Well done
I like the investigative journolism slant.
If you want to understand what is wrong with public policy when it comes to IT in the UK, look no further than the recent tragic case of the letter sent by a school to the parents of dead schoolgirl Megan Gillan, demanding that she improve her attendance. It was one of those bleak and bitter accidents that inevitably occur …
It may not be correct as its in the Daily mail report I read, but it said that the letter was SIGNED by the deputy head. If that's true, what is he doing signing stuff robotically like this, there surely must be many reasons why attendance has dipped and to blindly accept what a computer says without checking on each pupil is idiotic. Surely there aren't that many letters going out that he cant check each one ? ( or if there are the school has another massive problem that wont be fixed by a computer ).
"So why is a County Council putting out a statement that is at best misleading, and at worst calculated to divert attention from an individual failure or even a more systemic failure by the Education Authority to ensure that users of the SIMS system are properly trained?"
Gee, I dunno, maybe everyone's covering their asses instead of stepping forward and allowing as to how nobody has been properly trained on the system because of bureaucratic cockups, including the top level management who approved the system?
Nah. That's too easy ... It must be a conspiracy.
What the media want is two things:
someone (or thing) to apportion blame to, and
a short, not too technical, explanation - preferably in 4 words or less.
Now so far as the media is concerned, large corporations are BAD (the bigger, the badder). Schools are all staffed by either saints or child-molesters - there's nothing in-between. Any mention of science, technology, maths or politics loses readership/audience. Long or technical sounding words (i.e. anything more than a slow eleven-year-old, or your great granny would use) alienate readers
So given these parameters, it's not hard to see how a long, detailed, technically correct and accurate explanation - complete with diagrams, listings, screen captures and 3 pages of text could be sub-edited down to "computer error". Even better if a big, bad corporation (preferably foreign, even better if funded or working for the E.U.) can be associated with the issue. Slap in a few pictures of grieving kids/relatives and some pointless reporting from outside the establishment - even though it adds nothing to the sum of human knowledge and what have you got?
The perfect anti-technology story. Feed enough of these through and you end up with a anti-technology readership, who have been brainwashed into the idea that science is hard, nasty, socially unworthy and that change. or progress, is to be avoided. The drip-drip effect of all of these is to create a climate where people don't "do" science - and are proud of the fact. Where children learn "cool" subjects: such as wikipedia, football and meeja studies and all our technology is outsourced to cheap-labour countries where having a GCSE in physics is a major boost to your job prospects, rather than something you keep quiet about, if you don't want disapproving looks from the other shelf-stackers in ASDA.
A more disturbing possible element in that tale, is that Capita are able to bill for phantom works and/or identities which are not accurately entered in their databases and which would also give connecting databases erroneous, spurious inflated fictional data, creating a False Picture/Virtual Reality/Inflated Accounts.
While studying for an IT course at my local uni I had a part time job in a nearby supermarket. The number of people I would come across who would proudly proclaim they knew nothing about computers was staggering.
One woman insisted she had a copy of Windows 93 installed on her pc. Must have been a windows 95 beta then?
Another had seen a pc but was eager to tell all and sundry that she didn't know how to use one.
Would they be happy to tell everyone if they were crap at writing, reading, driving a car? What is this pc myth as if the pc is a sentient being.
If my pc was sentient it'd be cowering in a corner after all the swearing i've given it for f*cking up now and again.
Alien - computers aren't!
I believe that it was this system that sent me an automated text informing me that young Stacey was not at school.
Which was nice. But I don't have a child called Stacey. And the school in question was in a different county. The nice (but confused) lady at the other end of the phone didn't seem to understand that someone had put a wrong number in, and genuinely seemed to believe that it was a computer error that had generated it - despite me confirming with her that it was my number on the entry, in error.
*sigh*
This reminds me of the stories we used to see, back in the day, of some elderly person's pet budgie, or whatever, being sent a poll tax reminder. These sorts of stories were always portrayed in whichever section of the press reported them as being great examples of local authorities and their useless computer systems. In these type of cases no one ever mentions that the doddering old fool must have stuck their budgie's name on a form at some time.
I'd suggest there is a third group of interested parties who think the debate matters.
These are members of the general public, like me, who want the benefits of appropriate IT & automation to be exploited fully but who deplore the poor specification, implementation and testing of nonsense applications or applications that add cost not value. It doesn't matter in the end whether errors derived from these applications are "human" or "System". If they weren't used there would be no "error" for the ignorant journalists of the "Red Top" rags to waste peoples time with.
Too often, in my opinion, IT and automation is thrown in as a substitute for thinking. Also, too often new IT & Automation is not used because those who should seek guidance on its application fear the consequences of "system failures" landing at their door and press on with current crap paper and other systems. The results are massively sub-optimal systems, extremely large opportunity costs and an increasingly pervasive "it'll never work" attitude!
I don't run Capita's Unit-E but I do work for a college running a competitors product.
I would say the problem is this.
