As usual, the UK is first, then worst.
Since the late 70s.
Politicians have been told to help the UK's flailing Government Digital Service gain new momentum by unpicking Whitehall power structures. At an event discussing GDS – Whitehall's experiment in putting tech at the centre of government – in London yesterday, panelists had to admit the service had lost its mojo and was …
I'm familar with someone who was tasked to bring modern methodologies to a Govt department.
They were left in an office without network connectivity for two weeks, then when they did get connectivity, every single move they made was deliberately misconstrued, they were threatened, lied about, anything to get them kicked out. And we're not talking about questioning their timesheets, we're talking full on fucking slander.
The hundreds of thousands of man hours on a major govt project that had literally nothing to show for it, not even an alpha, didn't matter - someone was upsetting the balance of power in that department, and the head honcho there didn't like it.
It's not my story to tell in detail (for a variety of reasons) but I imagine many, many people have similar experiences.
My own direct experience back in the DirectGov days - when £100,000 a year was spent on salaries for sysadmins and web devs to maintain a handful of HTML pages for dozens of departments and even smaller fiefdoms, and Directgov was taking that waste and replacing it with centralised content management - means that I'm quite, quite certain that this is just the tip of the iceberg and as time goes on, we'll hear a lot more of this coming out.
Anon, because of fucking course.
"I'm familar with someone who was tasked to bring modern methodologies to a Govt department."
The TL;DR from this is that whoever's tasked to implement this has to ensure the power structure above them is short and that they have the ability not only to bring in modern methodologies but to root out the psychotic control-freak micromanagers who are usually the biggest problem in any structure.
"Computerisation" only addresses a tiny aspect of the malaise. If there's active resistance to change then you have either the above or a fear of entrenched corruption being brought to light.
Both happen, but the latter is definitely a major factor in some places - bear in mind that being 'high up in IT' in the civil service (even if what you're doing is a load of crap - it's all about your grade) can lead to big salaries, big pensions and easy access to cushy appointments down the line if you grease the right wheels - people are extremely protective of that, and if they're told that the project they've had sixty people working on for eighteen months and has produced nothing, could have been drafted out in two weeks by a half dozen coders (and then have that proven to them), then they see that salary and pension disappearing, and then the fucking knives come out.
There are dozens of fiefdoms like this in central govt, and they're deeply, deeply entrenched.
Same anon as above.
They can come up with these wanktocracy excuses as much as they like but they were a bunch of clueless idiots sucking at at the teat of Web 2.0 fuckwittery, overpaid, below par competency and with a vocabulary so full of bullshit that the digital arseflies grew fat and expired from gorging on it.
It was overstuffed with Ex-BBC cockwombles as I recall and small boys in sharp suits just out of college who looks just a bit too pleased with themselves. More obsessed with the use of the latest 'I' Thang than with national security. I did get to see a GDS staffer shit their trousers when an MOD security guy got proper angry at the their spouted shite.
I was in a review meeting 18 months ago overseeing a [technical] performance review of a legacy transactional platform. GDS had invited themselves because they could and for some reason had sign off on the project.
The presenter got to the section on NFRs and one of the hipsters put up their hand after a few minutes and asked a question that summed up GDS for me in 4 words: "What is an NFR?"
Why would you assume they should know your acronyms?
I've worked in IT since the 1980s from DB to DevOps, including both public sector (NHS, Police) and private sector (telecoms, internet), and I know the usual techie TLAs plus such stuff as CLI, BNA, KPI, RFS and so forth, but I've no idea what NFR stands for.
Couldn't agree more. I remember visiting a few years ago near Holborn to see a bunch coming in from lunch in espadrils and shorts. The idea of them doing actual work or driving a decision was anathema to most of them.
Sounds like a guaranteed way to fail.
50 people who have near enough no clue about the business, how it works and doesn't work, how and how not to influence people, all the social and organisational problems that are the main reason IT projects fail...
They might be better off hiring 50 people over 50 who know what an API is. There's not a shortage of them.
