UK Government Mandates 'Preference' For Open Source

An anonymous reader writes "ComputerWeekly reports that the U.K. government 'has, for the first time, mandated a preference for using open source software for future developments.' This comes from the newly released version of the Government Service Design Manual, which has a section about when government agencies should use open source. It says: 'Use open source software in preference to proprietary or closed source alternatives, in particular for operating systems, networking software, web servers, databases and programming languages.' The document also warns against vendor lock-in. This policy shift comes under the direction of government CTO Liam Maxwell, who said, 'In digital public services, open source software is clearly the way forward.' He added, 'We're not dogmatic about this – we'll always use the best tool for the job – but open source has major advantages for the public sector.'"

Comments

Is this real?

[rsilvergun]

anyone on the other side of the pond know if this is a real attempt to push OSS software or if it's just another attempt to get discounted Microsoft software?

Re:Is this real?

[K. S. Kyosuke]

it's just another attempt to get discounted Microsoft software?

Of course it is! What else did you think?

-- Sir Humphrey Appleby

Re:Is this real?

[benilov]

You betcha. The Government Service Design Manual comes from GDS, a part of the Cabinet Office. GDS also created GOV.UK - the new single domain for the UK government. The GOV.UK stack is almost entirely open-source software, which can be found on Github under the Open Government License.

Re:Is this real?

[Anonymous Coward]

Posting AC.

There are certain public services in the UK that have real issues at the moment, IT-wise, due to the general austerity measures in place to reduce the deficit.

There are large sections of the UK police force stuck using IE6 due to dependancies on ActiveX.
XP is being EOL'ed next year.
The money isn't there to deal with the situation.

There's a lot of people campaigning for a move to open-source so nothing like this happens again.

Re:Is this real?

[Dr_Barnowl]

Yeah, my lot have to manage the transition to Windows 7 for a whole bunch of bespoke applications. We got shot of IE6 and heavens, we were glad, because our stupid timesheet software used ActiveX so we had to ditch that too.

The only thing really holding us back from moving to Linux is MS Office. The NHS had an enterprise-wide license, which a back-of-napkin estimate says must have cost on the order of £100M per year. That got dropped a while ago, I'm guessing because it was a big fat line item in the budget and made a ripe target for people saying "hey, what if we spent some small fraction of that on LibreOffice development?".

A lot of our bespoke apps are Java and thus don't really need Windows to work. Web apps are web apps.

But we, like everywhere, I suspect, have a large number of things cobbled together from VBA and spit, not to mention the things people do with Access. Any coherent plan to move to Linux, or even LibreOffice, needs a department dedicated to migrating VBA and Access applications.

Re:Is this real?

[rtb61]

You have to consider that for other countries M$ is a dead loss, tons of money going out with no return. Pushing FOSS means that if an offshot of a major campaign contributing company sets up in that market you can readily funnel money to them and look really good in the polls when doing so. Basically FOSS in also going to be a double plus win for pollies.

It's only a suggestion, not a mandate

[menot]

"We're not dogmatic about this – we'll always use the best tool for the job".
That's one of the most interesting points in the article. More people should think like that. In the end, software is just a tool.

Re:Cost for software vs skill set

[DamonHD]

Things have changed for the better for Windows I am quite sure, but back in the days when I was a UNIX sysadmin for a living you needed 10x as many Windows admins as UNIX admins for the same number of machines / user seats, so a simple salary ratio would be misleading!

Rgds

Damon

It's not enough

[Stormwatch]

Governments should be forbidden from using non-Free software. Go ahead and get your company into whatever vendor lock-in you want, but public data should never be subjected to it.

Re:It's not enough

[Stormwatch]

Open Source cannot compete on bribes with proprietary software

Fix'd.

Re:It's not enough

[whoever57]

Governments should be forbidden from using non-Free software. Go ahead and get your company into whatever vendor lock-in you want, but public data should never be subjected to it.

No. This is wrong. Governments should be required to use open standards. Thus allowing open and closed source offerings to compete.