The school wrote their own report (they all have reporting services bolted on) which directly referenced the data held in tables with no thought of removing anyone dead or excluded from the list.
If the report was user generated then thats a user error. If the report was based on a datasource that was supplied by Capita with the system. Then that's Capita's problem.
I suspect the first. The trouble is there are a lot of institutions out there that do not have real experiance with caring for their databases or data in general. Often they completely lack an experianced database administrator/developer. Anyone can write a few reports using a drag and drop report writer can't they ...
Simples
"So the system, for some peculiar reason, holds multiple representations of the same data? If true, that would be an accident waiting to happen."
In my day, the data for a small group of people was centralized in a PERSON, who took register, noted the reasoning and adjusted accordingly.
So for example it (or rather 'she') would *not* send out a letter if the child had Leukemia and was often away... you know, H-U-M-A-N interaction, the flexibility of people.
It didn't need a special flag for a special exception, because the teacher could handle all special exceptions using a very flexible thing known as a brain.
I notice Brown is always "name and shame", his response to everything is to "name and shame local councils who....", "name and shame failing schools who..."
And that this is often built into the automated systems and even in the attitudes of the people who run them. Even the government adverts constantly threaten and harrass (e.g. you can now pay your car tax by phone, or your car will be crushed).
Automated systems that send out threatening letters are the norm for Britain and when people complain, the downtrodden people whose lives are sh*t tell them to shut up, enjoying the fact that someone else life is as sh*t as theirs is.
What kind of a life is that? Joyless people making each other unhappy to drag others down to their own level of unhappiness?
IMHO, They should stop sending out automated threatening letters, when someone is absent, they should ask the person's teacher to investigate or telephone, or explain about the death or the Leukaemia or whatever. People should treat each other decently.
Same goes for the 'squeal on your neighbour for your own suspicions' Jacqui Smith stasi idea.
Same goes for the attempt to replace people with Biometric measuring computers and 'behavioural analysis software'.
Same for every government department that sends threatening letters as the first default for every little thing their computer tells them is wrong.... do you ever imagine the fault is in your computer? Or the data it has? So why is the default to blame others?
If there is only one dataset and the deceased flag is set to yes or no, I can only suppose that either the query used on the database was poorly constructed and didn't actually read the flag, or the flag was read then ignored. Either way, it's a human error by the developer and also by the testers, given that the scenario which occurred was clearly not tested for either.
On a wider note, I often have to point out that computers rarely make errors. Human error is very common. When software handles a floating point operation and gives an answer which is fractionally off, this is not the computer's fault. It is simply doing what it was told to. When a computer sends a letter to a dead girl's parents, it is simply doing what it was told to, by humans who have made an error.
It's very convenient, especially in such appalling circumstances, to offload the blame onto an inanimate tool, but when you hit your thumb with a hammer it isn't the hammer's fault.
A lot of the issues i have to fix are holes in the software where no one ever considered that a user would try to do something like that, as soon as you do something that wasn;t anticipated, the whole thing comes crashing down.
<bit of a rant>
i think the problems with a lot of software stem from people not really knowing the difference between writing code and developing software.
A lot of the problems i have to deal with during what should be small, insignificant, changes stems from the original design of the software, or complete lack of one. A small change, such as allow this group of people to access this function, can end up as a major re-write of chunks of the code if someone just hard coded user id's in originally as it did the job. Or if someone just tacked a check in for something all over the place instead of writing a dedicated function as 'it'll never change'
Unfortunately, the easier it is to write code, and modern languages are easy, the less understanding you need of what you're doing and the more likely you are to write it badly. Just because an application does what it's supposed to do, doesn't necessarily mean its actually fit to be released
who wants to be blamed for making such a big cock up as the deceased girl? much better to blame the system, keep your head down and get on with it.
Of course the other side is equally something that you've somewhat missed. Although the system might be designed to handle this case, how easy is it for the user to do such? If they can open the student record and tick a box clearly marked (or change a drop down status) then this is fine. However we're talking capita here - if it takes 5 screens, 20 questions and 2 levels of managerial approval just to mark the student as deceased, then you're asking for human error.
Sometimes the parts of a system least pushed/focussed on are simplicity for user interaction, as I've seen time and time again in the business. Sadly new features are always prioritised over small quick changes that would in the long term make user lives easier and happier.
Add to that the fact that most users wouldn't come back to you and say 'couldn't you do it like this? wouldn't it be easier?' they wouldn't be any the wiser, and would assume the way it does it is the only way it can be done. It's not just a case of badly written systems/procedures.
I've always read 'System error' or 'computer problem' to actually mean 'we royally f**ked up'.
It's a fob off to get rid of unwanted press attention and to lay blame at a non-human entity.
Of course the general public are then left with impression that computers are unreliable.
I would personally have more respect for these organisations if they just told the truth, that there was a procedural error, or that somebody made a mistake, then apologise and move on. But no, they instead choose to dig themselves into a hole the size of, well you know.....
"Our instant reaction was that this was unlikely; most system errors usually turn out to have a very human origin."
