No Boats for larry!
So they've been taken to the cleaners, or should I say the taxpayer has funded Larry's boat obsession for long enough. For fuck sake!
Whitehall bean counters have ordered government departments to find fresh ways to end their reliance on Oracle. The Cabinet Office is understood to have formally contacted central agencies within the last month and asked them to look for ways to “get rid of Oracle". No. 10 is believed to be concerned about the amount civil …
And fighter aircraft not that I would not do the same type of thing if I had the boodle
Oh well, Google just rented the Valley of Temples in Sicily for their own party - maybe they didn't try to buy it because if wasn't on sale (anyway, it was rented for a mere 100,000 euro, pocket money, for them).
Guess megalomania is becoming a typical job-related illness for many CEOs...
Ahh someone who has not come into 'conflict' with the Oracle licensing proposition. It is not as simple as 'one licence per user' - it can be anywhere from that up to 'one licence per user per customised app per application per oracle install per process cores per colour of CTOs underwear' and other variations on the theme.
It is tempting to suggest that the licences were agreed by legal beetles (actually not beagles have more sense) without projecting costs with increased system complexity, growth in the user group, and expansion of database applications.
It makes TP & NG's cry for sanity look simple....
“Along with the standard computer warranty agreement which said that if the machine 1) didn't work, 2) didn't do what the expensive advertisements said, 3) electrocuted the immediate neighborhood, 4) and in fact failed entirely to be inside the expensive box when you opened it, this was expressly, absolutely, implicitly and in no event the fault or responsibility of the manufacturer, that the purchaser should consider himself lucky to be allowed to give his money to the manufacturer, and that any attempt to treat what had just been paid for as the purchaser's own property would result in the attentions of serious men with menacing briefcases and very thin watches. Crowley had been extremely impressed with the warranties offered by the computer industry, and had in fact sent a bundle Below to the department that drew up the Immortal Soul agreements, with a yellow memo form attached just saying: 'Learn, guys...”
This reminds me of a bit out of Good Omens:
Along with the standard computer warranty agreement which said that if the machine
1) didn't work,
2) didn't do what the expensive advertisements said,
3) electrocuted the immediate neighborhood,
4) and in fact failed entirely to be inside the expensive box when you opened it, this was expressly, absolutely, implicitly and in no event the fault or responsibility of the manufacturer, that the purchaser should consider himself lucky to be allowed to give his money to the manufacturer, and that any attempt to treat what had just been paid for as the purchaser's own property would result in the attentions of serious men with menacing briefcases and very thin watches.
Crowley had been extremely impressed with the warranties offered by the computer industry, and had in fact sent a bundle Below to the department that drew up the Immortal Soul agreements, with a yellow memo form attached just saying: "Learn, guys..."
Oracle probably employs the same people (for licensing and/or selling your soul. My understanding is that they're roughly equivalent)
The number of licenses Oracle require for their systems is ridiculous and, probably intentionally, very confusing. They probably get less customers because of the difficulty but those they get usually end up paying vastly more than they should so Oracle's happy.
Here in Denmark, an Oracle-based database was used for a traffic information system while expanding one of the main highways around Copenhagen.
Oracle initially tried to license it "per user", meaning every car that could potentially pass by one of the electronic information signs. With roughly 1 million cars in the greather Copenhagen area, this would have contributed nicely to some Oracle serfs' bonus account.
They tried the same stunt on the database behind (at the time) one of the biggest e-commerce sites in the country until we told them to go forth and multiply. They wanted to count each unique visitor as a licensed user with thousands visiting per month.
Move away from each and every proprietary piece of junk you're running, switch to FS.
Get PostgreSQL on GNU/Linux, switch all desktops to Linux ...pay a flatrate for support, I am sure they would end up paying 1/1000 of the price for licensing, compared to what they pay now.
They would have to train everybody for Windows 8/10 anyway, might as well train them to use Linux.
