34 comments on “For fans of government IT fuck ups

  1. From the article “…the fundamental problem, which is that the people who were describing this amazing system seemed to think that if something could be easily described in English, it must not be that hard to do”

    In 30 years of programming you’d not believe how many times I hit that problem. It’s been the curse of many an IT project where the deadlines aren’t set by the techies but by management, and management just aren’t interested in database incompatibilities or data security or much else that’s going to impact development times.

  2. “affordable coverage (defined as 9.5 percent of your income or less)”: that’s 9.5%, not 10%. If you used 10% people might be able to do the arithmetic at a glance. We can’t have that, can we?

  3. In 30 years of programming you’d not believe how many times I hit that problem.

    And in engineering, and technical provision, and construction- and no doubt in numerous other areas too. But I would imagine there is a particularly cute problem at the moment in IT because of a general idea that computers can “do anything”.

  4. It’s weird isn’t it? Right across the private sector there are companies introducing IT systems. And by & large, they do so successfully. They have to. Their business depends on it. Yet it’s hard to think of a government sponsored IT system that works. There must be something fundamentally different happening.
    Any ideas?

  5. Just thought of an example. While ago had to complete a lengthy on-line questionnaire for a social services matter. Almost every page requires info to be entered, takes considerable finding. And you can’t get to the next page without completing the current one, so you can’t preview the questions & prepare the data. Take too long over it & the page times out & you have to go right back to the beginning again & there seems to be no cookies stored with the previous answers. So it’s enter in full. There’s not even a “back” button. Completing it full takes hours & must be tying up server capacity all the while.
    Who’d design such a thing? There’s no fundamental difference to doing an on-line shop at Tesco. So why’s the Tesco experience quick & intuitive & this a nightmare?

  6. To be fair, lots of projects involving IT* fail in the private sector too, but we rarely hear about that because they aren’t particularly newsworthy (a factor here is scale or the sum of money involved, another that they don’t involve people we’re interested in). Likewise there are government projects involving IT that work, but that doesn’t make a particularly interesting story.

    As an observer, something frustrating is that the same problems occur time and again. Some time ago the Office of Government Commerce observed that the common causes of project failure are (and the private sector causes are similar):

    A. Lack of clear link between the project and the organisation’s key strategic priorities, including agreed measures of success.
    B. Lack of clear senior management and Ministerial ownership and leadership.
    C. Lack of effective engagement with stakeholders.
    D. Lack of skills and proven approach to project management and risk management.
    E. Lack of understanding of and contact with the supply industry at senior levels in the organisation.
    F. Evaluation of proposals driven by initial price rather than long term value for money (especially securing delivery of business benefits).
    G.Too little attention to breaking development and implementation into manageable steps.
    Inadequate resources and skills to deliver the total portfolio.

    You can see those in the Bloomberg article Tim links to.

    (*said in this pedantic way because IT projects are often really change management projects that involve IT – unfortunately that is not how people think about them.)

  7. “To be fair, lots of projects involving IT* fail in the private sector too, but we rarely hear about that because they aren’t particularly newsworthy”
    Big projects by big players? Banks have had f/u’s a couple times, hit the front page headlines. And the press love the quirky ones. The airline that was selling long hall flights for $0 last week. Something like the Amazon interface must be more complex than most government projects. Hell, they know more about my shopping habits than I do. Amazon glitch would hit the “newsflash” line on the bottom of the rolling news channel screen.

  8. Of course a failure at Amazon would make the news. There is a story about Nationwide online banking today. But the failing government projects we hear about tend to involve billions of pounds and millions of potential users. The local authority failures won’t tend to reach the nationals. The successes won’t tend to reach the non-specialist news. We’re more interested in failures than successes.

  9. “But the failing government projects we hear about tend to involve billions of pounds and millions of potential users.”
    And Amazon’s didn’t?

    government IT fuck ups

    This was fun. Was over Tim’s side of the border when my companion has his passport stolen. Recorded advice on the Passport Office’s lost passport number was to consult their website. (Don’tcha just love how they take it for granted any Brit coming to grief in deepest Foreign will have internet access?) What we found there had us chasing along to the consulate in Portomao, the next day.
    “Can you help, my passport’s been stolen.”
    “All the information you need is on our website. There’s a computer over there in the corner, if you wish to use it.”
    “OK. Can you show me how?”
    Bint reluctantly comes from behind counter, calls up passport Office front page, clicks “lost passports”

    404 page not available

    Been like that for 3 days we were told by a woman who’d lost hers.