Forgive me if I am wrong here but last I looked, Skynet had not yet taken over and humans were still a required part of most IT systems.
So, if a human is part of the system and makes an error then that's still a system error...!
It's still no excuse for blaming the code! They should have put the blame squarely on laziness, sloppiness, lack of training, the user being addicted to crack, user being distracted by their itchy arse or whatever the real reason for the problem was...!
According to recent reports 'shielded' data in ContactPoint became visible again when new data trawls (updates) re-wrote the formerly shielded data . Does this not imply that in ContactPoint shielding is implemented by deleting the field contents to render them invisible, rather than through the access controls which determine who can read what? Updating the data should not normally alter the access control lists or whatever equivalent is being employed.
Please ... if you get 150 letters to "sign" are you going to look into the providence of each of them? More likely the signature was just slapped on the page as a .gif file.
Also you seem to have a strange idea of what checking would entail. The person would be refering to a report which said that such and such students hadn't made 87% attendance (or whatever) that would be the same sort of list that generated the attendance letter in the first place.
I suggest you STOP blogging and get a real job where you will come into contact with real situations.
As an ex school IT monkey I've had the misfortune of working with SIMS and the story here is very plausible. The SIMS software was written (IIRC) in the early 90s and has been developed ever since as a "bitsa". Some bits are now 32-bit at last while others are still 16-bit. Capita do themselves no favours with attacks of glaring stupidity such as releasing an update for the software less than a week before the exam results were due (the one time of the year that SIMS MUST NOT FAIL) which nuked it and other attacks of bad software writing.
"If my pc was sentient it'd be cowering in a corner after all the swearing i've given it for f*cking up now and again.
Alien - computers aren't!" .... By Anonymous Coward Posted Friday 3rd April 2009 08:34 GMT
Hmmm? I shall not comment on the assertion shared there by an Anonymous Coward Being, who would appear quite deranged/alien to anyone human who would witness their behaviour with Programming Machines ...... and thus, if the Machines are built by such Beings, are they undoubtedly Alien.
* Which one should note, is not posed as a Question.
I used to support Capita SIMS in schools not so long ago. There's just one database that can be viewed in various ways depending on which menu option the user is accessing. There is a deceased flag on each pupil record and a reporting module for generating letters, attendance reports etc accesses the database based on choices made by the user (which class/year group, report type etc).
If a member of admin staff has been asked to generate attendance letters for a class/year group etc and the deceased flag has NOT been set on the pupil account, then the report will be generated for this pupil along with the others.
The head should have spotted the name for a deceased pupil among the letters, but if they had tens or hundreds of letters to sign chances are that they were just focussing on the bottom of the letter to write their signature.
The way I see it this is down to human error because; a) someone forgot to flag the pupil as deceased and b) the head and/or whoever printed the letters and put them into envelopes failed to spot the fact that the parents of the deceased child were being written to.
I suggest that you have misunderstood the purpose of a signature. Applying a signature is the same as accepting responsibility for the content of the paper being signed.
And as for the checking required surely it is as simple as calling the students form teacher, tutor, or whatever they are called these days and asking. In fact why doesn't the form teacher take charge of the thing in the first place?
If my children are the subject of any letter sent home it is personally written and signed by the teacher responsible for their well being at school so I don't see why the school in question can't do it.
Oh! Of course, now I come to think of it I do know: it's because my children go to school in Norway where the staff are expected to treat the children like human beings and expect similar treatment in return (I expect there are schools like that in the UK too but you never hear of them).
And 95% of the errors / issues are caused by the users inputting faulty data or not being trained how to use it (or other schools exporting faulty data and another school importing it).
The other 5% is down to every school management information system being an evil bastard of a piece of software altered regularly to comply with the whim of the local government requirements or programmers fixing their cockups and creating more.
We use SIMS where I work its a pain in the bottom. We have been asking for several years for the exam marks modules to be able to handle decimal numbers. All we get back is "It isnt possible to change the marks to deal with decimal numbers". How hard can it be to change a field type from integer to a real?
Our crocky home brew M$ Access database could do it. We would still be using that if SIMS hadnt been forced on us by central admin.
Perhaps the most disturbing element of this story is that amanfromMars makes a very pertinent point... "A more disturbing possible element in that tale, is that Capita are able to bill for phantom works". Indeed. Crapita ripping off the taxpayer. Who'd have thunk it?
But to return to the article, it is hardly news that people are blaming their tools, but nevertheless it is always healthy for journalists to flag up when this happens. It is bad for the health of public life if people get away with this too often because it is only a matter of time before the more unscrupulous figures in public life realise that they can use the same technique to force through unpopular decisions and bury the bad news of their own incompetence. Oh wait...
1. For insisting on the recording and collation of vast quantities of data about school pupils; this is impractical without a computer-based system.
2. For allowing the creation of ever bigger schools. With 1000 or even 2000 pupils in some secondary schools, 'care' of pupils is impractical without a computer-based system.