Instead of actually running Linux desktops, an even better plan would be to switch users to remote Linux desktops. Easier to manage, and you avoid the hardware compatibility hell of installing Linux on diverse desktops and laptops, whose manufacturers care less than nothing about Linux compatibility. The clients could in fact be cheap "landfill" Windows laptops, because the only application of interest is the remote client (VNC or similar) that requires very little memory or CPU power. If the machine breaks, it is simply recycled and the employee goes to get a new one at the office depot. Support costs would plummet.
To be fair Libre/Open Office isn't as good as MS Office, I've been trying for years to reconcile the issue but it doesn't come out in the wash. I did recently spend two years using mostly Google Docs, it justifies the lack of features because the document collaboration is so immensely good but the formatting and working options aren't nearly as good.
But in contra to my own argument my mother has been using Linux as a desktop for the past few years and is barely concerned that it isn't Windows.
Who cares how well internal government documents are formatted? Let them use plain text.
...but then how will the legions of civil servants while away their days, without those halcyon hours spent mindlessly stabbing at [B] and [I] buttons and fighting with autoindentation, you psychopath?
"...but then how will the legions of civil servants while away their days, without those halcyon hours spent mindlessly stabbing at [B] and [I] buttons and fighting with autoindentation, you psychopath?"
Put'em to useful work! Are there no streets that need cleaning? No chimneys that need to be swept?
"But for 99% of users it is perfectly adequate"
In any sort of normal work environment, I think you mean more like for 9.9% of users it is perfectly adequate.....
For starters there is no VBA support, very flaky ODF support, loads more bugs, it is much slower, and is missing lots of features and integration options compared to MS Office.
Ok so LO is not perfect - but the quantity of cash governments are throwing at $CORPS, it could be fixed using target funding.
The central problem is there is a vested interest in having gigantic companies run these contracts, so the middle manager machinery justifies its existence.
Having an enterprise Linux installation with customisation as necessary funded from the deployment costs, returned to the software packages.
Let's stop paying for problems that were solved decades ago, and the whole boat rises....
P.
"ODF was invented in Libre/Openoffice"
Nope, OpenDocument is not the same as the older OpenOffice.org XML format and these formats are not directly compatible. ODF was actually developed by OASIS.
"It's MS'es implementation of it which is flakey."
Nope - MS have by far the best ODF implementation that is currently available - and it fully support the latest ODF version - even in Office 365. If you look through LO support forums and bug trackers you will see that there are loads of issues with it's implementation of ODF. For instance crashes saving a file that can be opened / saved in other Office packages.
Just a few examples:
https://bugs.launchpad.net/ubuntu/+source/libreoffice/+bug/1292360
https://bugs.launchpad.net/ubuntu/+source/libreoffice/+bug/1412448
https://bugs.launchpad.net/ubuntu/+source/libreoffice/+bug/1072033
Also there are vast numbers of bugs when handling MS Office formats - that most businesses rely on.
The malware part is true, but a lot of organizations rely on heavily "programmed" document templates, especially in the spreadsheet world. I've seen banks with programmers dedicated to it - and those templates generated document to manage a lot of money. I guess in many government departments is more or less the same.
Do you really believe saving some dollars/euro to get rid of MS Office, lose those functionalites and need to re-implement them, looks a good move for them?
Sure, maybe not every office needs them, but as long as some offices need them, to simplify deployment, management and interoperability you choose a single product.
Open/LibreOffice needs to cover this requirements also, if it aims to become a full replacement for MS Office.
"But for 99% of users it is perfectly adequate."
It's that statement which has Microsoft shitting housebricks.
Word and their ilk weren't as good as WordPerfect. But they were good enough - and cheaper.
Postgresql isn't as good as Oracle - but its good enough for 99% of usage cases.
And the raw facts are that Oracle make a great database, but _everything_ else they touch turns to a steaming pile of shit - ESPECIALLY their financial packages. My employer just moved from a terrible Oracle product to an even worse one.
Oracle, lilke other big companies in the same arena, suffers from the "Indian syndrome" - applications built out of a lot of Java (or the like) code written by lame Indian programmers paid a few rupees, while the applications are sold for many $$$$$.
Not that our SAP system is built by more competente programmers...
To be fair Libre/Open Office isn't as good as MS Office, I've been trying for years to reconcile the issue but it doesn't come out in the wash.