Furthermore, if it turns out that a supplier claimed compliance with an open standard but did not deliver this, there should be serious penalties levied against the supplier (and not just a slap on the wrist that the supplier will see as merely "cost of doing business"). The penalties could include requiring the supplier to make their version of the standard open to all.

Re:It's not enough

[maxwell demon]

Another requirement should be that the supplier allows the government to inspect the source code in order to make sure there are no backdoors in the code. With Open Source, this is automatic; for Closed Source solutions, it would be an additional requirement in the contract.

Openness is the customer requirement

[Anonymous Coward]

You are clearly not basing this on what is the best tool (even if the open source happens to be the best tool).

You're contradicting yourself.

Openness has clear and undoubted benefits for the public sector, and so not surprisingly this customer made openness a default requirement. He's not mandating against proprietary software, but if a software company can't give him the desired openness then it's not fulfilling his requirement. Given his requirement, open source tools are the best tools by default, but not the only ones.

The customer decides the requirements, not the provider. Live with it.

Also, closed source and backdoors

[Anonymous Coward]

Governments should be forbidden from using non-Free software.

Here's another reason which underlines your point:

- A government has no mandate to entrust the country's data to a corporation nor to allow it to leak. It is therefore simply not permissible to allow that data to be processed by closed source software which by definition cannot be trusted.

The above should be self-evident, but in case it's not, objectors would do well to ponder the acknowledged backdoors in Skype and in a variety of Chinese routers. With open source, this cannot easily happen.

Re:Open file formats should be mandatory

[Dr_Barnowl]

I have to disagree. Most of the formats I see developed this way end up horrible messes because they hire a whole bunch of consultants to do the work.

The difficulty with that is that contractors are paid by the hour, so you don't get

* Re-use of other standards where appropriate

I've seen people reinvent the wheel so many times it's not true. This is true from simple little things like time values in XML (xsd:time sensibly uses ISO8601, this lot made up their own format, with ensuing hilarity when implementers think that their standard XML tool kit date / time types will produce valid documents), diagram formats (they just copied another standard verbatim into their documents rather than saying - "Hey, lets use this standard and say so"), and document formats (they didn't like the ability of XHTML to have script tags in it, so they copied THAT as well).

* Simplicity

Simple designs that work don't generate billable hours. Complex monsters that require hours of argument over the finer points of what they actually mean, do.

* Implementations

Implementations are essential for the development of standards. If you don't implement them, you don't get any kind of feel for the actual needs of the problem domain and how well your design is solving them. Alas, standards developed by publicly funded committee in my experience don't bother with this, and typically don't include any actual software engineers to tell them what problems they might be causing for implementers down the line.

Things like pretending an identifier is an integer when all the handling means you have to treat it like a string (it consists of four separate fields, one of them optional, but as a stream of digits and not bytes). Or taking a set of metadata that you have to understand to read the data, and .. embedding that data inside the data itself. Or creating an abstract data type with a contract and then insisting that people store it without thinking about it's concrete requirements.

Formats thought up by corporations at least have the benefit of their creators not wanting to spend as much time as possible debating the finer points of the thing. They want something that works, but as evidenced by MOO-XML, practicality often means they end up with a real mess as well - but at least it's a real mess, and not just a theoretical mess.

I think "Open" is more important than "Standard". "Standard" gives the appearance of authority, but "Open" means you have a chance of things being useful.

MOO-XML is a horrifying mess. Not even MS Office implements it. It's a "standard", having been ratified by ISO, but nothing about it's development was "open".

FreeMind is a small java mind-map program. FreeMind format isn't a "standard", but it is "open". And it is useful - useful enough that most of the other mind-map programs will import it. You can open the files up in a text editor, or feed them through XSLT, or consume them with a program and do interesting things with them. And if you want a feature implemented in it, you can patch the sources, and even feed the patch back upstream.

I think collaboration on trying to solve a problem benefits from some actual problem solving, rather than just talking about what the problem might be and how it might be solved if so.