3. For loading such a quantity of administration on teachers (see 1. above) that they don't have the time or energy left to get to know their pupils properly.
Well you're a nice, unfeeling, little shite-hawk aren't you? My father still occasionally gets letters addressed to my late mother who died over 4 years ago, but any reminder of a lost loved one just makes memories flood back. Hopefully you'll grow-up you nasty little 12 year old troll, but I'm not holding my breath!
I do love this attitude these days that computer systems are somehow always right and that a "computer glitch" is the problem. No! No such thing as "computer error"! Systems very rarely make errors, they can't, they operate to do exactly as the software instructs them. With complete twats like EDS and Crapita and their a two-a-penny developers who obviously can't even write a simple, small scale CMS system without ballsing it up, no wonder we all fear Wacky J and her wonderful ID card system and kiddie-DNA databases!
There is a balancing act between ensuring that bad data cannot be entered and bloating the code with complex validation (complexity increases issues, you can either make a system so simple that there are obviously no errors or some complex that there are no obvious errors).
So either your users are competent and understand garbage in->garbage out or are incompetent and assume that "the magic box" knows everything, so if it's wrong it must be its fault.
" "hows about the school and educational system just admit they made a mistake and apologise."
Because that would mean admitting culpability and opening the door to being sued."
sure, i know that. however, we're all human and make mistakes from time to time, and this does not make it ok to tell lies, for any reason, just because you don't want to face any concequences. thats what a cowardly liar does (and this includes corporations).
if we make a mistake we should admit it and face the concequences. this is the way of life.
the way it is 'normal' and accepted behaviour to tell lies apon lies these days just makes me very dissapointed at people's general lack of understanding (does not apply to everyone though, there are a few honest people out there).
..and it's also dishonest and greedy for the 'victim' in this case to sue for money. they too should accept that mistakes do happen, and get over it.
I once worked for team A at a company. We recieved data from team B's system each day, and one day it all went titsup.
Team B got in first. They raised a severe incident report stating that team A's system had "failed to properly error check" the data it recieved and it was therefore its fault.
Team A responded that the source had "failed to error check its output" so it was at fault. The bun-fight escalated into a virtual Ok-corral.
Actually, neither was the source of the problem. Miles upstream of both, a user had entered bad data....
But that didn't stop each team blaming the other team's system, the database vendors, the web designers, the messaging protocol, Unix, Windows, sunspots, quantum chromodynamics and so on. Never mind that a user had entered her age as minus twelve or something. In everyone's minds it was a "system error", and as long as it wasn't their system they didn't care what the real problem was....
Not enough detail in the article to support a number of assertions made here, except one.
SIMS has been implemented in many LAs with very little user training - this refers to the updated version which is SIMS.net, not the older SIMS which was hosted within the school.
So, as Admin staff are very stretched, they will use the most familiar methods that they know to complete a task.
It is likely (my 6-pennorth of speculation) that:
The deceased flag was set on the database as stated by the head,
The pupil list was exported from SIMS so that a Word mail-merge could be completed in the time allowed.
i know reports are available, but my experience in schools has shown that if staff don't know how to use them they won't be used.
Time to get back to lurking...
As somebody who used to work on a helpdesk supporting SIMS on both the old foxpro system and the new SQL system. It is truly shockingly bad. Weird data errors cropped up all the time, response times to fix said issues often meant we ended up fixing them ourselves instead. And errors like this are definatly WELL within the bounds of things I've seen before with my own eyes.
However, that's still not a system error - that's the person/people who programmed the systems error. Computers don't make mistakes, people do. Good data in -> Junk Program -> Junk data out
I for one am glad I have nothing to do with the software at all - maybe that's why I haven't fully developed my hatred of MS yet - compared to my experiences with SIMS, Windows Vista is a dream come true :o
The problem here is the meeja's habit of trying to apportion blame the second a story comes to light. It is no longer good enough to report the facts, journalists (and I use that word in the loosest possible sense) feel that they have to do the job of a public enquiry in no more than three hurriedly scribbled paragraphs.
Public bodies have got used to this and as soon as anything goes wrong or the slightest criticism is leveled in their direction they issue a hastilly constructed statement laying the blame at anybody else's door. The real problem here appears to be that somebody at Capita wasn't so media savvy and fessed to a bug rather than contacting somebody who knew the system.
Instead of admitting that a rewrite was necessary a more sensible response from Capita would be to threaten the council with legal action.
There are two standard responses from school's when they f*ck up with pupil's data; if letters are written or data is released when it shouldn't happen then they blame the computer system; if letters are not written or data is not released when it should be then they blame the data protection act. The majority of people employed as journalists have no knowledge of the DPA or computer systems and are too lazy to do any research and just print whatever press officers tell them. Twats!
Re: How hard is is to change...
Well that depends on how complex the system is, and how well it is written (a badly written system can still be easy to change if small enought...)
Given the nature of government IT, my suspicion is that it is large AND bady designed / written (in part due to poor requirements analysis, for which the fault would lie on both government and supplier, IMV) - so probably it IS difficult to change.