  10. @UKL

    ‘To be fair, lots of projects involving IT* fail in the private sector too, but we rarely hear about that because they aren’t particularly newsworthy (a factor here is scale or the sum of money involved, another that they don’t involve people we’re interested in).’

    We rarely hear about them because it’s not our money (unless you’re a shareholder).

    ‘As an observer, something frustrating is that the same problems occur time and again. Some time ago the Office of Government Commerce observed that the common causes of project failure are (and the private sector causes are similar):’

    The one they missed out, and the reason they ‘occur time and again’, is simple: it’s not their money they are pissing awsay (and in fact, there are jobs for the boys in fixing the screw ups, or overseeing the fixing thereof).

  11. Simple: no one employed by big state ever gets fired for incompetence or even manslaughter and lynching by the irate citizen is illegal. Protectionism by state trades union for underlings, promotion for managers, pensions for all. It’s a parallel universe where different consequences apply, all run on wishful thinking and boxticking.

  12. @blokeinspain
    Actually, you’d be surprised just how many private sector systems go live with major problems. We just don’t normally hear about them because they have to be catastrophic before the media pay attention.

    Part of the problem is that software development is still using basic tools and hand crafting everything. Things have improved a lot in the last decade or so, but we’re still a long way from the ‘engineering’ part of software engineering.

    You’ll know things have changed when it’s possible for software can be assembled from a catalog of standard parts and has an MTBF.

  13. “…the fundamental problem, which is that the people who were describing this amazing system seemed to think that if something could be easily described in English, it must not be that hard to do”

    It’s not just the sciences, engineering and technology.

    I work in publishing and this describes most screw ups. The key phrase is, “can’t you just…?”

    It’s the “just” that tells you you’re dealing with someone who has no conception how things work. Usually a liberal arts graduate who, as the original quote says, believes that because it is possible to formulate the words, reality must conform.

  14. Two points – private sector failures rarely make news headlines because the company that makes the mistake has to make good any losses suffered – or at least make a gesture that keeps guys from complaining too loudly.
    ljh and ukliberty make the second point – those who waste a fortune in the private sector get sacked and replaced by someone who either is more capable or makes vast efforts not to make the same mess but in the public sector no-one is held responsible, unless they try to seduce a programmer, so public sector IT disasters are a recurring phenomenon.

  15. Stuck-record is correct.

    Normally it is a BYT – Bright Young Tosser – that throws the spanner in the works, irrespective of the degree course, but a shithead who has been told he is bright and on the fast track management programme. He will come up with “can’t you just make it work like my Iphone shit to show everybody how smart and clever and the ability to think out of the box and forward thinking he is. Then he fucks off and the programmers and department is left with his shit, get on track, get going and then the next BYT comes along and all the shit starts again.

  16. I was in a group that was invited to tender for one of the large govt IT projects. We decided not to. It was almost as though it had been structured to fail.

    The best thing for them to do would be to work in terms of developing small components that do simple, useful things well and can talk to each other.

  17. Steve Crook,

    Part of the problem is that software development is still using basic tools and hand crafting everything.

    Software development is about making unique things (if it’s not, you just reuse a library). Therefore it has to be hand crafted.

  18. @Steve Crook

    You’ll know things have changed when it’s possible for software can be assembled from a catalog of standard parts and has an MTBF

    As a veteran software engineer (now retired) I have my doubts that off-the-shelf software components will ever constitute more than a small part of any innovative system. For bog standard commercial applications I would have thought we were already there.

    Having worked for software contractors most of my career I have yet to work on one where we were not committed contractually to a MTBF usually under pain of financial penalty.

  19. Yes, unfortunately large scale software development involves re-inventing the wheel each time. Like the Galgafrinchiams most of the arguments are about its colour.
    It is as if the entire project team has been wiped clean before sitting down in front of the whiteboard and forgotten everything that they once knew about running a project.

    In certain mainframe environments, development can be done automatically. One writes a “factory” program and it can produce the software modules based on inputted logic. These are quite complex and can take a few months to write, but the results will be bug-free and ready in a blink of any eye.

    Project Managers don’t like them because there’s nothing “to show” while the factory is being built and because usually they have no experience of such techniques, they consider them risky.

  20. I assume a “long hall” flight is anything involving terminal 1 at LHR, in which case $0 seems a reasonable price, seeing as you have to walk to your destination.

