A publicly funded Linux project which cost UK taxpayers half-a-million pounds has flopped.
Birmingham City Council began the project — one of the largest public-sector Linux projects in the UK — in May 2005 to evaluate the potential of open-source software. The council, the largest local authority in the UK, intended to deploy open-source software on 1,500 PCs in libraries across the city.
But the project has fallen vastly short of expectations, with just 200 Linux PCs being deployed. Even some of those have been migrated back to Windows, council executives have told ZDNet UK.
We will continue with a mixed economy...I'm not an open-source fanatic.
Glyn Evans, Birmingham City Council
"We have deployed open source in some libraries. We have worked on the basis of 200 PCs. In some cases, we have migrated back to Windows," said Les Timms, project manager at the city council. "1,500 was the original plan. It was a figure plucked from the air at the time," Timms told ZDNet UK.
Timms said the council had compared the cost of the Linux desktop migration with an upgrade to Windows XP, and had found that a Microsoft upgrade would be cheaper. Most of the difference was made up of costs attributed to "decision making" and "project management", largely brought about because of a shortage of skills in open-source networking and the changes to IT processes that would result.
The Linux project cost £534,710, while the equivalent XP upgrade would have cost the council £429,960. There were a range of problems with the open-source implementation, Timms said, including desktop interfaces and lack of support for removeable drives.
In the light of the findings, the council has taken the decision to mothball the project.
Timms has now moved jobs to work for Service Birmingham, a joint venture between Birmingham City Council and Capita, which is focusing on increasing business efficiency. Responsibility for the day-to-day running of the council's IT now rests with transformation chief Glyn Evans. Evans told ZDNet UK: "We will continue with a mixed economy [Microsoft and open source]." But he warned, "I'm not an open-source fanatic."
Birmingham's project was funded through the £1.3m central government-backed Open Source Academy (OSA), which has itself faced mixed fortunes.
Although the OSA has supported successful projects such as a 5,500-desktop PC project run by Bristol City Council, it has also been blighted by criticism. The Office of the Deputy Prime Minister withdrew its support from the OSA in March after accusations of poor management and after its dissemination programme flopped spectacularly.
Talkback
Some facts have been omitted from this article which shed further light on the appalling waste of taxpayers money that was the Birmingham City Council's Linux trial:
1) A trial of 4 differently configured Linux desktops (Ubuntu-based) and one Sun Java Desktop machine was held at Birmingham's central library in the summer 2005. A local research company was employed to measure the outcomes of the double-blind trial, specifcally which configuration was viewed as the best by participants. The Linux desktops took the top four spots with Sun's Java Desktop coming in last. Unsurprisingly the report was never published. BCC are a major Sun client.
2) The Open Source community, especially the Open Source Consortium (others included the Gnome Foundation), was entirely excluded from the project after the initial trial. BCC IT's department thought they could undertake the deployment themselves. The failure of this project proves this was not the case.
3) BCC selected an obsolete version of Suse Linux rather than the Ubuntu desktops that won the Library trial. They were unable to replicate the winning desktop configuration because the IT department accidentially erased it.
4) Open Forum Europe managed the Open Source Academy and were responsible for the dissemination programme.
This story just demonstates how the public sector has completely lost focus as to who they serve. If this amount of money had been wasted on such a project in the private sector the persons responsible would have been quite rightly sacked. I doubt those responsible will even get a written warning and no doubt at the time of the project being announced they were courting all the different IT journals about their great experiment.
Maybe ZDNET could publish the names of the inviduals are who are responsible for this collective failure?
I think that the title is comprehensively misleading. If four or five systems administrators at a school in Skegness can roll out 120 computers, which use far more demanding programmes that a library would, for the measly sum of £6,000, then why can the half-million only setup 200 computers?
I think BCC needs a severe dressing down, and instead of moving the project back to XP, hire the Skegness Teachers during their summer break!
http://schoolforge.org.uk/index.php/Skegness_Grammar
It looks like they have not done a proper evaluation.
E.g. it says
"There were a range of problems with the open-source implementation ... including desktop interfaces and lack of support for removeable drives."
If someone can't even use simple desktop interfaces like KDE or GNOME then there could be a case of a computer illiteracy and such a person shouldn't be in IT management in the first place.
And indeed, the article admits:
" costs ... largely brought about because of a shortage of skills in open-source networking and the changes to IT processes that would result."
JNeuhoff
"1,500 was the original plan. It was a figure plucked from the air at the time"
This project was doomed from the start. No competent manager would ever 'pluck figures from the air'. They should have an estimate of the number of PCs required.
I can't believe that this level of incompetence actually exists in BCC. It is truely shocking. Other companies have managed to do this type of project, so why not BCC? Incompetence. Simple.
I just wish the management would get punished for it.
Namibia was able to service their entire school system with linux on desktop. Why cannot a 1st world country with more money and resource not do it?
I think this smacks of poor planning and even worse implementation. The majority of the cost was down to bad management and from a previous posters information exclusion of people that could have made this a success.
Unfortunately this is bad reporting and worse headlines.
http://www.schoolnet.na/
http://www.theregister.co.uk/2002/10/31/namibia_wisely_spurns_m_gift/
Bristol Clueless Council?