'Impossible' is simply a way of saying 'Too Expensive' (IE Goverment won't pay the ammount Capita would want to charge)
There is a blame culture in the UK. As a result, there is continued effort to eliminate human error in many 'systems' (I'm using system in the generic sense). Too often, this is done by taking the decision making away from real people, and codifying it according to unvarying rules.
This is absolutely fine, but only as long as you catch EVERY possible situation that the system has to cover.
But when you get a situation that you don't cover, chances are that you will get an inappropriate result. In the meantime, your human people, who have become de-skilled (either by accident or design) because they don't have to make these decisions, are less likely to spot the inappropriate response (they keep being told that the 'system' will do the checking, so they don't have to). So they blame 'the system', and are in many cases correct in doing so.
The failure is in the creation of the rules in that poor requirements and systems analysis has been performed. This makes the 'system' flawed, but as a result of a human failing (it could be a systemic failure in the process that created the system in question - recursion here we come!)
Please note that this is not limited to computer 'systems' but can happen to any process. It just so happens that so many complex systems nowadays are centred on computers enforcing the rules.
That's why big IT systems like the NHS database or the identity card/passport database cost hundreds of millions to develop. If you really do expect to be able to analyse every single possible exception or special case that might occur in a system in advance it's a huge effort - and you'll probably still get it wrong!
Do you really expect schools to have a single all-encompasing system that codes EVERY piece of data about a student? If so, then presumably that precludes the drama teacher from keeping a handwritten list of students that expressed an interest in performing in the summer play in the back of her notebook - she'd better use the PotentialDramaActivity module in the school database instead (or wait for the IT consultants to implement it) before she can organise the play otherwise there's a risk the data might be inconsistent.
Of course the database needs to have a deceased flag in the core student record, but presumably also one for currently in prison/on extended leave of absence with parents sailing round the world for 6 months/recuperating in hospital after serious accident/missing presumed abducted, and so on, in case any of those circumstances might lead to a distressing letter being sent out too. Or they'd better employ someone who's sole job it is to compare every communication that goes out against a list of sensitive names.
"Too often, in my opinion, IT and automation is thrown in as a substitute for thinking."
It's a problem which is common in IT. Rather than producing clear and unambiguous business rules and a clear specification and then not changing any element of the rules or the specification, the business rules will be a hotpot of sometimes contradictory rules, the rules change as the product is being coded and require hasty (=buggy) fixes will to be included, and on top of all that the managers of the project will require the product the be ready in less time and for less money than is feasible and will demand the product is released early in order to meet an unrealistic delivery date.
What's really required is management with big bollocks who are prepared to politely ask the customer to stop changing their minds and/or interfering.
At the end of the day it´s hardly surprising that this kind of thing is common place. After all, most of us own a computer, and a fair few of us think we "get" computers, especially when it comes to data management. How many companies out there use software in ways that the developers didn´t originally expect?
For example, I mistreat Excel horribly at work, I have it performing basic database roles, I even force the program to give me forms that convert the end output into MySQL ready to import into the webserver. It clearly wasn´t intended to do this, but it still does it. I don´t have to know why, or even much of the how, just how to trace through the start to finish to check it´s working as intended.
Yes, getting painfull reminders of a lost loved one is not a pleasant experience, but that doesn't justify running to the press screaming and pointing is if the school themselves had murdered the poor child.
It's a completely inapropriate responce - no different to Mr and Mrs Boring running straight to the courts about their house being captured on streetview, instead of reporting it to google, who would have immediately removed it from the system.
I'll wager there is a law suit being prepared by the family right now for "emotional damage" caused with a ludicrous figure attached to it. That sum of money will come out of the schools funding, damaging our kids education.
If they had instead reported it directly to the school, they would have had a very sincere appology for the cock-up, and the member of HR respinsible for this would have been reprimanded, along with appropriate actions taken to ensure it didn't happen again.
Instead, they instigated the typical media circus witch hunt. The ONLY recourse for the school is to keep their heads down and try to pass the buck. I agree that the shirking of responsibility is a major problem, but it cannot, and will not go away until we do something about the underlying blame and sue culture that has caused it in the first place.
I heartily agree that SIMS is an appallingly designed peice of software, containing a miriad of 3rd party apps, some of which are still 16bit. But it doesn't make mistakes like this. People do.
This mailshot was not generated directly from SIMS, I can assure you. One of the HR staff will have had an independant access database exported from SIMS to do the mail shot.
Now their reason for doing this may well be very reasonable. Perhaps the particular formatting of the data, central database connectivity issues (I knew of several staff who worked on the data from home).
I'd also wager that erroneus data discrepencies such as this are very much the exception to the rule, dispite the dual database store.
Either way, mistakes happen. They always have, and always will. They won't get fixed though if everyone is too afraid to put their hand up and say "actually, I think that was my fault" because they'll get shot.