I beg to differ. It is nothing but a matter of what you have grow used to. I have mostly been using OpenOffice and LibreOffice for years, and now trying to MS Office for some project is driving me nuts. LibreOffice just fits my way of working better, and gives more support for producing a consistently formatted document. Maybe there are some places where MS Office wins, but I don't know of any other than being more compatible with the MS Office file formats. This advantage would disappear if more people used OpenOffice, LibreOffice or other tools that properly process OpenDocument formats.
Maybe some day...
Only if you absolutely have to have the dancing bear at the bottom of evey page. Everyone learns some "cute" trick in MS office and then pesters the world. I simply want text and some MINOR formatting for readability If you are saying MS is better because it has more "features" then you are right, but if you are claiming to be better because the product produced is clear and presentable you are far from correct.
OpenOffice Writer is great. I had to start using it a few months ago for a course I'm doing and it suddenly made so much more sense having the same application on my MacBook and my Windows desktop.
Running Office 2011 on my MBP, Wordpad on my Windows desktop (on the few occasions I needed to write something) and MS Office at work was just horrible. If I could use Writer at work I would, it doesn't randomly f**k up my document formatting like Word does.
You fail to notice that a lot of those Oracle instances already run on Linux, for which both Red Hat and Oracle itself make you pay happily. Also Oracle makes maybe more money from "mandatory" support than the licenses themselves.
The GPL doesn't enforce any price or support model - what make you believe if everybody switch to FS you will be offered only cheap flatrate support?
Once you get locked in some "enterprise" supported version of Linux/Postgres/whatever you like from which you can't simply switch away easily - and being FS doesn't make it easier at all (unless you're Google, maybe) - do you believe you will still get a cheap flat rate support price? Do you have ever read Red Hat support agreement, for example?
I guess they will start to make you pay per seat, per user, per processor per whatever they like to make more money - most people like money, a lot of them. And less competition, the more you can charge - that's why Oracle can charge so much and still be able to sell obscenely priced products.
@Hans 1.Yes indeed, Oracle moved to Linux desktops internally years ago. I have worked in fairly big and fairly well run IT companies and I know it's a hell of a problem to keep track of all the licenses used. Invoices are cleared by people who don't know enough about it. And developers/programmers feel it's not their job to report on what's in use and what has been scrapped years ago. It's not just Oracle it's all of it including even hardware. The more people involved the more difficult it gets, and the invoices just keep coming.
"Get PostgreSQL on GNU/Linux, switch all desktops to Linux"
Oooh! The boom in consultancy for re-inventing all of the unique functions that have been created for the various government departments, and the training at the desktop.
"They would have to train everybody for Windows 8/10 anyway, might as well train them to use Linux."
a) what makes you think that they are going to move off XP/7?
b) The few that i am aware of are/have already decided that ClassicShell is the way to go.
Sure, but replace it with what ?
I cannot imagine the sheer volume of data stored and managed in a government-level group of Oracle databases, even taking into account the fact that there is no oversight or global planning of any kind (good one there, way to ensure economies are impossible).
You don't just get rid of Oracle. You need to find a suitable substitute, redesign the applications, create a parallel environment, migrate the data onto said parallel platform, design a test suite to validate the transfer and test the new environment and validate it.
When all that is done, you can start migrating the users. You'll be doing that bit by bit, so as not to have everything blow up in your face with the inevitable unexpected complications that will arise.
Just planning this kind of thing will be a major project which will require external expertise.
Hmm, do I smell the presence of yet another pork barrel for yet another vast, government-wide IT failure ahem, project ?
Some databases may be easy enough to migrate to other RDBMS. Others could be very, very difficult - a lot depends on how they were implemented, and what Oracle features they use. It is true Oracle is absurdly expensive and its licensing is close to blackmailing, but it is also true it offers some very advanced features other databases often lack, and you can't easily replace if your rely on them.
Also, moving to another database may mean to redesign the applications built upon it because transaction/concurrency model are often not identical, and if you don't care about it, bad things will happen.
It is also true there are also DBAs so worried about upgrades they would still run Oracle 6 on a VAX system if they only could find the hardware - forcing you to buy even more expensive support (and Oracle knows it and doesn't make upgrades as easy as you could wish...).