But better yet, hire a mix of people of different ages with different talents and specialities, but make damn sure there's a good understanding of the real reasons projects fail. Its rarely the technology per se, technology failures are just a symptom of a project that's already gone badly wrong.
To my mind the biggest problem with huge government IT projects is that they are huge IT projects. Everything has to be huge. There's no sense that you can transform the service just as well with 50 small projects done right (in a framework so they can intersect in the future) as with one large project.
The problem is that the source systems that need transforming are so big and complex. Trying to build a new system whilst at the same time parallel run the old system is very hard and expensive. All of the complexities of synchronisation, interface routing, master/slave data stores is a pain.
This all leads to "big bang" deployments. We have tried arguing for MVP deployments of the new system with an ongoing build out of function only for the MVP scope to be insisted to be what the old system does in order to avoid business change.
And don't start me on how government procurement neuters any chance of these big systems being successful. Most big programmes I have been on have been hobbled from the start by a really shit procurement process that drove a race to the bottom on some very shoddy requirements.
"All of the complexities of synchronisation, interface routing, master/slave data stores is a pain....We have tried arguing for MVP deployments of the new system with an ongoing build out of function only for the MVP scope to be insisted to be what the old system does in order to avoid business change."
And how would your less than big bang MVP remove those complexities? Surely it's the cause of them.
Quite often big bang is the *only* way to do these deployments because the the complexities I referred to.
The MVP comment was describing an approach to try reduce the scope of an initial release to reduce the risk and the time to implement. But as these systems are often critical to the operations of a department and/or implementing an archaic, pre-computing, business process then the MVP is typically not an option.
"Hire 50 people under 30 who know what an API is."
For a start that is age discrimination but lets put that aside. When hiring for the Treasury the last thing we need is a bunch of millennial fuckwits who think an API can solve the country's problems. What the Treasury needs is to hire people who understand economics and taxation.
Personally, I would have thought that hiring a few clue full 50 or 60 year olds to give the under 30s helpful advice and erm.. "guidance" would also have helped. Apart from anything else, the CS's power structures are aged based and having some wet behind the ears junior oiks (as the upper reaches of the CS would see it) telling CS mandarins what to do, or how things are about to change rather radically, simply would never have worked.
Trying to do it again with "New GDS" will also fail unless it has a structure and credible people in charge that speaks the senior CS's language and is prepared to engage with them on that basis.
That means hiring technical people of substance, experience and age that are actually worth the sorts of salaries that were being paid out.
Apart from anything else, the CS's power structures are aged based and having some wet behind the ears junior oiks (as the upper reaches of the CS would see it) telling CS mandarins what to do, or how things are about to change rather radically, simply would never have worked.
Indeed - you'd have more luck trying to turn a supertanker with a row boat.
As far as I can tell, GDS has largely been characterised by doing some nice data vis and shiny web2.0 odds and sods.
Nothing whatsoever to do with getting to grips with the underlying business practices. That might be unfair but is the impression I get from MLF.
Change management isn't done by parachuting in and saying "look at us, aren't we magic, we're the future.". It's a gradual process, finding ways to incrementally improve a department's workflow, with the buy-in of the staff who are going to actually use your system after you've fucked off.
Usually the biggest change you can make to an old-school department is to give them a NAS with file and version management. It's not sexy, but it centralises and consolidates documentation and eliminates the number of "Briefing v12_FINAL_CHECKED_ACTUALLY_FINAL_2.doc" type files floating around on USB sticks.
If it hadn't been for the quasi-religious faith in everything in open-source they may have achieved something.
I kicked the can down the road of flipping 1,000s of MS Office seats to use ODF 'standard' files knowing it would all blow over once Mad Frankie Maude went. Who wanted their staff up on arms that Word mark-up had gone or Excel couldn't have macros?
Still, the contractors on the gravy train made out like gang-busters!
"Hire 50 people under 30 who know what an API is"
That's what GDS was full of. Unfortunately they didn't hire a single person of any age who knew what 'Integration' was.
They ignored all the 'difficult to do' and 'requires knowledge' backend stuff and just skinned the frontend using the latest fad on hackernews.