For about 8 years I was in the military. I spent a big chunk of that time working alongside civil servants. As far as I could tell the main competence that the civil service generates in it's staff is the ability to justify your own existence. The personnel codes are designed to make it pretty much impossible to sack anyone (I had a storekeeper caught with his fingers in the till who was suspended on full pay for a year after a criminal conviction before I could finally sack him). There is then a massive blame culture where the incompetent keep passing the buck.
It is pretty clear in this case that whoever signed the letter before it left the school failed to check it for accuracy. My suspicion is that because, "Not sending letters about attendance of dead kids" isn't a performance target the deputy head who signed it was too busy chasing targets to worry about basic things like accuracy of letters. In my day (which isn't that long ago) if the deputy head was about to send a letter about attendance of a kid the first thing he would do is grab the kid's form tutor and year head in the staffroom at break-time and ask them what was going on. The teacher would either point out a known problem, or get an earful for not knowing what is going on with their pupils. In this case, "little Joey is dead" would have been a rather obvious response.
And now welcome to the real world where a signature at the bottom means very, very little.
In your world when the bank fu*ks up a transaction, say charges me when then should not. I go direct to the name on the bottom of the sheet.
Pleeeease.
I generate thousands of letters per year and whilst we have never sent anyone that had died a note about their attendance, we have sent one to a student that had been excluded.
The problem is human error/ad hoc reporting/poor applications in roughly that order. That the organisation decided not to fess up is a sad indictment of human nature and nothing really we haven't seen before.
Move along please, nothing to see here.
Rather than all this rubbish about "they should have checked" "no they shouldent" we should look at the fact that a child died. Others will have been off ill, with good reason.
The Head should have at least known about ones that died. All it would have taken was some basic thought to have run a report with the letters and checked this. They should ahve had a list of pupils who had left or been cronicly ill that year to check against. 150 lines? I have just checked my SAP reports and that is about 2 1/2 sides of a4. Not much to check against a list of probably 10 people at most.
Infact I do the same sort of thing morning.
Regardless of whether one thinks stuff like this should be open source, we'd be able to settle the issues of possible bad systems architecture once and for all if the source code were available. Just like other areas of public concern, like electronic voting, without the source it otherwise boils down to outsider speculation and the vendor denying any problem and not letting anyone check to see if they've done a reasonable job.
Nuff said..
Though probably, if the system was for a prom invite, the data had probably been exported to either excel or access and used as a mailing list, which the non-technical beleive to be part of the same system.
That would explain 2 instances of the user data, but not how it got the attendance records, unless the IT bod decided to write an Access routine to query the SIMS database (probably a lot more accurately than SIMS does it) - I seem to recall that the SIMS databases aren't particularly secure.
This sort of piss poor software writing goes on all the time for school applications, no standards are adhered to except 'the writer always uses a capital letter at the start of a variable name' type of standards, and schools have to pay silly sums of money for this sort of crap because there's no other choice. (I know, used to be an IT bod in a school, got out as soon as they tried to give me a final warning for not doing something that was technically impossible with the resources we had).
</rant>
Paris, well soft ware Hurr Hurr , ahem.
One reason I do know about this sort of thing is that before taking up with El Reg I designed and managed a number of fairly large db projects.. One such project - involving the tailoring of a package solution for a financial services company - suffered from an abrupt funds shortage half way through the build.
So certain things we knew we needed suddenly had to take second priority to important stuff: like getting the data loaded and the function we had tested.
The whole deceased...do-not-mail area was one that needed far greater subtlety.
But without a budget, we came up with a cunning plan to manage the function through the data.
So we had a single goneaway flag...and decided to code it according to levels of goneaway-ness.
One was DNM (there but goneaway for comms purposes): 2 was goneaway; 3 was deceased(as in gone away with extreme prejudice).
Then the marketing bods added a requirement that we were never to mail anyone in the channel isles. So, er, 4 became "living in the Channel Isles".
An interesting extension to the karmic cycle, with spiritual progress being measured through different addresses, on to the afterlife, finally arriving at Guernsey.
Or perhaps a different take on the old saying: "when good americans die...they go to paris".
Paris... for obvious reasons...an cause she is very happy to see any good americans coming to her.
"What's really required is management with big bollocks who are prepared to politely ask the customer to stop changing their minds and/or interfering."
I agree, but only to a point. If we were to adopt that kind of rigid approach, we'd soon find we had no jobs, while the customers would go to someone prepared to be flexible.
I like it as much as you do, but that kind of flexibility is a necessary evil.
So start signing everything as anonymous coward if it means fuck all. What you don't want to do that because it looks unprofessional?
I do actually try to go to the person who's name is on the letter as it's there for a reason. The best is telling a PR droid in a company that you will forward every bit of spam to them if they don't remove you, works a treat.
People aren't held accountable anymore because idiots like you believe that it's ok to be completely unaccountable.
Probably the system and operator marked her as Dead. Normal code likely includes something similar to WHERE status <> Moved and status <> Dead, maybe hidden in a #define for NOTPRESENT. Then mail merge code was hurriedly tacked on later, and used WHERE status <> Moved. Anyway, it was something equally boring and discrete.