Moving away may not be easy, and Oracle takes advantage of it. An effort from some big customers to move away from Oracle is welcome, maybe someone there will rething its licensing strategy - but I'm afraid if someone else just believe to slap a Postgres/<put your preferred RDBMS here> database here and there and import data, it could become soon another expensive bloodbath.
Moreover, it looks what they really need is a good database to manage licenses to avoid buying millions they don't need...
"expensive bloodbath" - a touch alarming!!!
If there are problems, write a RFP for modifications for $RDMS or $OFFICESUITE to address them.
Look at the software that runs on supercomputers, that has been funded to solve very complex problems.
It can be done, and *stays* done.
P.
Guess you never saw - and thereby worked on - a large, complex database with a lot of complex data types used (which may not be easily mapped to other platforms types), a lot of data logic (correctly) implemented as packages/stored procedures, and a lot of highly optimized complex queries using often some database specific SQL features. And that without speaking of some advanced features which may be available on some high-end databases and not on others - and could require third party tools or be reimplemented maybe at the application leve.
Trying to move this kind of databases to a different engine is a very long, complex task. Sure, if you have small "data dumps" - data written in simple tables, and nothing else - and simple CRUD applications, porting could be easy enough.
Software that runs on supercomputer has nothing to do with large databases - but maybe some big data analytics that work on data *extracted* from a large database in a format suitable for processing.
But if you believe IT is an OS and an office suite, well, believe me, there is far far more to learn.... one day, maybe, you'll learn by experience...
"Just planning this kind of thing will be a major project which will require external expertise."
This is why oracle will be safe for a fair few years, I bet they will ultimately decline like IBM, because of current business model of screw all the current customers. Larry doesn't care about that far in the future, he wants his boats today.
Then I'll be able to make a fortune contracting to create training for the new systems. After that I'll be able to do the same again when half of them move back to Oracle because the new systems don't do what they wanted. Retirement plan sorted!
"How can you be too cynical?"
Two years of hell first trying to stop Oracle being brought in by themanagement lads that wanted Big Kit, and then trying to get it to do what we needed it to do (a zillion kilometres of bespoking) and watching the money in the budget evaporate into Oracle's coffers while we limped on... I left, underwent intensive therapy, and am a abetter person for it (the twitches are pretty infrequent now).
Hire more employees, then the software per employee cost will fall.
This is absurd, only a politician could come up with this nonsense.
I am all for removing Oracle, but seriously, license per employee? So in the "IT revolution" they pretend to spend less.. and expensive software that reduces headcount and saves money will be automatically discarded.
Bloody hate 'Orable - had to put up with it during my stint at UK Gov. The story, which should sound depressingly familiar to any IT bod, goes:
1 - New HR Director arrives
2 - HR Director wants to be seen to make changes
3 - HR Director, an 'Orable fan, gets approval to replace the existing, perfectly fine Employee Database and Holiday system
4 - ??? (I suspect that various meetings happen during Golf with greasy salesmen)
5 - New system gets implemented, massively over budget
6 - Staff hate it, absence records become sporadic at best, everyone gets a bonus holiday day due to some weird quirk about it being a leap year, paper forms are soon introduced to help track manager approvals, appraisals (Ha!) are no longer complete... I could go on, but you get the idea
7 - Next round of the Employee Satisfaction Survey, the new system is Public Enemy Number 1, 2 and 4 out of 5 for all staff
8 - HR Director inexplicably gets a bonus and pay raise (seen in the next annual report).
And now I am getting a headache as I just found myself grinding my teeth in frustration and acrimony
Management (I reckon 3 got blurred with 4 somewhat). This was mid 2000's sometime, during the heady days of the Blair era of New Labour, where cash was just splurged around and many empires sprung up and some real dubious, big brother style stuff got done (Heard about something on the grapevine called the ILR from another agency that was rumoured by some staff to be part of the ID card scheme. Not sure if it is true or not though). Government jobs seemed to be everywhere and easy to get...
We have Oracle and Peoplesoft.