Most of the difference was made up of costs attributed to "decision making" and "project management",
I read that as we spent most of the money on lunches, to decide when to have the next freebie.
"largely brought about because of a shortage of skills in open-source networking"
Too busy having lunch to find anyone! Besides anyone who's worked with unix can easily deal with "open-source networking".
As others have commented, pure incompetence.
Just few quotes:
"We have deployed open source in some libraries..."
"...shortage of skills in open-source networking".
I think that it's a real evidence of the reason - why mentioned expensive project was failed.
It looks like BCC project managers split world into two pieces: OpenSource VS Windows.
But in fact Windows is proprietary operation system and OpenSource - model that defines rules for software production and development.
It's very funny - "open-source networking...". Do we have open source
IP protocol ?
I think that shortage of IT skills caused very poor results.
Just few quotes:
"We have deployed open source in some libraries..."
"...shortage of skills in open-source networking".
I think that it's a real evidence of the reason - why mentioned expensive project was failed.
It looks like BCC project managers split world into two pieces: OpenSource VS Windows.
But in fact Windows is proprietary operation system and OpenSource - model that defines rules for software production and development.
It's very funny - "open-source networking...". Do we have open source
IP protocol ?
I think that shortage of IT skills caused very poor results.
The government has a terrible record on delivering IT projects on time and on budget. It has more to do with poor management, analysis and consultancy than the technical suitability of Linux. It's a shame the article doesn't reflect this!
Coming to what we know of the project, the implementation of the Linux pilot I believe was a success, and various members of the linux community, opensource advantage, south birmingham linux group and the open source academy were initially involved in the project.
When I have visited Birmingham Central Library, the linux machines were always in use with a waiting queue and appeared to be more available then the other machines in the library at the time.
The version of linux distribution was chosen as it had all the functionality they required as well as being available and backed by a company like Novell. At the time, 9.3 was the current and supported version and not out of date as suggested. For a big organisation it was important for official support to be available to reassure the rest of the business. According to the presentation given to SBLUG, Ubuntu was assessed and did do well in the initial trails on a equal par to Suse, but the implementation team had prior experience of SuSE and some other distros but not of Ubuntu. In addition certain required enterprise grade applications Birmingham required were only supported by their manufacturers on either SuSE or RedHat.
I believe it was the Birmingham Libraries powers that be that pulled the plug on the rollout as they didn’t feel comfortable about supporting another OS, as they thought it be a steep learning curve for library staff to support.
Cost of SuSE licence, £zero, cost of xp licence £120ish.
Cost of OpenOffice £zero, cost of officexp £150ish.
Cost of Linux AV software £not-required, cost of xp AV £20ish.
Assuming costs of support would be broadly similar once in-house open source experience was at the same level as windows, doesn’t take much comparison to see which is the better value for money.
The project was not a failure due to the OS distribution choice or because the choosen distribution did'nt work I've been told it was simply due to the business (Libraries) not wanting to implement a solution which they saw as a too big change for them, and more work, training for their library staff. The trial dogstar is on about was not the pilot but an initial trail run by libraries to see what the public prefered as the look and feel for the desktop. But Ubuntu (who clearly dogstar is an advocate of) was not chosen as the maturity of the distribution and official support was not proven.
I suggesst people read the report that was published for the Opwn Source Academy before making up their minds and placing blame.
http://www.opensourceacademy.gov.uk/solutions/casestudies/birminham-city-council
The project was not a failure due to the OS distribution choice or because the choosen distribution did'nt work I've been told it was simply due to the business (Libraries) not wanting to implement a solution which they saw as a too big change for them, and more work, training for their library staff. The trial dogstar is on about was not the pilot but an initial trail run by libraries to see what the public prefered as the look and feel for the desktop. But Ubuntu (who clearly dogstar is an advocate of) was not chosen as the maturity of the distribution and official support was not proven.
I suggesst people read the report that was published for the Opwn Source Academy before making up their minds and placing blame.
http://www.opensourceacademy.gov.uk/solutions/casestudies/birminham-city-council
Having watched this rather ill-informed debate for a week or two, I feel I must add my thoughts.
I was at a presentation given to SouthBirmingham Linux Users Group at the inception of the project. At this event the team presented their state of the project and reasons for selecting the distributions for server and client use.
Despite comments to the contrary, Suse 9.3 was the current, supported release at that time. Suse 10 only became supported part way through the project and imho, would not off a sigificant improvement to justify swapping to as part of a pilot. For a production rollout, yes, pilot, no.
Ubuntu was not chosen as the desktop for a number of reasons, all of which were justified by facts, many of the people who have criticised the chouce of SuSE over Ubuntu clearly either support Ubuntu commercially or are involved in some other way, so I suspect these comments represent sour grapes and as such should be disregarded.
The team certainly did not give the impression of wanting to exclude the Linux community, in fact quite the opposite. The intention to set up a website to track the progress was proposed subject to management approval. It's non appearance is probably down to being too "unofficial" for a local authority.