Okay, somebody screwed up - a peon that wrote the spec, checked the documentation, added the feature, performed an unvetted change, maybe the operator, whatever - but peons are not allowed to take responsibility and apologize. In politics and news, problems and errors and terrorist attacks all have real value.
So some unprofessional manager used the marketable affront to the parents to smack-down another manager in revenge for the Christmas party. Another used it as proof that their empire needed a larger budget. Someone - internally or externally - realized that it had additional value as 'confrontational news', and peddled it to countless reporters whose usual situationally selective conceptual grasp helped incite and excite their tabloidees at grocery checkout lines and multipage web spreads everywhere.
This poor kid and her family have probably generated 6-7 figures in turnover. No, it's no 9/11 or 7/7, but it definitely helps with the mortgage payments.
Finally, when people think "the computer", they DO include the unknown designers, programmers, Bozo the CIO, and to a lesser extent the more empathetical data entry staff - in the same way "the government" includes council planners, highway engineers, Bozo the MP, and one's letter carrier.
The wisdom of these words has stuck in my head and formed the basis of numerous software development reviews...
Author Unknown:
"I really hate this damn machine
I'd really love to sell it
It never does quite what I want
But only what I tell it"
no one takes into account the minimum wagers who are inputting this data, who *may not* have job satisfaction and purposely enter erroneous details.
I remember the good old days of temping, at 18 with other 18-25 year olds, for an unsecure loans company. People (i never did) would change names to Mr cock, Mr Dick. etc etc. or what their loan was for (buying an elephant, penis enlargement) and then someone else in another office would have that on their system when the computer auto called for them.
'sorry it whats the computer said'
I work at a hospital (not in the UK) supporting a horrible niche Electronic Patient Record system, and yes, some users make stupid mistakes, but a lot of the mistakes are not exactly helped by a bloody patient system with no data validatation whatsoever above that reqired by law. The system (incredibly) provides basic SQL-injection protection, and validates a few fields that will be exported to other applications, but it will happily re-admit a patient who has formerly been discharged as deceased (that flag is in the case table and not the patient table).
Dates are stored as integers where 0 is 1/1/1800. If no date is entered it used to default to 0 (we actually got that "fixed" after a few years of pestering the vendor. It took them than long to add "WHERE date > 0" to the relevant queries).
It allows a patient to be discharged before it is admitted.
It allowed duplicate patients (for 15 years) untill the latest patch. Lucklily we have been riding herd on the monstrous database (constraints? Never 'eard of those.) and have periodically merged the duplicates (by updating the DB directly). I know several other hopitals running the same system have not, and they now have a real bitch of a job to do.
The point? All this could be classified as "human error" on the part of the users, because it works just fine if all data entered are valid, and all the reqired data are present. It could be classified as "human error" on the part of the vendor, but this system has... evolved since it first entered production round 1993, the vendor has changed twice and is about to change again, and there is never enough money to throw it away and start from scratch (or better yet, extend the other EPR in use at the hospital to cover the needs of this niche medical profession as well). It could be classified as a "human error" on the part of hospital mangement for not finding this money, and not pressuring the vendor for more functionality before the existing one works better.
But altogether I call this a system error. Not a computer error mind, that is not the same thing.
The system is not just the computers, the system most definately also includes the users, the maintenance org, the supplier, government regulations and local mangement.
(1) The technology industry assumes users are experts by not documenting how their stuff actually works.
(2) The technology continues to berate users for cocking things up.
(3) The reality is that all users, even now technically competent ones, start life as novices. Until the engineers in the tech industry wise up to this fact and actually start writing good documentation, not just the crap "online help" which seems to have permeated recently we are going to get stuck...
That said - managers of the purchasing company are generally to blame here too - most are willing to spend £500K on some uber project, and then refuse to sign off a £10K budget to actually train their staff on how to use it. All about the kudos of getting in the expensive system, and not very much accountability for it actually working...
I'm involved in maintaining and supporting the social care database for a local authority. We've seen practically every cock-up a human could possibly make...
Duplicate people - often because social workers and admin staff can't get their head around searching using wildcards, so create a second copy of the person, or they can't spell the name provided over the phone... Sometimes they notice the mistake early, so it's fairly easy for us to eliminate the duplicate. Sometimes it's only noticed after several years, when both copies have assessments attached. And as the person who created the duplicate has probably left by then, the teams unsurprisingly don't want to concact the data onto one record. (We've even had a couple of triplicates and a quintuplet - now imagine what chaos will ensure when ContactPoint goes live...)
Assessments ended prematurely, with the wrong dates, or the wrong outcome.
People "moved" to the wrong address - or even with no address (that impresses us - it's supposedly impossible to delete a client's main address, but our social workers can and do - whereupon the record cannot be found through the database front-end, and we have to put the address back on through the DB back-end)
Incorrectly entered service packages.
Not to mention...