The Oracle timesheet won't let you enter a time of zero hours. But when you add an activity it fills evey line Mon-Friday with a '0'.
Then you can't actually put in the hours that you work. Only the standard work week hours.
They have the nerve to actually charge $$Lots$$$Lots$$$Lots$$ for this POS?
Then the expense claims section has error message that bear no resemblance to the real problem. The last three expanse claims I submitted got 'lost'. I got the ack email but it never reached the person who was suppose to approve them. POS.
Peoplesoft is so slooooooowwwwwwww it hurts. Even if you access it when no one else is using it (04:00 Uk) it takes an age to do anything.
At the moment, I hope ORacle goes belly up
"I've been banging on at work to transition to EnterpriseDB for years. It's still massive amounts of work, money, training etc... Oracle really have you over the barrel once you've gone down that particular rabbit hole."
One day you will have to move on from what you have now simply because Oracle won't actually support it any longer - however much money you throw at it. You then have the choice of throwing more money at Oracle to help you move to another Oracle system or move on to something else. The *real* transition cost should work out the same either way because you are having to fund replacement of the system & retraining.
The sooner you make the decision to move to something else, preferably something that can be sourced from multiple vendors so you aren't held hostage to lock-in, the lower that transition cost will be.
Deferring the decision break the lock-in simply ensures that your profit margins and business decisions will be dictated by what suits the vendor. Meanwhile the vendor is aiming to maximise their profits, which in practice means charging as much as their customers can bear in return for delivering as little as possible. Of course if the vendor decides to cut you off because they aren't making enough money out of your organisation you are SoL.
Folks can argue about the details until they are blue in the face it doesn't change the big picture.
Last time I touched the car-crash that is Universal Credit, it was an Oracle salesperson's wet dream. The business rules engine, Oracle Policy Automation, was at the throbbing heart of UC and was licenced to the value of Xty meellion dollars even after very very deep discounts.
But that was probably a special case - OPA is designed for big fish government projects.
I'm sure some beard face from GDS thought he could replace it with a javascript library.
Is Oracle the problem, or is it the symptom of the problem?
It is somewhat frustrating to think that if a decent FOSS solution existed for any of the many IT tasks that governments have to perform, then governments all over the world could probably adapt it to their local laws (quite possibly for free, since it is in citizens' interest to keep their own government's costs down).
Of course, you'd need a decent spec. That's probably why it hasn't happened and why the commercial sector (who don't have to publish the spec, let alone the code) are the only ones who can cope with government contracts.
Perhaps that's what really needs to change here. If the Cabinet Office insisted on publishing the detailed *requirements* for every software contract, that would give the contract winners less wiggle-room to produce over-priced rubbish. They might even get some free submissions, and then contractors would have to compete with those (or support-contracts-for-those).
LDS: That's why the "O" is important. By starting with the requirements in the public domain and by encouraging FOSS solutions where the solution would be in the public domain, we reduce the risk of waking up in ten years time having spent all the money and developed nothing at all of any value, but it has all been messed up in secret so we don't know who to blame.
A marketing statement, yours, which actually says nothing. The quality of requirements, and the quality of outcome, has nothing to do with being FOSS or not - depends only on the quality and skills of those working on them.
A government funded FOSS project can easily suffer from all the same issues government projects suffer from - i.e. assigning the funds to entities for political reasons and advantages instead of quality/skills reasons, and in turn this could keep away other people who could contribute on the project. Sure, you would have a public domain project - which is still crap and of no use.
From where I work, I already see it, with government projects assigned to university X in city A only because the minister in charge is from city A too, and gets his or her votes there. If the minister changes, then funds could be reassigned....
Ah, government specs. Always good for a laugh those. They often read something like:
1. New system should do everything current system does, only faster. No, we don't have a feature list of current system
2. New system should include all the things we think we'd like.
3. Including the things we've not realised yet.
4. Oh, and a bunch of things we had assumed as well
5. Plus whatever the committee(s) wants added too.
6. This should cost x. Yes, we know it will end up costing x^y but for now we'll say x.
Then once someone has made an attempt based of the partial requirements list, repeat items 2 through 5 until it becomes clear the system needs to be redesigned from the ground up.