Apart from the ridiculous 1500 desktops quote "plucked from the air", 200 desktops out of 400 is probably not far off the mark, my maths are as follows:-
41 district libraries and 1 central library.
40 public access computers in the central library (10 each on 4 floors) and average of 4 in each district makes 161+40 = 201.
Assuming more or less the same for staff gives a total of 400 ish. So deploying 200 for public access is not far off what I would expect, so the "with just 200 Linux PCs being deployed" quote in the ZDnet article smacks of the "not enough research school of journalism".
As for the cost, anyone who has worked on any local government project will know that if £x is allocated to a project, £x will be spent if the project runs to budget, it may however overspend in which case it could be called a waste of money. Also every minute of time will be recorded on a timesheet to claim the funds. Given that the staff were employed by BCC (from the presentation), had the grant not been forthcoming and the staff done this as part of their daily work, the cost excluding hardware would have been close to £zero (I think the cost of the new hardware was included in the project).
The disgrace lies with the way public funding is allocated, not how much is spent.
As for the "range of problems" mentioned, on the ocasions I visited the Central Library, all the computers were in use and a booking system was in use to reserve time. Unsurprisingly the number of Linux computers not working the the Windows ones not working was rather low. So I am not sure what the problems were but I did not find any. Internet access was fine, usb devices worked fine and there was a good range of software available.
Just as important on a public access machine, the user settings were cleared between sessions, so saved work / internet links from previous users were erased in the interests of privacy. I could only find one place in the filesystem which had not been locked down and that wasn't obvious.
In conclusion, Dogstar really needs to get his facts in order, point two is totally wrong, I have spoken with several members of the team at different times about the best way to do something, just because dogstar wasn't asked doesn't mean no one else was. Point 3 is again wrong, Suse was current during the development process, and from what I heard at the presentation, the initial trial was looking at the desktop not the distribution (ie. Gnome vs KDE). As it happens, I heard that the split was about equal between those favouring Gnome over KDE.
Having watched this rather ill-informed debate for a week or two, I feel I must add my thoughts.
I was at a presentation given to SouthBirmingham Linux Users Group at the inception of the project. At this event the team presented their state of the project and reasons for selecting the distributions for server and client use.
Despite comments to the contrary, Suse 9.3 was the current, supported release at that time. Suse 10 only became supported part way through the project and imho, would not off a sigificant improvement to justify swapping to as part of a pilot. For a production rollout, yes, pilot, no.
Ubuntu was not chosen as the desktop for a number of reasons, all of which were justified by facts, many of the people who have criticised the chouce of SuSE over Ubuntu clearly either support Ubuntu commercially or are involved in some other way, so I suspect these comments represent sour grapes and as such should be disregarded.
The team certainly did not give the impression of wanting to exclude the Linux community, in fact quite the opposite. The intention to set up a website to track the progress was proposed subject to management approval. It's non appearance is probably down to being too "unofficial" for a local authority.
Apart from the ridiculous 1500 desktops quote "plucked from the air", 200 desktops out of 400 is probably not far off the mark, my maths are as follows:-
41 district libraries and 1 central library.
40 public access computers in the central library (10 each on 4 floors) and average of 4 in each district makes 161+40 = 201.
Assuming more or less the same for staff gives a total of 400 ish. So deploying 200 for public access is not far off what I would expect, so the "with just 200 Linux PCs being deployed" quote in the ZDnet article smacks of the "not enough research school of journalism".
As for the cost, anyone who has worked on any local government project will know that if £x is allocated to a project, £x will be spent if the project runs to budget, it may however overspend in which case it could be called a waste of money. Also every minute of time will be recorded on a timesheet to claim the funds. Given that the staff were employed by BCC (from the presentation), had the grant not been forthcoming and the staff done this as part of their daily work, the cost excluding hardware would have been close to £zero (I think the cost of the new hardware was included in the project).
The disgrace lies with the way public funding is allocated, not how much is spent.
As for the "range of problems" mentioned, on the ocasions I visited the Central Library, all the computers were in use and a booking system was in use to reserve time. Unsurprisingly the number of Linux computers not working the the Windows ones not working was rather low. So I am not sure what the problems were but I did not find any. Internet access was fine, usb devices worked fine and there was a good range of software available.
Just as important on a public access machine, the user settings were cleared between sessions, so saved work / internet links from previous users were erased in the interests of privacy. I could only find one place in the filesystem which had not been locked down and that wasn't obvious.
In conclusion, Dogstar really needs to get his facts in order, point two is totally wrong, I have spoken with several members of the team at different times about the best way to do something, just because dogstar wasn't asked doesn't mean no one else was. Point 3 is again wrong, Suse was current during the development process, and from what I heard at the presentation, the initial trial was looking at the desktop not the distribution (ie. Gnome vs KDE). As it happens, I heard that the split was about equal between those favouring Gnome over KDE.
I know all about this type of management,
these people are hierachical, most stupid at the top.
no, not kidding. typical realworld example !.
they are obviously in search of some bright minds which are now completely lacking.
oh yeah, F/OSS is about openness, that is NOT being locked out.
yes it is cheaper and available, but it is not about the money folks.