
He was unable to immediately respond to the concerns expressed by Mulqueeny and others, however. We'll update this story if the Home Office gets back to us with a further statement. ®
Not a surprise
Having spent some time within NPIAS's IT group, at several levels, I suspect I can explain what happened:
• an original project, very lightweight and set up as a quick experiment turned out to be far more successful than thought (and it really has done - for all the 'oh god is there that much crime' there appear to be ten 'there is much less than I thought' with subsequent benefits for public confidence)
• an inhouse project was encouraged to scale it up (and just in case the openistas complain about costs here - there is a little issue of using the google service for large numbers of invocations - you pay and boy will there be a large numer of invocations)
• the system was sized based on expected averages and a discussion was held with the (non technical) senior management about loading - startup, mean, peaks - who did not understand it
• someone warned them of a similar problem with the UK census online (qinetiq I believe screwed that up) and someone suggested it would be a good idea to either use an admittance system or simply roll it out county by county over several weeks/months to avoid this type of loading failure
• the bean counters in alliance with the 'marketing department' (yes NPIA have one) ignored them
• the results are there for everyone to see - and it will be the architects who warned about this - who will be blamed and another public sector service gets unnecessary shtick on a failure that should never have happened
A few questions
1. 300k? Seriously? I've built similar sites/apps for a fraction of that. Does the Gov. automatically go to the highest bidder or something?
2. As it obviously doesn't work, will the Home Office be getting any of OUR money refunded from the useless company they hired? Aside from the fact that it doesn't work, how could they not anticipate and plan for the massive amount of interest something like this will have on day 1?
3. When will someone be getting fired for this?
I was told that...
Apparently somebody typed in "SW1A 0AA" and the system tried to return a number greater than MAX_INT.
It has the expenses form in the pocket.