You've never seen a clusterfuck like a combined public/private sector clusterfuck.
"You've never seen a clusterfuck like a combined public/private sector clusterfuck."
The best way is to first of all work out what the current system does and needs to do better.
That isn't done by talking to management, nor is it done by committee.
A classic example is the businessman who set a design committee to provide a device to allow hikers to be rescued if they got into trouble. It ended up as an all-singing, all-dancing map, smart-phone device that would have lasted half a day on batteries - not the EPIRB unit he usually chucked in his backpack when hiking that has batteries lasting 6 months or more.
He knew they'd do that. The object of the exercise was to establish why committees are one of the worst possible ways to design systems.
I've yet to see researchers for any "overhaul" of a system sit down individually with people who actually have to work with the existing system and ask what they need to do and what's needed to do it better.
"Is Oracle the problem, or is it the symptom of the problem?"
Well you might ask.....
The primary reason that projects are screwed up is because targets are poorly defined and goalposts constantly being moved (not to mention being changed along the way form simple wooden uprights to something not out of place on a Mod's scooter)
Most of the time, the project won't work as specified, however any vendor who says that will be summarily ejected from the bidding process. Those who keep quiet can hit the original target on budget, then charge a fortune for all the extra work required to make it go.
Dealing with incompetents is a wet dream for any large vendor. They get to write their own contract conditions and then charge whatever they want for as long as they want. Whether it works at the end or not is irrelevant - what matters is getting the most profit and as long as you hit the initial poorly designed targets you're on the gravy train for life.
If you want to fix this kind of thing, you need to sack the career civil servants who skip merrily from clusterfuck to clusterfuck with glowing recommendations because that's the only way to get rid of them. Ordinarily a set of glowing recommendations should be a big red flashing warning sign and klaxon for any potential employer, but these people are hired according to who they know not how good they are (a really good employee can pretty much command his/her own salary if they're that critical)
Having proper service definitions and competent people working out what's actually required is more expensive up front, but better long term. Because of that, it will never happen in the UK.
On the one hand, Oracle pricing astonishes me. On the other hand, a point that seems lost on many is that a dollar spent on consultants or otherwise on support weighs as heavily as a dollar spent on licensing or hardware. I wish the UK all luck in cutting loose, but the process of doing so could be extremely expensive in the short term.
Not consultants. RFP, the result, source code back into the pool.
Apply some proper project management, but once you have a proper list of subject neutral specifications (e.g. must work with MS Office or $TOOL is a common way to crowbar in specific licenses).
I have a bias. sure. I am scientist and we would never make progress if we had these crappy license issues.
But these are solved problems and not cutting edge research.
P.
"I am scientist and we would never make progress if we had these crappy license issues."
The scientists I work with insist on using "software" (IDL) manifestly unsuited to the tasks they're trying to do, because many years ago someone thought it was a good idea to use it for displaying results and then extended that to doing heavy math processing data behind the displaying. The software company exists pretty much to service this science discipline worldwide, despite the software being unsuitable to the task.
They use it because they've always used it and quite frankly the coding quality of most "scientists" is abombinable.
My classic example is the IDL mirroring program - which called wget 2.5 million times to retrieve files from a webserver, instead of simply using wget with the right flags - once.
The person behind that particular pecadillo writes code for systems which is used extensively by the insurance industry to calculate risks due to extreme weather events (This is worth a few tens of billions of pounds each year to the industry). When he was "indisposed" recently, the system had to be shut down as all the inputs were hard coded for each run and was written so badly that noone could understand what the code was trying to do.
At least half the code I run across that's written in-house uses tools which aren't so much a hammer being used to bang in screws as a carpenter's plane being used as the hammer for that task.
And of course they won't hire in actual analysts or coders to implement things properly because that would cost too much.
Anon, for obvious reasons.
At last the UK Government is using common sense and intelligence, in first adopting Open Document Format (ODF) standard for full compatibility with and to replace Microsoft Office lockin and costs, and now (rightly) kicking ridiculously priced and cumbersome Oracle technology to the curb.