First name: Frederick
Surname: Bloggs (prefers to be called Fred)
The system has "Other name" fields (previous name, alias, preferred name etc.), but they don't print out on reports...
Requests to add an address to the database's gazetteer - often with an incorrect postcode, incorrect town name, or a complete mess-up. (OK, so this is the client giving dodgy information, but it's still a challenge to validate the address - unsurprisingly, the Council Tax Valuation List is often more up-to-date than the Royal Mail's database...)
There was only need for one secretary because she didn't spend her day sending letters for non-attendance. She knew who wasn't attending, and when it got sufficiently bad she'd ring the parents and make an appointment for them to visit the headmaster.
Someone, somewhere, decided it would be efficient to replace this person with a machine. Later on, someone else decided to recruit a person to deal with all the computer-work, but they never had the breadth of responsibility of the school secretary. And last of all, someone decided to send letters home (lovely traceable, blame shifting paperwork) rather than dealing directly with the problem (hard, messy dealing with people and their lives).
That you can export the database for reports etc
So what are the odds they were using a very old export of the database, or possibly they had two copies (one probably unlicensed) and they HAD set the deceased flag, but only in one DB and not the other.
Smacks of user error, the "second database" statement pretty much confirms that I think.
Anonymous Coward said "Please ... if you get 150 letters to "sign" are you going to look into the providence of each of them?"
And as I said, in my initial response, if there were so many letters you couldn't check them, e.g. if you are getting 150 letters to sign saying that the student has too poor an attendance record to attend the prom, (a) you might as well cancel it as most of the students wont be attending, and (2) you have such a massive issue with poor attendance it's WAY beyond sending out form letters
They were actually auto sending out gcse english pass certificates to everyone who had ever registered for the school newsletter, thus proving they could read, to boost their league table standings....(the fact it was opt-out was irrelevant)...
..and they would've gotten away with it too,...
If the same developer made the software that handles the database and the software that handles the mail merge, shame on them! Do one thing well. One thing! Allow the user to export the data/reports in a variety of types, and let professional mail merge software handle the letters.
If the user made an export of the form: While Attendance<n, export(Name_Student, Attendance, Name_Parent_Custodial, Address_Parent_Custodial), shame on the user for (in order) not (a) Checking the results. (b) Accepting responsibility for the error. or (c) Considering the deceased flag in the query.
Yes, the lack of technical skill is the LEAST important thing here- The user knew his or her level of skill, and should have run a common sense check against the results.
The system is still in error, and more than a software patch is needed.
I worked in IT support for schools for 7 years until very recently. In that time I supported SIMS, Facility CMIS and just before I left, RM's Integris G2. All 3 of these pieces of school management software were completely and utterly atrocious and it's very, very believable that this was indeed a system error.
Giving these systems the benefit of the doubt that they were programmed correctly and that there was no data corruption shows utter ignorance as to how terribly bad these systems are.
Judging by the issues faced with these systems I'd say it's highly likely that this was indeed a system error as they truly are that bad and that unreliable.
I do agree with your point in the article John, that systems are often unfairly blamed but you really could and should have found a far better example than this because this is one of those cases where the system almost certainly is to blame.
Although information may be stored in one database, there are issues surrounding the database being developed to a reasonable normal form (3NF or upwards) and tables not being sensibly cross-referenced and changes not being correctly cascaded.
SIMS and all other School Management Software out there is abysmal. We're talking about software so badly written that even Windows ME appears like a gem of perfect software development when compared relatively. We're talking about a system where updates or fixes might come as a bunch of files copied straight from the developers computer into a zip file including their local configurations which you have to extract and mangle back by hand to suit your configuration even if it's more standard as per their documentation than the developer's often customised setup! You'll probably then have to run some dodgy db fix app or similar that then runs through the database and mangles it to suit the update, but of course you'll be reminded to back up your database first because this may not work and then you'll have to get their support to fix your data by hand too.
See if Capita will send you a copy of SIMS to evaluate and have a go at setting it up and installing it, play around with it, see what you think. I think you'll be dissapointed when you realise how bad it is and how tax payers money is being wasted on this crap. If I thought I stood a chance of selling the stuff to schools (which I don't because backhanders and government buddies etc. ensure a select few companies get all the contracts) I'd say the market is ripe for the picking for a team of 3 or 4 competent developers to make a small fortune.
The other software, Facilicty CMS is just as bad, and RM's integris G2 is web based and horrifically bad also - the system regularly keels over under the load (apparently RM don't believe in stress testing) their database server crashed once or twice and was down for hours (apparently redundancy and failover devices aren't something they believe in either) and simple errors such as users not logging out of the system properly leaving pupil records locked which you can only get unlocked for staff to access again by contacting the Integris helpdesk were all too common. Oh, and it was advertised as working in Windows 2000 with IE5.5 and Firefox upwards except it never worked properly in IE6 at the time, so it was even sold on lies.
Make no mistake, the UK school management software market is an embarassment to IT, it is an utter fucking shambles. It is an abomination that should be erased from the planet and that we should start from scratch on with competent development teams.