Only going to be taken seriously by the clueless.
I have a lot of sympathy for the point about crap and fusty ministerial leadership. Sure, they aren't interested in success unless they are rewarded for it, and their currency is political capital.
But, and there is one, the biggest resister of change and diversity, in my anonymous experience, is getting anything past the equally noxious technical committees who resist change, and encourage mono-culture, just as solidly as their masters and in similar fashion, albeit more faith-based.
Some of us are actually fixing this, and not with GDS' brain-dead la-la double-speak, but with solid incremental engineering. Wish us luck, we'll need it.
So Whitehall isn't optimally organised. Who knew?
Nevertheless, Whitehall is where GDS has to operate. That's the job. To say that Whitehall needs to be reorganised, we need to start again, we need a new world, is to say that GDS can't do its job. Who knew?
Never forget that some people in Whitehall can do their job. The Cabinet Office of old produced the Government Gateway, for example.
The alternative model for Whitehall advocated by Tom Loosemore, Stephen Foreshew-Cain and Mike Bracken is to "smash the silos", create canonical registers which constitute the single source of truth and then let algorithms design services by processing the (big) data. Parliamentary democracy would die and the power would shift to GDS, in charge of the registers and the algorithms and with the final say on what users need. No thank you.
The question arose at the Institute for Government event on 4 July what progress has been made since the advent of GDS. Answer, government IT failures are a thing of the past, please watch between 6'20" and 6'40". Who knew?
If you cannot do a task yourself, at least in principle, then what chance do you think I have of automating it?
I can't imagine a process in they have that anyone can actually fulfil. There is rarely an accountable individual for a given process or data, and even if you think you have found them there is a clone in another area running the same process for some other purpose.
Do you write software that builds in all the inefficiency, overlap, double standards and confusion of government operation? Or do you attempt to write relatively tightly scoped processes?
I expect the easier tasks are done now, and the rest need to change the way they do business. This is a challenge, not only because of natural inertia, but a vast tangled web of outsourcing contracts that will also have to change. I wish them good luck, and give thanks that I am not involved...
"This is a challenge, not only because of natural inertia, but a vast tangled web of outsourcing contracts that will also have to change. "
Those come with their own baggage. One example in my own organisation being procurement:
We had a brilliant procurement department. Fantastic attention to detail, supervision of contracts and flexibility to adjust things as times changed.
Unfortunately the archittect of all that retired and whilst he wrote up _everything_, the first thing his replacements did was do a horizontal sweep of 40 years of knowledge and procedures off the desk and into the bin, then claim that he'd left nothing behind (those of us who knew he'd spent the previous 5 years ensuring continuity and were still in contact with his retired life knew otherwise, but hey ho)
Now what we have is a bunch of box-tickers who are focussed on "procedures" and "goals", resulting in suppliers running rings around them and prices being significantly higher than they used to be - in many cases higher than retail figures where they used to be 30% below that - and supply terms being significantly inferior. Of course it's nearly impossible to have them removed because _everything_ is aimed at CYA first and getting the best deals second.
Unsurprisingly, several of these staff have gone on to work for the suppliers....
"The analogue mechanism of Whitehall is not fit for a digital world," said Bracken. "The culture of propriety and in departments is quite phenomenal."
Do they just string words together and hope there is a message therein? Is Whitehall really running Babbage Engines? Is impropriety something that should be strongly supported?
'"Politicians need to tackle this – it needs political nous. We're going to have to get some political grip," Bracken said – a view that was echoed in the familiar lament from Watmore that Francis Maude was the first to give digital "ministerial drive" without it necessarily being in his brief.'
The first two sentences suggest very little nous behind them.
GDS is handcuffed by the CS, they see it as something foisted on them to make things harder.
The CS is the last UK industry that has an 80s style Management structure and a Union that has any power. The intentions of GDS were noble, but if we want change thats going to continue through more than one parliment, someone has to takle the working practices of the CS.
Biting the hand that feeds IT © 1998–2019