Fortunately for them, the very latest Free/Open Source Software (FOSS) PostgreSQL database suite of applications offer probably about 90% plus of the functionality of top Oracle products, and the similarities in admin and configuration language tools make it significantly easier to migrate with minimum of pain. Such an adoption of PostgreSQL would also allow the UK government to customize the applications specifically the the needs of government agencies without any significant pounds outlay to Larry Ellison and Co.
It is about time the UK stop mindlessly following much of the fading USA technology dominance.
"PostgreSQL database suite of applications offer probably about 90%"
Do you have any reference to support this claim? Moreover, government databases are often on the large, complex side - I may agree that Postgres covers most of the needs for small-medium sized databases, and some large ones, but its coverage for very large, complex database is still not so complete.
Even if Postgres offers the easiest path to migrate from Oracle, it could not be the right choice if you really want to migrate away from Oracle, depending on your needs.
Customizing a database engine requires skilled programmers in the field - not just a DBA...
"you avoid the hardware compatibility hell of installing Linux on diverse desktops and laptops, whose manufacturers care less than nothing about Linux compatibility. "
There is no hardware compatibility hell. I'm serious, you can find plenty of stories of people running into problems with some computer or other, and it makes it sound like hardware support is an utter basket case. But you're ONLY seeing the reports of people who have something not work out of the box.
The reality is, you're FAR more likely with Linux to be able to just pop in the install CD (or USB), install it, click "alternative hardware drivers" thing IF it shows up (mostly this installs nvidia's or ATI's provided video drivers instead of the default free & open source ones), and you are done, then I've ever seen with a Windows install.
The Windows install, you'll be able to get to work too, assuming your hardware isn't too new or too old for your WIndows version, but you're almost guaranteed to have to download extra drivers instead of having them included and possibly fiddle with things.
All that said, I'd heartily recommend ditching Windows for their desktops and going to a nice Linux distro, Microsoft probably is a large part of their costs after Oracle, but, if they have all these Windows-specific apps I could see not doing it. And I certainly wouldn't bother to rip-and-replace just to do it, I'm all for transitioning things over in an orderly manner in a case like that.
"There is no hardware compatibility hell. I'm serious, you can find plenty of stories of people running into problems with some computer or other, and it makes it sound like hardware support is an utter basket case. "
The only time there are problems is when people buy systems blindly and then try to make them work afterwards. People do this at home because they buy as cheap as possible. If you do that when you're purchasing a fleet of systems then you deserve to have your arse bounced on the sidewalk outside the front door as security eject you from the building.
Those defra licenses will be for "RM". I work for the DWP and have to use RM for personel related things like overtime and expenses claims and logging staff sickness absences. If I need to reset my password I email a defra email address. I'm pretty certain HMRC use RM too and I guess most other departments do too.
Although it's a shitty way to account for the licences it's not quite as scandalous as 200 licences per member of defra staff.
*I'm just guessing all of this but it makes pefect sense.
I'm sure they realize the only sane alternative is to go full out open source, and redirect spending to salaries and educational benefits for top notch in-house talent.
Governments and companies have been stupidly throwing money at big firms like Oracle, Microsoft and IBM for decades, leaving little to build and maintain an in-house team -- leading to further dependence on the big boys for support that never really materializes.
Of course this is probably just a feint to allow the Microsoft lovers at No. 10 to flood the country with PFY subcontractors whose CVs come up in a bulk search for ".NET development". I heard HMRC is now using tax form that can only be filed using Adobe's proprietary Reader software -- someone should go take a look around management offices there and see if they find any undeclared Adobe branded swag.
The definition for insanity comes to mind at this moment: doing the same thing over and over but expecting a different result. But then we already knew the people running the governments in the UK and US are completely mad, so no real news there.
"Makes you wonder if they dumped all the IT solutions and went back to pen, paper, interdepartmental memo's and people filling in forms in triplicate it would save money and probably be more efficient...."
Honestly: It would do.
UK civil service computerisation has almost entirely been done in a way as to not reduce staffing requirements.
If done right, then staffing numbers should be slashed by half without affecting the coalface and without imposing more crap on them.
Any time more work is dumped on the coalface or ends up requiring more people to make it all go, the project has failed.