  21. Almost all failure in IT projects at the actual development level comes from a lack of clear specifications and the ability to enforce adherence to them. Feature creep has killed huge numbers of systems. In government IT provisioning, there is the same problem that afflicts a lot of other reform efforts: that the bloke ultimately in charge will be shuffled off somewhere so developing a coherent long-term strategy is next to impossible. When it’s the latest wizard wheeze to shake up nursery schools or whathaveyou, it’s one thing, but steering a large project to completion is another. And of course there’s very often the Yes Minister syllogism: something must be done/this is something/therefore we must do it. Why is this project being created seems to be a question that is only ever asked half-heartedly.

  22. By the same token, failure can also be down to specifications that are just wrong. Usually because they have not been tested on a computer first. Often the result of poor understanding by the designer of the system and the requirements ( or even what the tools available are ) and sometimes by customer interference.
    They look good on paper and the whiteboard ( or Powerpoint ) but actually program the thing in and it’ll crash and burn.
    Overengineering is usually the reason, because “keeping it simple” is too… simples.

  23. when someone uses the word “just” as in can’t you just or don’t we just….then the project is on track for failure…another instance of Hodge’s Law

  24. “..because they have not been tested on a computer first.”
    Subject of an e-mail exchange with the guy designing a website for a pal. When I was given the opportunity to test fly it, I was stunned by the amount of data it ate. Using a 1Gb/month shit mobile connection makes you notice things like that.
    “But your target user is the third world & flaky net connections. They’ll never load the pages”
    “Works alright in the office.”

  25. Good hackers are about 100 times more productive than average developers. Good hackers want excellent computers and lots of stock options. Government work offers crap computers and no stock options. QED. The spec thing is a red herring, because a good hacker will run up a working prototype in a few days and that enables the users to get their ideas clear.

  26. @ Thomas Gibbon
    Good programmers or hackers don’t need excellent computers.
    When I was 17 I had the great good fortune to work for 8 months (with two other guys who also had relatives* working for the company and places to read maths) in a section that had a computer with 30k bytes of hard drive and a few bytes of RAM. Since it was to small for an Algol 60 compiler the section head had devised a “Conventional Language Code” – a stripped-down Algol – and single-handed written a translator that was slim enough to run on this! Accounts department had a much “more powerful” computer with a Fortran compiler but our “little” (size of this room) computer and CLC could – did – do more
    *In case anyone thinks nepotism, this was as more about potential future recruits with some cause for loyalty: two fathers, one chemical engineer, one process worker, one sister, a secretary and we were paid (£6 per week) which was decent pocket money for lads living at home with an interesting job but less than we could have got from labouring jobs.

  27. It is as if the entire project team has been wiped clean before sitting down in front of the whiteboard and forgotten everything that they once knew about running a project.

    Our major oil and gas projects are like that. I’m currently working on the company’s 8th or 9th giant FPSO, and we still haven’t standardised any of the design. Every one is unique, and from what I can tell, that means it’s shit in its own special way.

  28. “a good hacker will run up a working prototype in a few days and that enables the users to get their ideas clear.”

    Aaargh! No! The number of instances of “the prototype becomes the product” is too large to calculate. Storyboarding is one thing, but knocking up a proof-of-concept has a distressing tendency to mean technological lock-in, failure to scale and poor maintainability.

  29. “Part of the problem is that software development is still using basic tools and hand crafting everything.”

    Uh, what?

    Much of software development today is picking the right third-party libraries and hooking them together with custom code to make them do what you want to do.

    Of course that’s in private industry, where they want something that does the job for the lowest possible cost. If the government was willing to hand us billions, obviously we’d write everything from scratch to justify the cost.

  30. You can write anything in software. That, alas, is the problem. You can write shit far, far more easily than you can write good.

    Couple this with lack of communication, lack of understanding of complexity, the whole show being run by arts graduates and rampant arse covering at all levels and you end up with government IT fuck ups.

  31. If you get your info via fax you need inputters, Chris. Inputters who will then vote for you. It’s genius, actually.

  32. @John77
    In truth I don’t know whether the gifted folks I work with really need lots of kit and bandwidth, or whether it’s a status thing. But that (and the stock options) is what they want, so that’s what they get. And it works.
    @David G
    Of course the prototype should never become the product! But, like I said, customers benefit hugely from seeing their inchoate thoughts crystallized & the faster the iteration, the faster they converge towards a solid spec that you can properly engineer. If you do that you’ll make a lot of money.

Leave a Reply

Name and email are required. Your email address will not be published.